introduction to mineral exploration

499

Upload: claudia-lopez

Post on 16-May-2015

3.303 views

Category:

Education


23 download

DESCRIPTION

exploracion minera

TRANSCRIPT

Page 1: Introduction to mineral exploration
Page 2: Introduction to mineral exploration

Introduction toMineral Exploration

Second Edition

Edited by Charles J. Moon, Michael K.G. Whateley& Anthony M. Evans

With contributions fromWilliam L. Barrett

Timothy BellAnthony M. Evans

John MilsomCharles J. MoonBarry C. Scott

Michael K.G. Whateley

Page 3: Introduction to mineral exploration
Page 4: Introduction to mineral exploration

introduction tomineral exploration

Page 5: Introduction to mineral exploration
Page 6: Introduction to mineral exploration

Introduction toMineral Exploration

Second Edition

Edited by Charles J. Moon, Michael K.G. Whateley& Anthony M. Evans

With contributions fromWilliam L. Barrett

Timothy BellAnthony M. Evans

John MilsomCharles J. MoonBarry C. Scott

Michael K.G. Whateley

Page 7: Introduction to mineral exploration

Copyright © 2006 by Charles J. Moon, Michael K.G. Whateley, and Anthony M. Evans

BLACKWELL PUBLISHING350 Main Street, Malden, MA 02148-5020, USA9600 Garsington Road, Oxford OX4 2DQ, UK550 Swanston Street, Carlton, Victoria 3053, Australia

The rights of Charles J. Moon, Michael K.G. Whateley, and Anthony M. Evans to be identified as the Authorsof this Work have been asserted in accordance with the UK Copyright, Designs, and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted,in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permittedby the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher.

First edition published 1995 by Blackwell Publishing LtdSecond edition published 2006

1 2006

Library of Congress Cataloging-in-Publication Data

Introduction to mineral exploration.–2nd ed. / edited by Charles J. Moon, Michael K.G. Whateley & Anthony M.Evans; with contributions from William L. Barrett . . . [et al.].

p. cm.Evans’s appears first on the previous ed.Includes bibliographical references and index.ISBN-13: 978-1-4051-1317-5 (pbk. : acid-free paper)ISBN-10: 1-4051-1317-0 (pbk. : acid-free paper) 1. Prospecting. 2. Mining geology. I. Moon, Charles J. II.

Whateley, M.K.G. III. Evans, Anthony M.

TN270.I66 2006622′.1—dc22

2005014825

A catalogue record for this title is available from the British Library.

Set in 9/11pt Trump Mediaevalby Graphicraft Limited, Hong KongPrinted and bound in Indiaby Replika Press PVT Ltd

The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, andwhich has been manufactured from pulp processed using acid-free and elementary chlorine-free practices.Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmentalaccreditation standards.

For further information onBlackwell Publishing, visit our website:www.blackwellpublishing.com

Page 8: Introduction to mineral exploration

Contents

Contributors ixPreface to the Second Edition xiPreface to the First Edition xiiUnits, Abbreviations, and Terminology xiii

Part I – Principles 1

1 Ore, Mineral Economics, and Mineral Exploration 3Charles J. Moon and Anthony M. Evans

1.1 Introduction 31.2 Mineral economics 31.3 Important factors in the economic recovery of minerals 111.4 Structure of the mining industry 121.5 Some factors governing the choice of exploration areas 131.6 Rationale of mineral exploration 141.7 Further reading 18

2 The Mineralogy of Economic Deposits 19Anthony M. Evans

2.1 Introduction 192.2 Mineralogical investigations 252.3 Further reading 32

3 Mineral Deposit Geology and Models 33Anthony M. Evans and Charles J. Moon

3.1 Nature and morphology of orebodies 333.2 Wall rock alteration 393.3 Geological models of mineral deposits 403.4 Further reading 453.5 Summary of Chapters 1, 2, and 3 463.6 Appendix: mineral deposit model classification of Cox and Singer (1986) 46

4 Reconnaissance Exploration 52Charles J. Moon and Michael K.G. Whateley

4.1 Exploration planning 534.2 Desk studies and reconnaissance 57

Page 9: Introduction to mineral exploration

4.3 Sustainable development 674.4 Summary 694.5 Further reading 69

5 From Prospect to Prefeasibility 70Charles J. Moon and Michael K.G. Whateley

5.1 Finding a drilling target 705.2 When to drill and when to stop 845.3 Recycling prospects 935.4 The Bre-X Minerals saga 1005.5 Summary 1025.6 Further reading 102

6 Remote Sensing 104Michael K.G. Whateley

6.1 Introduction 1046.2 The Landsat System 1056.3 Other imaging systems 1166.4 Photogeology 1176.5 Summary 1256.6 Further reading 126

7 Geophysical Methods 127John Milsom

7.1 Introduction 1277.2 Airborne surveys 1277.3 Magnetic surveys 1307.4 Gravity method 1337.5 Radiometrics 1357.6 Resistivity 1367.7 Spontaneous polarization (SP) 1387.8 Induced polarization (IP) 1387.9 Continuous-wave electromagnetics 1427.10 Transient electromagnetics (TEM) 1447.11 Remote source methods (VLF and magnetotellurics) 1457.12 Seismic methods 1477.13 Ground radar 1487.14 Borehole geophysics 1497.15 Geophysics in exploration programs 1507.16 Integration of geological and geophysical data 1507.17 The role of the geologist in geophysical exploration 1517.18 Dealing with contractors 1527.19 Further reading 153

8 Exploration Geochemistry 155Charles J. Moon

8.1 Planning 1558.2 Analysis 1598.3 Interpretation 1628.4 Reconnaissance techniques 1648.5 Follow-up sampling 177

vi CONTENTS

Page 10: Introduction to mineral exploration

8.6 Summary 1788.7 Further reading 178

9 Mineral Exploration Data 179Charles J. Moon and Michael K.G. Whateley

9.1 Data capture and storage 1799.2 Data integration and geographical information systems 1869.3 Integration with resource calculation and mine planning software 1979.4 Further reading 198

10 Evaluation Techniques 199Michael K.G. Whateley and Barry C.Scott

10.1 Sampling 19910.2 Pitting and trenching 21810.3 Drilling 21810.4 Mineral resource and ore reserve estimation, grade calculations 23010.5 Geotechnical data 24310.6 Summary 24910.7 Further reading 25010.8 Appendix: use of Gy’s sampling formula 250

11 Project Evaluation 253Barry C.Scott and Michael K.G. Whateley

11.1 Introduction 25311.2 Value of mineralisation 25311.3 Cash flow 26011.4 Study definitions 26211.5 Mineral project valuation and selection criteria 26511.6 Risk 27011.7 Inflation 27111.8 Mineral project finance 27211.9 Summary 27611.10 Further reading 276

Part II – Case Studies 279

12 Cliffe Hill Quarry, Leicestershire – Development of Aggregate Reserves 281Michael K.G. Whateley and William L. Barrett

12.1 Introduction 28112.2 Phase 1 – order-of-magnitude study 28212.3 Phase 2 – pre-feasibility study 28312.4 Phase 3 – feasibility study 28412.5 Phase 4 – detailed engineering 29112.6 Quarrying and environmental impact 29212.7 Summary 293

13 Soma Lignite Basin, Turkey 294Michael K.G. Whateley

13.1 Introduction 29413.2 Exploration programs 295

CONTENTS vii

Page 11: Introduction to mineral exploration

13.3 Geology 29813.4 Data assessment 30313.5 Geotechnical investigation 30713.6 Lignite quality 30913.7 Lignite reserve estimates 31413.8 Surface mine evaluation 31513.9 Summary 31913.10 Further reading 319

14 Witwatersrand Conglomerate Gold – West Rand 320Charles J. Moon and Michael K.G. Whateley

14.1 Introduction 32014.2 Geology 32014.3 Economic background 33014.4 Exploration models 33314.5 Exploration methods 33414.6 Reserve estimation 34414.7 Exploration and development of the southern part of the West Rand Goldfield 34514.8 Conclusions 357

15 A Volcanic-associated Massive Sulfide Deposit – Kidd Creek, Ontario 358Anthony M. Evans and Charles J. Moon

15.1 Introduction 35815.2 Volcanic-associated massive sulfide deposits 35815.3 The Kidd Creek Mine 37015.4 Conclusions 381

16 Disseminated Precious Metals – Trinity Mine, Nevada 382Michael K.G. Whateley, Timothy Bell and Charles J. Moon

16.1 Background 38216.2 Trinity Mine, Nevada 38616.3 Exploration 38716.4 The geology of Pershing County and the Trinity District 39116.5 Development and mining 39516.6 Mineral processing 39716.7 Environmental considerations 39716.8 Grade estimation 39916.9 Ore reserve estimation 40916.10 Summary and conclusions 412

17 Diamond Exploration – Ekati and Diavik Mines, Canada 413Charles J. Moon

17.1 World diamond deposits and exploration 41317.2 Canadian discoveries: Ekati and Diavik 43117.3 Summary 443

Internet Links 444References 447Index 469

viii CONTENTS

Page 12: Introduction to mineral exploration

Contributors

WILLIAM L. BARRETT Tarmac Quarry Products Ltd, Millfields Road, Ettingshall, Wolverhamp-ton WV4 6JP. Previously Geological Manager, Estates and Environment Department of Tarmac(now Consultant) but he has enjoyed a varied career in applied geology. His first degree was takenat the University of Wales and he obtained his doctorate from Leeds University. Much of his earlycareer was spent in base metal exploration and underground mining in East Africa and this cul-minated with a spell as Chief Geologist at Kilembe Mines. Latterly he has been concerned withall aspects of Tarmac’s work in the field of industrial minerals.

TIMOTHY BELL Bell Hanson, Foss Road, Bingham NG11 7EP. Graduated from West London Col-lege with a first class degree (cum laude) in Geology and Geography before obtaining an MSc degreewith distinction in Mineral Exploration and Mining Geology at Leicester University in 1989. Hejoined CP Holdings (opencast coal contractors) as geologist in charge of computer applications inthe mining division before moving to SURPAC Software International where he was TechnicalManager in their Nottingham office and is now running his own business.

ANTHONY M. EVANS 19 Hall Drive, Burton on the Wolds, Loughborough LE12 5AD. FormerlySenior Lecturer in charge of the MSc Course in Mining Geology and Mineral Exploration andthe BSc Course in Applied Geology at Leicester University. A graduate of Liverpool University inPhysics and Geology he obtained a PhD at Queen’s University, Ontario and worked in Canada forthe Ontario Department of Mines spending much time investigating uranium and other mineral-isation. In Europe he has carried out research on many aspects of mineralisation and industrialmineralogy and is the editor and author of a number of books and papers in these fields. For anumber of years he was Vice-President of the International Society for Geology Applied to MineralDeposits.

JOHN MILSOM Department of Geological Sciences, University College, Gower St, London WC1E6BT. Senior Lecturer in Exploration Geophysics at University College. He graduated in Physicsfrom Oxford University in 1961 and gained the Diploma of Imperial College in Applied Geophysicsthe following year. He joined the Australian Bureau of Mineral Resources as a geophysicistand worked in the airborne, engineering, and regional gravity sections. He obtained a PhD fromImperial College for his research on the geophysics of eastern Papua and joined the staff therein 1971. Then came a spell as Chief Geophysicist for Barringer Research in Australia followed byone as an independent consultant when his clients included the UNDP, the Dutch Overseas AidMinistry, and BP Minerals. In 1981 he joined the staff at Queen Mary College to establish theExploration Geophysics BSc Course which later transferred to University College. He is the authorof a textbook on field geophysics.

CHARLES J. MOON Department of Geology, University of Leicester, Leicester, LE1 7RH. Agraduate in Mining Geology from Imperial College, London in 1972, he returned some years later to

Page 13: Introduction to mineral exploration

take his PhD in Applied Geochemistry. He has had wide experience in the fields of mineral depositassessment and exploration all over the globe. He has worked for Consolidated Gold Fields search-ing for tin deposits in Cornwall, the South African Geological Survey, Esso Minerals AfricaInc., with whom he was involved in the discovery of three uranium deposits. Besides his generalexpertise in mineral exploration he has developed the use of raster and vector based GIS systemsand courses in these techniques, particularly in ArcGIS for exploration integration. His presentposition is that of Senior Lecturer in Mineral Exploration at the University of Leicester.

BARRY C. SCOTT Knoyle House, Vicarage Way, Gerrards Cross, Bucks SL9 8AS. His most recentposition was Lecturer in Mining Geology and Mineral Exploration at Leicester University fromwhich he retired in 1993. Previously he had occupied senior positions with a number of miningcompanies and academic institutions. These included Borax Consolidated Ltd, Noranda MinesLtd, G. Wimpey plc, the Royal School of Mines, and Imperial College. Dr Scott obtained his BScand PhD from Imperial College and his MSc from Queen’s University, Ontario. He was a Memberof Council of the Institution of Mining and Metallurgy for many years, was its President andTreasurer.

MICHAEL K.G. WHATELEY Rio Tinto Technical Services, Principal Consultant-Manager. Hegraduated in Geology and Geography from London University in 1968, gained an MSc from theUniversity of the Witwatersrand in 1980 and a PhD from Leicester University in 1994 for his workon the computer modeling of the Leicestershire Coalfield. He worked in diamond exploration inBotswana and on gold mines near Johannesburg before joining Hunting Geology and Geophysics inSouth Africa. He then joined Utah International Inc. in charge of their Fuel Division in southernAfrica. In 1982 he returned to the UK to work for Golder Associates before taking up a post at theUniversity of Leicester as Lecturer in Mining Geology. Since 1996 he has been a consultant withRio Tinto Technical Services based in Bristol.

x CONTRIBUTORS

Page 14: Introduction to mineral exploration

Preface to theSecond Edition

Since the first edition was published much has changed in the mineral exploration business. It hasbeen through one of the sharpest downturns since the 1930s and the industry has become truly glo-bal, with many mergers and acquisitions of mining and exploration companies. Indeed, conditionsat the time at the time of writing appear to have moved back to boom driven by increasing demand.

We have tried to maintain a balance between principles and case histories, based on our experi-ence of teaching at senior undergraduate and postgraduate levels. We have attempted to reflect thechanging mineral exploration scene by rewriting the sections on mineral economics (Chapter 1)and the increased emphasis on public accountability of mineral exploration companies (Chapters10 and 11), as well as updating the other chapters. One of the major areas of advance has beenthe widespread use of information technology in mineral exploration. We have added a chapter(Chapter 9) dealing with the handling of mineral exploration data. Although we do not advocategeologists letting their hammers get rusty, the correct compilation of data can help expeditefieldwork and improve understanding of deposits for feasibility studies.

The case histories are largely the same as the first edition although we have replaced a studyon tin exploration with one on diamond exploration. This details one of the major explorationsuccesses of the 1990s near Lac de Gras in arctic Canada (Chapter 17).

Another of the trends of the last few years has been the decline in the teaching of mining-relatedsubjects at universities, particularly in Europe and North America. We hope that the contents ofthis book will help to educate the geologists needed to discover the mineral deposits required forfuture economic growth, whether or not they are studying at a university.

The two of us have taken over editorship from Tony Evans who has completely retired. We areboth based in western Europe, although we manage to visit other continents when possible, andone of us was commuting to southern Africa when this edition was completed. Inevitably theexamples we mention are those with which we are familiar and many readers’ favorite areas will bemissing. We trust, however, that the examples are representative.

We owe a debt of thanks to reviewers of the additions: Ivan Reynolds of Rio Tinto forcomments on Chapter 2; Jeremy Gibbs of Rio Tinto Mining and Exploration Limited for his helpwith Chapter 6; Peter Fish, John Forkes, and Niall Weatherstone of Rio Tinto Technical Servicesfor their helpful comments on Chapters 10 and 11; Paul Hayston of Rio Tinto Mining and Explora-tion Limited for comments on John Milsom’s update of Chapter 7; and Andy Davy of Rio TintoMining and Exploration Limited and Gawen Jenkin of the University of Leicester for comments onChapter 17. They are, of course, not responsible for any errors that remain our own. Lisa Barber isthanked for drawing some of the new figures.

Finally, we would like to thank our families for their forbearance and help. Without them thecompilation of this second edition would not have been possible.

Charlie Moon and Mike WhateleyLeicester and Bristol

Page 15: Introduction to mineral exploration

Preface to theFirst Edition

It is a matter of regret how little time is devoted to teaching mineral exploration in first degree uni-versity and college courses in the United Kingdom and in many other countries. As a result the fewavailable textbooks tend to be expensive because of the limited market. By supplying our publisherwith camera ready copy we have attempted to produce a cheaper book which will help to redressthis neglect of one of the most important aspects of applied geology. We hope that lecturers willfind here a structure on which to base a course syllabus and a book to recommend to their students.

In writing such a book it has to be assumed that the reader has a reasonable knowledge of thenature of mineral deposits. Nevertheless some basic facts are discussed in the early chapters.

The book is in two parts. In the first part we discuss the principles of mineral exploration and inthe second, case histories of selected deposit types. In both of these we carry the discussions rightthrough to the production stage. This is because if mining projects are to progress successfully fromthe exploration and evaluation phases to full scale production, then exploration geologists mustappreciate the criteria used by mining and mineral processing engineers in deciding upon appropri-ate mining and processing methods, and also by the financial fraternity in assessing the economicviability of a proposed mining operation. We have tried to cover these and many other relatedsubjects to give the reader an overall view of mineral exploration. As this is only an introductorytextbook we have given some guidance with regard to further reading rather than leaving thestudent to choose at random from references given in the text.

This textbook arises largely from the courses in mineral exploration given in the GeologyDepartment of Leicester University during the last three decades and has benefitted from the ideasof our past and present colleagues and to them, as well as to our students and our friends in indus-try, we express our thanks. They are too many to name but we would like in particular to thankDave Rothery for his help with Chapter 6; Martin Hale for discussion of Chapter 8; Nick Laffoley,Don Moy and Brendon Monahan for suggestions which have improved Chapter 9; and John Rickusfor a helping hand with Chapter 12. Steve Baele of US Borax gave us good advice and Sue Button andClive Cartwright performed many miracles of draughting in double quick time. In addition we aregrateful to the Director-General of MTA for permission to reproduce parts of the maps used toprepare Figs 12.2 and 12.3, and to the Director-General of TKI for permission to publish the dataon Soma.

Page 16: Introduction to mineral exploration

Units, Abbreviations,and Terminology

NOTE ON UNITS

With few exceptions the units used are all SI (Système International), which has been in commonuse by engineers and scientists since 1965. The principal exceptions are: (a) for commodity pricesstill quoted in old units, such as troy ounces for precious metals and the short ton (= 2000 lb);(b) when there is uncertainty about the exact unit used, e.g. tons in certain circumstances mightbe short or long (2240 lb); and (c) Chapter 16 where all the original information was collectedin American units, but many metric equivalents are given.

SI prefixes and suffixes commonly used in this text are k = kilo-, 103; M = mega-, 106 (million);G = giga-, 109.

SOME ABBREVIATIONS USED IN THE TEXT

AAS atomic absorption spectrometryASTM American Society for Testing MaterialsCCT computer compatible tapesCIF carriage, insurance, and freightCIPEC Conseil Inter-governmental des Pays Exportateurs de Cuivre (Intergovernmental

Council of Copper Exporting Countries)CIS Commonwealth of Independent States (includes many of the countries formerly in

the USSR)d.c. direct currentDCF ROR discounted cash flow rate of returnDTH down-the-hole (logs)DTM digital terrain modelEEC see EUEIS environmental impact statementEM electromagneticERTS Earth Resources Technology SatelliteEU European Union sometimes still referred to as EECFOB freight on boardFOV field of viewGA Golder AssociatesGIS Geographical Information SystemsGPS global positioning satellitesGRD ground resolution distance

Page 17: Introduction to mineral exploration

HRV high resolution visibleICP-ES inductively coupled plasma emission spectroscopyIFOV instantaneous field of viewIGRF International Geomagnetic Reference FieldINPUT induced pulse transient systemIP induced polarizationIR infraredITC International Tin CouncilMS multispectralMSS multispectral scannersMTA Maden Tetkik ve AramaNAA neutron activation analysisNAF North Anatolian FaultNPV net present valueOECD Organization for Economic Co-operation and DevelopmentOPEC Organization of Petroleum Exporting CountriesPFE percent frequency effectPGE platinum group elementsPGM platinum group metalsppb parts per billionppm parts per millionREE rare earth elementsRMR rock mass ratingROM run-of-mineRQD rock quality designationSEM scanning electron microscopySG specific gravitySI units Système International UnitsSLAR side-looking airborne radarTDRS Tracking and Data Relay SystemTEM transient electromagneticsTKI Turkiye Komur Isletmekeri KurumuTM thematic mappert p.a. tonnes per annumt p.d. tonnes per diemUSGS United States Geological SurveyUSSR the geographical area that once made up the now disbanded Soviet RepublicsVLF very low frequencyVMS volcanic-associated massive sulfide (deposits)XRD X-ray diffractionXRF X-ray fluorescence

NOTE ON TERMINOLOGY

Mineralisation: “Any single mineral or combination of minerals occurring in a mass, or deposit, ofeconomic interest. The term is intended to cover all forms in which mineralisation might occur,whether by class of deposit, mode of occurrence, genesis, or composition” (Australasian JORC2003). In this volume we have spelt the term mineralisation to be consistent with the JORC codealthough it is usually spelt mineralization in the USA. For further discussion see section 1.2.1.

xiv UNITS, ABBREVIATIONS, AND TERMINOLOGY

Page 18: Introduction to mineral exploration

Part I

Principles

Page 19: Introduction to mineral exploration
Page 20: Introduction to mineral exploration

1Ore, Mineral Economics,

and Mineral ExplorationCharles J. Moon and Anthony M. Evans

1.1 INTRODUCTION

A large economic mineral deposit, e.g. 200 Mtunderlying an area of 2 km2, is minute in com-parison with the Earth’s crust and in mostcountries the easily found deposits croppingout at the surface have nearly all been found.The deposits for which we now search arelargely concealed by weathered and leachedoutcrops, drift, soil, or some other cover, andsophisticated exploration methods are requiredto find them. The target material is referredto as a mineral deposit, unless we use a morespecific term such as coal, gas, oil, or water.Mineral deposits contain mineral resources.What sort of mineral deposit should we seek?To answer this question it is necessary to havesome understanding of mineral economics.

1.2 MINERAL ECONOMICS

1.2.1 Ore

Ore is a word used to prefix reserves or body butthe term is often misused to refer to any or allin situ mineralisation. The Australasian JointOre Reserves Committee (2003) in its code, theJORC Code, leads into the description of orereserves in the following way: “When the loca-tion, quantity, grade, geological characteristicsand continuity of mineralisation are known,and there is a concentration or occurrenceof the material of intrinsic economic interestin or on the Earth’s crust in such form andquantity that there are reasonable prospectsfor eventual economic extraction, then this

deposit can be called a mineral resource.”“Mineral Resources are subdivided, in order ofincreasing geological confidence, into Inferred,Indicated, and Measured categories.” A morecomplete explanation is given in section 10.4.1.

The JORC Code then explains that “an orereserve is the economically mineable part ofa Measured or Indicated Mineral Resource.It includes diluting materials and allowancesfor losses that may occur when the materialis mined. Appropriate assessments, which mayinclude feasibility studies (see section 11.4),have been carried out, and include consid-eration of and modification by realisticallyassumed mining, metallurgical, economic,marketing, legal, environmental, social, andgovernmental factors. These assessmentsdemonstrate at the time of reporting thatextraction could reasonably be justified.” “OreReserves are sub-divided in order of increasingconfidence into Probable Ore Reserves andProved Ore Reserves.”

“The term ‘economic’ implies that extrac-tion of the ore reserve has been establishedor analytically demonstrated to be viableand justifiable under reasonable investmentassumptions. The term ore reserve need notnecessarily signify that extraction facilities arein place or operative or that all governmentalapprovals have been received. It does signifythat there are reasonable expectations of suchapprovals.” An orebody will be the portion ofa mineralized envelope within which orereserves have been defined.

Ore minerals are those metallic minerals,e.g. galena, sphalerite, chalcopyrite, that formthe economic portion of the mineral deposit.

Page 21: Introduction to mineral exploration

4 C.J. MOON & A.M. EVANS

These minerals occur in concentrations thatrange from parts per million (ppm) to low per-centages of the overall mineral deposit.

“Industrial minerals have been defined asany rock, mineral or other naturally occur-ring substance of economic value, exclusive ofmetallic ores, mineral fuels and gemstones”(Noetstaller 1988). They are therefore mineralswhere either the mineral itself, e.g. asbestos,baryte, or the oxide, or some other compoundderived from the mineral, has an industrialapplication (end use). They include rocks suchas granite, sand, gravel, and limestone, thatare used for constructional purposes (theseare often referred to as aggregates or bulkmaterials, or dimension stone if used forornamental cladding), as well as more valuableminerals with specific chemical or physicalproperties like fluorite, phosphate, kaolinite,and perlite. Industrial minerals are alsofrequently and confusingly called nonmetallics(e.g. Harben & Kuzvart 1997), although theycan contain and be the source of metals, e.g.sodium derived from the industrial mineralhalite. On the other hand, many depositscontain metals such as aluminum (bauxite),ilmenite, chromite, and manganese, which arealso important raw materials for industrialmineral end uses.

The JORC definition covers metallicminerals, coal, and industrial minerals. This isthe sense in which the terms mineral resourceand ore reserve will normally be employed inthis book, except that they will be extended toinclude the instances where the whole rock,e.g. granite, limestone, salt, is utilized and notjust a part of it.

The term mineralisation is defined as “anysingle mineral or combination of mineralsoccurring in a mass, or deposit, of economicinterest” (IMM Working Group 2001). Thisgroup used the term mineral reserve for themineable part of the mineralisation (for furtherdiscussion see section 10.4). The AustralianJoint Reserve Committee has a similar ap-proach but prefers the term ore reserve.

Another useful discussion on the economicdefinition of ore is given by Lane (1988). Thistext explains the principles of cut-off gradeoptimization used in mining and processing.This is an essential part of extracting max-imum value out of finite geological resources.

Gangue material is the unwanted material,minerals, or rock, with which ore minerals areusually intergrown. Mines commonly possessprocessing plants in which the run-of-mine(ROM) ore undergoes comminution before theore minerals are separated from the gangueminerals by various processes. This provides asaleable product, e.g. ore concentrates and tail-ings which are made up of the gangue material.

1.2.2 The relative importance of metallic andindustrial minerals

Metals always seem to be the focus of atten-tion for various reasons, such as their use inwarfare, rapid and cyclical changes in price,occasional occurrence in very rich deposits(e.g. gold bonanzas), with the result that thegreat importance of industrial minerals to ourcivilization is overlooked. As flints, stone axes,bricks, and pottery they were the first earthresources to be exploited by humans. Todayindustrial minerals permeate every segment ofour society (McVey 1989). They occur as com-ponents in durable and nondurable consumergoods. The use of industrial minerals is obviousbut often unappreciated, e.g. the construc-tion of buildings, the manufacture of ceramictables, and sanitary ware. The consumer isfrequently unaware that industrial mineralsplay an essential role in numerous other goods,ranging from books to pharmaceuticals. Indeveloped countries such as the UK and USA,but also on a world-wide basis (Tables 1.1 &1.2), industrial mineral production is far moreimportant than metal production from boththe tonnage and financial viewpoints.

Graphs of world production of the tradi-tionally important metals (Figs 1.1–1.3) showinteresting trends. The world’s appetite forthe major metals appeared to be almost insati-able after World War II. Postwar productionincreased rapidly. However, in the mid seven-ties an abrupt slackening in demand occurredtriggered by the coeval oil crisis. Growth formany metals, such as copper, zinc, and ironore, resumed in the 1980s and 1990s. Lead,however, shows a different trend, with overallproduction increasing slightly. Mine pro-duction has declined with more than 50% ofoverall production now derived from recycling.Other factors affecting the change in growth

Page 22: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 5

TABLE 1.1 World production of mineral resourcesin 1998 by quantity; in 000 tonnes (ore is given inmetal equivalent, natural gas in million m3).(From Wellmer & Becker-Platen 2002.)

Commodity 000 tonnes

Diamonds 0.02Platinum group metals 0.35Gold 2.5Electronic metals 3.2Silver 16Cobalt 24Columbium 26Tungsten 32Uranium 35Vanadium 45Antimony 118Molybdenum 135Mica 205Tin 209Magnesium 401Kyanite and related minerals 431Zirconium 456Graphite 648Boron 773Nickel 1100Asbestos 1970Diatomite 2060Titanium 2770Lead 3040Chromium 4180Fluorite 4810Barite 5890Zinc 7470Talc and pyrophyllite 7870Manganese 8790Feldspar 8800Bentonite 9610Copper 12,200Magnesite 20,100Peat 25,300Potash 26,900Aluminum 27,400Kaolin 36,600Phosphate 44,000Sulfur 55,600Gypsum anhydrite 103,700Rock salt 191,100Industrial sand 300,000Clay 500,000Iron 562,000Lignite 848,000Natural gas 2,357,000Crude oil 3,578,000Coal 3,735,000Aggregates 4,100,000Sand and gravel >15,000,000

TABLE 1.2 World minerals production in 1998 byvalue in million euro. (From Wellmer & Becker-Platen 2002.)

Commodity 000,000a

Kyanite and related minerals 64Mica 69Zirconium 130Antimony 146Tungsten 158Graphite 194Electronic metals 200Columbium 237Titanium 345Bentonite 388Talc and pyrophyllite 439Diatomite 440Lead 556Fluorite 576Asbestos 618Peat 658Gypsum anhydrite 689Chromium 727Uranium 862Magnesium 864Vanadium 882Tin 918Molybdenum 926Cobalt 1040Boron 1560Silver 1590Manganese 1610Sulfur 1820Nickel 1890Zinc 2930Rock salt 3030Phosphate 3140Potash 3510Platinum group metals 3760Kaolin 3950Diamonds 3960Industrial sand 4910Barite 5890Feldspar 8800Clay 9500Copper 9800Iron 15,200Magnesite 20,100Gold 20,900Aggregates 23,100Lignite 30,300Aluminium 39,400Sand and gravel 97,000Coal 122,000Natural gas 161,000Crude oil 412,000

Page 23: Introduction to mineral exploration

6 C.J. MOON & A.M. EVANS

1900 1920 1940 1960 1980 20000

400

800

Ore

pro

duct

ion

(Mt)

1200

FIG. 1.1 World production of iron ore, 1900–2000.(Data from Kelly et al. 2001.)

1900 1920 1940 1960 1980 20000

5

10

15

20

25

Al

Mn

Met

al p

rodu

ctio

n (M

t)

FIG. 1.2 World production of aluminum andmanganese metal, 1900–2000. (Data from Kellyet al. 2001.)

rate include more economical use of metalsand substitution by ceramics and plastics, asindustrial minerals are much used as a fillerin plastics. Production of plastics rose by astaggering 1529% between 1960 and 1985, anda significant fraction of the demand behindthis is attributable to metal substitution. The

1900 1920 1940 1960 1980 20000

4

8

12

16

Cu

Zn

Pb

Met

al p

rodu

ctio

n (M

t)FIG. 1.3 World production of copper, lead, and zincmetal, 1900–2000. The lead data are for smelterproduction 1900–54 and for mine production1955–2000. (Data from Kelly et al. 2001.)

increases in production of selected metals andindustrial minerals provide a striking contrastand one that explains why for some years nowsome large metal mining companies have beenmoving into industrial mineral production.In 2002 Anglo American plc derived 8% of itsoperating profits and Rio Tinto plc in 2003derived 11% of its adjusted earnings fromindustrial minerals. An important factor in thelate 1990s and early 2000s was the increasingdemand for metals from China, e.g. aluminumconsumption grew at 14% a year between 1990and 2001 (Humphreys 2003).

1.2.3 Commodity prices – the marketmechanism

Most mineral trading takes place within themarket economy. The prices of minerals ormineral products are governed by supply anddemand. If consumers want more of a mineralproduct than is being supplied at the currentmetal and mineral prices price, this is indicatedby their “bidding up” the price, thus increas-ing the profits of companies supplying thatproduct. As a result, resources in the formof capital investment are attracted into the

Page 24: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 7

industry and supply expands. On the otherhand if consumers do not want a particularproduct its price falls, producers make a loss,and resources leave the industry.

World markets

Modern transport leads to many commoditieshaving a world market. A price change in onepart of the world affects the price in the rest ofthe world. Such commodities include wheat,cotton, rubber, gold, silver, and base metals.These commodities have a wide demand, arecapable of being transported, and the costs oftransport are small compared with the value ofthe commodity. The market for diamonds isworldwide but that for bricks is local.

Over the last few centuries formal organizedmarkets have developed. In these markets buy-ing and selling takes place in a recognizedbuilding, business is governed by agreed rulesand conventions and usually only members areallowed to engage in transactions. Base metalsare traded on the London Metal Exchange(LME) and gold and silver on the LondonBullion Market. Similar markets exist in manyother countries, e.g. the New York CommodityExchange (Comex). Because these marketsare composed of specialist buyers and sellersand are in constant communication with eachother, prices are sensitive to any change inworldwide supply and demand.

The prices of some metals on Comex and theLME are quoted daily by many newspapers andwebsites, whilst more comprehensive guides tocurrent metal and mineral prices can be foundin Industrial Minerals, the Mining Journal, andother technical journals. Short- and long-termcontracts between buyer and seller may bebased on these fluctuating prices. On the otherhand, the parties concerned may agree on acontract price in advance of production, withclauses allowing for price changes because offactors such as inflation or currency exchangerate fluctuations. Contracts of this nature arecommon in the cases of iron, uranium, andindustrial minerals. Whatever the form of saleis to be, the mineral economists of a miningcompany must try to forecast demand for, andhence the price of, a possible mine product wellin advance of mine development, and such con-siderations will usually play a decisive role in

formulating a company’s mineral explorationstrategy. A useful recent discussion of mineralmarkets can be found in Crowson (1998).

Forces determining prices

Demand and supplyDemand may change over a short period oftime for a number of reasons. Where one com-modity substitutes to a significant extent foranother and the price of this latter falls then thesubstituting commodity becomes relativelyexpensive and less of it is bought. Copper andaluminum are affected to a degree in thisway. A change in technology may increase thedemand for a metal, e.g. the use of titanium injet engines, or decrease it in the case of thedevelopment of thinner layers of tin on tinplateand substitution (Table 1.3). The expectation offuture price changes or shortages will inducebuyers to increase their orders to have moreof a commodity in stock.

Supply refers to how much of a commoditywill be offered for sale at a given price over a setperiod of time. This quantity depends on theprice of the commodity and the conditionsof supply. High prices stimulate supply andinvestment by suppliers to increase their out-put. A fall in prices has the opposite effect andsome mines may be closed or put on a care-and-maintenance basis in the hope of bettertimes in the future. Conditions of supply maychange fairly quickly through: (i) changes dueto abnormal circumstances such as natural dis-asters, war, other political events, fire, strikesat the mines of big suppliers; (ii) improvedtechniques in exploitation; (iii) discovery andexploitation of large new orebodies.

Government actionGovernments can act to stabilize or changeprices. Stabilization may be attempted bybuilding up a stockpile, although the merebuilding up of a substantial stockpile increasesdemand and may push up the price! With asubstantial stockpile in being, sales from thestockpile can be used to prevent prices risingsignificantly and purchases for the stockpilemay be used to prevent or moderate price falls.As commodity markets are worldwide it isin most cases impossible for one country act-ing on its own to control prices. Groups of

Page 25: Introduction to mineral exploration

8 C.J. MOON & A.M. EVANS

Aluminum 28 Cobalt 35 Copper 16Diatomite 29.1 Feldspar 81.5 Gold 8.8Gypsum 37.6 Iron ore 12 Lead −5.5Mica 18.9 Molybdenum 8.3 Nickel 16.8Phosphate 42.5 PGM 80.8 Potash 39.1Silver 13.9 Sulfur 19 Talc 44Tantalum 143 Tin −9.8 Trona 44Zinc 26

TABLE 1.3 Percentage increase inworld production of some metalsand industrial minerals 1973–88;metals are in italics. Recycledmetal production is not included.

countries have attempted to exercise controlover tin (ITC) and copper (CIPEC) in this waybut with little success and, at times, signal fail-ure (Crowson 2003).

Stockpiles may also be built up by govern-ments for strategic reasons and this, as men-tioned above, can push up prices markedly.Stockpiling policies of some leading industrial-ized nations are discussed by Morgan (1989).

An action that has increased consumption ofplatinum, palladium, and rhodium has beenthe adoption of regulations on the limitation ofcar exhaust fumes by the EU countries. Theworldwide effort to diminish harmful exhaustemissions resulted in a record industrial pur-chase of 3.2 million ounces of platinum and 3.7million ounces of palladium in 2003. Com-parable actions by governments stimulated byenvironmental lobbies will no doubt occur inthe coming years.

RecyclingRecycling is already having a significant effecton some product prices. Economic and particu-larly environmental considerations will lead toincreased recycling of materials in the immedi-ate future. Recycling will prolong resource lifeand reduce mining wastes and smelter efflu-ents. Partial immunity from price rises, short-ages of primary materials, or actions by cartelswill follow. A direct economic and environ-mental bonus is that energy requirements forrecycled materials are usually much lowerthan for treating ores, e.g. 80% less electricityis needed for recycled aluminum. In the USAthe use of ferrous scrap as a percentage of totaliron consumption rose from 35% to 42% overthe period 1977–87 and aluminum from 26%to 37%; but both copper and aluminum wereapproximately 30% for the western world in1999 (Crowson 2003). Of course the poten-

tial for recycling some materials is muchgreater than for others. Contrary to metals thepotential for recycling industrial minerals ismuch lower. Aggregate recycling is currentlybeing promoted within the European Union.Other commodities such as bromine, fluor-compounds, industrial diamonds, iodine andfeldspar and silica in the form of glass arerecyled but industrial mineral prices will beless affected by this factor (Noetstaller 1988).

Substitution and new technologyThese two factors may both lead to a diminu-tion in demand. We have already seen greatchanges such as the development of longer-lasting car batteries that use less lead, substitu-tion of copper and plastic for lead water pipes,and a change to lead-free petrol; all factors thathave contributed to a downturn in the demandfor lead (Fig. 1.3). Decisions taken by OPEC in1973 affected all metals (Figs 1.1–1.3). They ledto huge increases in the prices of oil and otherfuels, pushed demand towards materials hav-ing a low sensitivity to high energy costs, andfavored the use of lighter and less expensivesubstitutes for metals (Cook 1987).

In the past, base metal producers have spentvast sums of money on exploration, minedevelopment, and production, but have paidtoo little attention to the defence and develop-ment of markets for their products (Davies1987, Anthony 1988). Producers of aluminum,plastics, and ceramics, on the other hand, havepromoted research for new uses including sub-stitution for metals. Examples include tankarmour, now frequently made of multilayercomposites (metal, ceramic, and fibers) andceramic-based engine components, widely usedin automobiles. It has been forecast that by2030 90% of engines used in cars, aeroplanes,and power stations will be made from novel

Page 26: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 9

ceramics. A useful article on developments inceramic technology is by Wheat (1987).

Metal and mineral prices

MetalsMetal prices are erratic and hard to predict(Figs 1.4–1.6). In the short run, prices fluctu-ate in response to unforeseen news affectingsupply and demand, e.g. strikes at large minesor smelters, unexpected increases in ware-house stocks. This makes it difficult to deter-mine regular behavior patterns for somemetals. Over the intermediate term (severaldecades) the prices clearly respond to the riseand fall in world business activity, which issome help in attempts at forecasting pricetrends (Figs 1.4–1.6). The dramatic oil pricerises caused by OPEC in 1973, besides settingoff a severe recession, led to less developedcountries building up huge debts to pay for theincreased costs of energy. This led to a reduc-tion in living standards and the purchasingof fewer durable goods. At the same time manymetal-producing, developing countries suchas Chile, Peru, Zambia, and the DemocraticRepublic of Congo increased production irre-spective of metal prices to earn hard currenciesfor debt repayment. A further aggravation fromthe supply and price point of view has been

FIG. 1.4 Iron ore and manganese prices, 1950–2000,New York in US$ 1998. The prices have thereforebeen deflated or inflated so that they can berealistically compared. (Data from Kelly et al. 2001.)

1960 1980 20001000

2000

3000

4000

5000

6000

Cu

AlPri

ce (

$ t–1

)FIG. 1.5 Copper and aluminum metal prices, 1950–2000, New York in US$ 1998. (Data from Kelly et al.2001.)

1960 1980 2000500

1000

1500

2000

2500

3000

Zn

Pb

Pri

ce (

$ t–1

)

FIG. 1.6 Lead and zinc metal prices, 1950–2000,New York in US$ 1998. (Data from Kelly et al.2001.)

the large number of significant mineral discov-eries since the advent of modern explorationmethods in the fifties, many of which are stillundeveloped (Fig. 1.7). Others, that have beendeveloped since 1980, such as the large dis-seminated copper deposits in Chile, are pro-ducing metal at low cost. Metal explorationists

1960 1980 200020

40

60

80

100

200

400

600

800

1000

Mn

Fe

Fe

ore

pric

e ($

t–1

)

Mn

ore

pric

e ($

t–1

)

Page 27: Introduction to mineral exploration

10 C.J. MOON & A.M. EVANS

per ounce, a figure inconceivable at the begin-ning of the seventies; it then fell back duringthe late 1990s to a price lower in real termsthan that of the 1930s (Fig. 1.8). (Using the USConsumer Price Index the equivalent of $35 in1935 would have been $480 in 2004.) Citizensof many countries were again permitted to holdgold either as bars or coinage and many haveinvested in the metal. Unfortunately for thoseattempting to predict future price changes,demand for this metal is not determined somuch by industrial demand but by fashionand sentiment, both notoriously variable andunpredictable factors. The main destinationsof gold at the present day are carat jewelleryand bars for investment purposes.

The rise in the price of gold after 1971 ledto an increase in prospecting and the discoveryof many large deposits (Fig. 1.7). This trendcontinued until 1993 and gold productionincreased from a low point in 1979 to a peakin 2001 (Fig. 1.8). Many diversified miningcompanies adopted a cautious approach and,like the major gold producers, are not openingnew deposits without being sure that theycould survive on a price of around $US 250 perounce, whilst others are putting more empha-sis in their exploration budgets on base metals.

Industrial mineralsMost industrial minerals can be traded inter-nationally. Exceptions are the low valuecommodities such as sand, gravel, and crushedstone which have a low unit value and aremainly produced for local markets. However,minor deviations from this statement arebeginning to appear, such as crushed granitebeing shipped from Scotland to the USA, sandfrom Western Australia to Japan, and filtration

Num

ber

of d

isco

veri

es

1940 1950 19701960 1980 1990

30

25

20

15

10

5

0

All depositsGiant deposits(> $US 10 billionIGV)

FIG. 1.7 Mineral exploration discovery rate. (FromBlain 2000.)

have, to a considerable extent, become victimsof their own success. It should be noted thatthe fall off in nongold discoveries from 1976onwards is largely due to the difficulty explora-tionists now have in finding a viable deposit inan unfavorable economic climate.

During the 1990s the general outlookwas not promising for many of the traditionalmetals, in particular manganese (Fig. 1.4),lead (Fig. 1.6), tin, and tungsten, although thisseems to have changed at the time of writing(2004). Some of the reasons for this prognost-ication have been discussed above. It is theminor metals such as titanium, tantalum, andothers that seemed likely to have a brighterfuture. For the cyclicity of price trends overa longer period (1880–1980) see Slade (1989)and Crowson (2003). Nevertheless a companymay still decide to make one of the traditionalmetals its exploration target. In this case evalu-ate the potential for readily accessible, highgrade, big tonnage orebodies, preferably in apolitically stable, developed country – highquality deposits of good address as Morrissey(1986) has put it. This is a tall order but wellexemplified by the discovery at Neves-Corvoin southern Portugal of a base metal depositwith 100 Mt grading 1.6% Cu, 1.4% Zn, 0.28%Pb, and 0.10% Sn in a well-explored terrane.The discovery of such deposits is still highlydesirable.

Gold has had a different history since WorldWar II. From 1934 to 1972 the price of goldremained at $US 35 per troy ounce. In 1971 Pre-sident Nixon removed the fixed link betweenthe dollar and gold and left market demand todetermine the daily price. The following dec-ade saw gold soar to a record price of $US 850

Pro

duct

ion

($ o

z−1)

Pro

duct

ion

(ton

nes)

1200

1000

800

600

400

200

0

3000

2500

2000

1500

1000

500

0

Real pricesOutput

1900 1920 1940 1960 1980 2000

FIG. 1.8 Gold price (in US$ in real 1999 terms) andproduction. (Updated from Crowson 2003.)

Page 28: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 11

sand and water from the UK to Saudi Arabia.Lower middle unit value minerals from cementto salt can be moved over intermediate to longdistances provided they are shipped in bulkby low cost transport. Nearly all industrialminerals of higher unit value are internation-ally tradeable, even when shipped in small lots.

The cost to the consumer of minerals witha low unit value will increase greatly withincreasing distance to the place of use. Con-sequently, low unit value commodities arenormally of little or no value unless availableclose to a market. Exceptions to this rule mayarise in special circumstances such as thesouth-eastern sector of England (includingLondon) where demand for aggregates cannotnow be met from local resources. Considerableadditional supplies now have to be broughtin by rail and road over distances in excessof 150 km. For high unit value minerals likeindustrial diamonds, sheet mica, and graphite,location is largely irrelevant.

Like metals, industrial minerals respond tochanges in the intensity of business activit-ies, but as a group, to nothing like the extentshown by metals, and their prices are generallymuch more stable. One reason for the greaterstability of many industrial mineral prices istheir use or partial use in consumer nondur-ables for which consumption remains com-paratively stable during recessions, e.g. potash,phosphates, and sulfur for fertilizer production;diatomite, fluorspar, iodine, kaolin, limestone,salt, sulfur, talc, etc., used in chemicals, paint,paper, and rubber. The value of an industrialmineral depends largely on its end use and theamount of processing it has undergone. Withmore precise specifications of chemical purity,crystalline perfection, physical form, hardness,etc., the price goes up. For this reason manyminerals have quite a price range, e.g. kaolin tobe used as coating clay on paper is four timesthe price of kaolin for pottery manufacture.

Individual commodities show significantprice variations related to supply and demand,e.g. potash over the last 40 years. When sup-plies were plentiful, such as after the construc-tion of several large Canadian mines, priceswere depressed, whereas when demand hasoutstripped supply, prices have shot up.

According to Noetstaller (1988) alreadydiscovered world reserves of most industrial

minerals are adequate to meet the expecteddemand for the foreseeable future, and so nosignificant increases in real long-term pricesare expected. Exceptions to this are likely to besulfur, barite, talc, and pyrophyllite. Growthrates are expected to rise steadily, rates exceed-ing 4% p.a. are forecast for nine industrialminerals and 2–4% for 29 others. These figuresmay well prove to be conservative estimates.Contrary to metals, the recycling potential ofindustrial minerals, with some exceptions, islow, and competing substitutional materialsare frequently less efficient (e.g. calcite forkaolinite as a cheaper paper filler) or moreexpensive.

An industrial mineral, particularly one witha low or middle unit value, should be selectedas an exploration target only if there is an avail-able or potential market for the product. Themarket should not be from the exploration areato keep transport costs at a sustainable level.

1.3 IMPORTANT FACTORS IN THE ECONOMICRECOVERY OF MINERALS

1.3.1 Principal steps in the exploration andexploitation of mineral deposits

The steps in the life cycle of a mineral depositmay be briefly summarized as follows:1 Mineral exploration: to discover a mineraldeposit.2 Feasibility study: to prove its commercialviability.3 Mine development: establishment of theentire infrastructure.4 Mining: extraction of ore from the ground.5 Mineral processing: milling of the ore, sep-aration of ore minerals from gangue material,separation of the ore minerals into concen-trates, e.g. copper concentrate; separation andrefinement of industrial mineral products.6 Smelting: recovering metals from the min-eral concentrates.7 Refining: purifying the metal.8 Marketing: shipping the product (or metalconcentrate if not smelted and refined atthe mine) to the buyer, e.g. custom smelter,manufacturer.9 Closure: before a mine has reached the endof its life, there has to be a closure management

Page 29: Introduction to mineral exploration

12 C.J. MOON & A.M. EVANS

plan in place that details and costs the pro-posed closure strategies. Significant expend-iture could be incurred with clean up andremediation of mining and smelting sites, thecosts of employee retrenchment, and social andcommunity implications.

The exploration step can be subdivided asfollows:(i) Study phase: choice of potential target,study of demand, supply, commodity pricetrends, available markets, exploration cost,draw up budget.(ii) Reconnaissance phase: will start with aliterature search and progress to a review ofavailable remote sensing and photogeologicaldata leading to selection of favorable areas, ini-tial field reconnaissance, and land acquisition,probably followed by airborne surveys, geolo-gical mapping and prospecting, geochemicaland geophysical surveys, and limited drilling(see Chapter 4).(iii) Target testing: detailed geological map-ping and detailed geochemical and geophysicalsurveys, trenching and pitting, drilling (seeChapter 5). If successful this will lead to anorder of magnitude study which will establishwhether there could be a viable project thatwould justify the cost of progressing to a pre-feasibility study.(iv) Pre-feasibility: major sampling and testwork programs, including mineralogical ex-amination of the ore and pilot plant testing toascertain the viability of the selected mineralprocessing option and likely recoverability (seeChapter 11). It evaluates the various optionsand possible combinations of technical andbusiness issues.(v) Feasibility study: drilling, assaying,mineralogical, and pilot plant test work willcontinue. The feasibility study confirms andmaximizes the value of the preferred technicaland business option identified in the pre-feasibility study stage.

It is at the end of the order-of-magnitudestudy that the explorationists usually handover to the mining geologists, mineral proces-sors and geotechnical and mining engineers toimplement steps 1 to 9. Typical time spans andcosts might be: stage (i) 1–2 years, US$0.25M;(ii) 2 years, US$0.5–1.5M; (iii) and (iv) 2–3years, US$2.5–50M; (v) 2 years, US$2.5–50M(excluding actual capital cost for mine con-

struction). Some of these stages will overlap,but this is unlikely to reduce the time involvedand it can be expected that around 12 years willelapse between the start of the exploration pro-gram and the commencement of mine produc-tion. In a number of cases the lead-in time hasbeen less, but this has usually been the result ofthe involvement of favorable factors or a delib-erate search for deposits (particularly of gold)which would have short lead-in times. Furtherinformation on costs of mineral explorationcan be found Tilton et al. (1988) and Crowson(2003).

1.4 STRUCTURE OF THE MINING INDUSTRY

The structure of the mining industry changedgreatly in the 1990s and early 2000s with thedecline in government-funded mineral explora-tion, particularly in centrally planned eco-nomies of central Europe and the former USSR,and the merging and globalization of manymining companies. The producing section ofthe mining industry was dominated, in 2002,by three companies mining a range of com-modities (BHP Billiton, Rio Tinto, and AngloAmerican) and by Alcoa, an aluminum pro-ducer (Fig. 1.9). Other major companies con-centrate on gold mining (Newmont Mining,Barrick Gold, and AngloGold Ashanti),platinum (e.g. Anglo American Platinum andImpala Platinum), and nickel production(Norilsk, Inco). One major copper-producingcompany, Corporacion Nacional del Cobre(Codelco), is not shown as it is still owned bythe Chilean state. Other smaller mining com-panies produce at regional or national levels.

Junior companies are a major feature of themineral exploration industry. They are basedlargely in Canada, where more than 1000 com-panies are active, in Australia, and to a lesserextent in the USA and Europe. Their strategiesare varied but can be divided into two sub-groups: one is exclusively involved in mineralexploration and aims to negotiate agreementswith major companies on any deposits they dis-cover and the other to retain at least a share ofany discovery and to control the production ofany discovery (MacDonald 2002). The depend-ence of these small companies on speculativeactivities has led to some taking extreme risks

Page 30: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 13

US$ billion

AlcoaBHP Billiton

Rio TintoAnglo American

Norsk HydroAlcanCVRD

Anglo American PlatinumBarrick Gold

Newmont MiningWMC

Placer DomeAnglo Gold

PechineyPotash Corp

Homestake MiningNorilsk Nickel

Impala PlatinumFranco Nevada Mining

IncoNoranda

LonminPhelps Dodge

Gold FieldsMitsubishi Materials

Sumitomo Metal MiningFalconbridge

Normandy MiningFreeport-McMoRan Cu and Au

ImerysMitsui Mining and Smelting

BuenaventuraCameco

Antofagasta HoldingsTeckCominco

GencorDowa Mining

Outokumpu OyCoal Allied

Goldcorp

0 5 10 15 20 25 30

FIG. 1.9 Capitalization of major mining companies– a snapshot on 28 September 2001. Since thenHomestake and Barrick Gold have merged as haveNewmont, Franco Nevada, and Normandy. (FromMMSD 2002.)

and giving the industry a bad name (e.g. Bre-X,see section 5.4).

1.5 SOME FACTORS GOVERNING THE CHOICEOF EXPLORATION AREAS

In Chapter 11, consideration is given in somedetail to how mineralisation is converted intoan ore reserve and developed into a mine. Abrief introduction to some of these aspects isgiven here.

1.5.1 Location

Geographical factors may determine whetheror not an orebody is economically viable. In aremote location there may be no electric powersupply or water supply, roads, railways, houses,schools, hospitals, etc. All or some of theseinfrastructural elements will have to be built,the cost of transporting the mine product to itsmarkets may be very high, and wages will haveto be high to attract skilled workers.

1.5.2 Sustainable development

New mines bring prosperity to the areas inwhich they are established but they are boundto have an environmental and social impact.When production started at the Neves-Corvocopper mine in southern Portugal in 1989,it required a total labor force of about 1090.Generally, one mine job creates about threeindirect jobs in the community, in serviceand construction industries, so the impact isclearly considerable. Such impacts have led toconflicts over land use and opposition to theexploitation of mineral deposits by environ-mentalists, particularly in the more populousof the developed countries. The resolution ofsuch conflicts may involve the payment ofcompensation and planning for high closurecosts, or even the abandonment of projects.A Select Committee (1982) stated “. . . whilstpolitical risk has been cited as a barrier toinvestment in some countries, environmentalrisk is as much of a barrier, if not a greater inothers.” Opposition by environmentalists toexploration and mining was partially respons-ible for the abandonment of a major coppermining project in the Snowdonia NationalPark of Wales as early as 1973. Woodall (1992)has remarked that explorationists must notonly prove their projects to be economicallyviable but they must also make them sociallyand therefore politically acceptable. A majorattempt to understand the problem, and to sug-gest solutions, has been made by the Mining,Minerals and Sustainable Development projectsponsored by most major mining companies(MMSD 2002). From this it is clear that the con-cerns of local inhabitants must be addressedfrom an early stage if mine development is tobe successful (discussed further in section 4.3).

Page 31: Introduction to mineral exploration

14 C.J. MOON & A.M. EVANS

These new requirements have now led to atime gap, often of several years in developedcountries, between the moment when a newlyfound deposit is proved to be economicallyviable, and the time when the complex regulat-ory environment has been dealt with and gov-ernmental approval for development obtained.During this time gap, and until productionoccurs, no return is being made on the sub-stantial capital invested during the explorationphase (see section 11.2).

1.5.3 Taxation

Overzealous governments may demand somuch tax that mining companies cannot makea reasonable profit. On the other hand, somegovernments have encouraged mineral devel-opment with taxation incentives such as awaiver on tax during the early years of a miningoperation. This proved to be a great attractionto mining companies in the Irish Republic inthe 1960s and brought considerable economicgains to that country.

Once an orebody is being exploited it hasbecome a wasting asset and one day there willbe no ore, no mine, and no further cash flow.The mine, as a company, will be wound up andits shares will have no value. In other words, allmines have a limited life and for this reasonshould not be taxed in the same manner asother commercial undertakings. When this istaken into account in the taxation structure ofa country, it can be seen to be an importantincentive to investment in mineral explorationand mining in that country.

1.5.4 Political factors

Political risk is a major consideration in theselection of a country in which to explore. Inthe 1970s and 1980s the major fear was nation-alization with perhaps inadequate or even nocompensation. Possible political turmoil, civilstrife, and currency controls may all combineto increase greatly the financial risks of invest-ing in certain countries. In the 1990s and 2000sperhaps more significant risks were long delaysor lack of environmental permits to operate,corruption, and arbitrary changes in taxation.One of the most useful sources of information

on political risk in mining is the Fraser Insti-tute in Vancouver (Fraser Institute 2003). Itpublishes an annual review of the investmentattractiveness of many countries and regionsbased on a poll of mining company executives.The attractiveness is a combination of mineralpotential and policy potential. Some countries,for example Chile, rank at the top of bothindices, whereas others, such as Russia, have avery high mineral potential index but a verylow policy potential index.

1.6 RATIONALE OF MINERAL EXPLORATION

1.6.1 Introduction

Most people in the West are environmental-ists at heart whether engaged in the mineralextraction industry or some other employ-ment. Unfortunately many are of the “nimby”(not in my backyard) variety. These and manyother people fail to realize, or will not face upto the fact, that it is Society that createsthe demand for minerals. The mining andquarrying companies are simply responding toSociety’s desire and demand for houses, wash-ing machines, cars with roads on which todrive them, and so on.

Two stark facts that the majority of ordinarypeople and too few politicians understand arefirst that orebodies are wasting assets (section1.5.3) and second that they are not evenly dis-tributed throughout the Earth’s crust. It is thedepleting nature of their orebodies that playsa large part in leading mining companies intothe field of exploration, although it must bepointed out that exploration per se is not theonly way to extend the life of a mining com-pany. New orebodies, or a share in them, canbe acquired by financial arrangements withthose who own them, or by making successfultakeover bids. Many junior exploration com-panies are set up with the idea of allowingothers to buy into their finds (farming out is thecommercial term). The major mining com-panies also sell orebodies that they consider toosmall for them to operate.

The chances of success in exploration aretiny. Only generalizations can be made but theavailable statistics suggest that a success rating

Page 32: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 15

assessed on the global, national, and companyscale.

For a company, success requires a reasonablefinancial return on its exploration investment.Exploration productivity can be determinedby dividing the expected financial return bythe exploration costs, after these have beenadjusted to take account of inflation. For Soci-ety, on the global or national scale, the cal-culation is much more involved but has beenattempted by a number of workers.

Data on exploration success is rare andscattered. The study of Blain (2000), originallybased on a proprietary database, has attemptedto analyse the mineral exploration successrate. Such an analysis is complicated by manydiscoveries only being recognized some yearsafter initial drilling. The overall appearanceof Fig. 1.7 is however of a peak in explorationsuccess in the late 1960s and a distinct fallin the mid-1990s. Blain considers the discover-ies as a series of waves, offset over time, in dif-ferent commodities, uranium, nickel, copper,poly-metallic base metals, and gold (Fig. 1.10).Most of the discoveries in the 1980s and

of less than a tenth of a percent is the norm andonly in favorable circumstances will this riseabove one percent. With such a high element ofrisk it might be wondered why any risk capitalis forthcoming for mineral exploration. Theanswer is that successful mining can provide amuch higher profitability than can be obtainedfrom most other industrial ventures. Destroythis inducement and investment in mineralexploration will decline and a country’s futuremineral production will suffer. Increases in en-vironmental constraints as well as an impres-sion that the area has been thoroughly exploredled, for example, to a flight of exploration com-panies from British Columbia to Latin Americain the late 1990s.

1.6.2 Exploration productivity

Tilton et al. (1988) drew attention to this im-portant economic measure of mineral explora-tion success and rightly pointed out that itis even more difficult to assess this factorthan it is to determine trends in explorationexpenditures. It is a measure that should be

Num

ber

of d

isco

veri

es Copper price ($ lb

−1)

12

10

8

6

4

2

0

1.4

1.2

1.0

0.6

0.4

0.2

0

Price

1950 1960 1970 1980 1990

Num

ber

of d

isco

veri

es

LME

gold price ($ oz−1)

141210

86420

7006005004003002001000

Price

1950 1960 1970 1980 1990

(a) Copper (b) Gold

Num

ber

of d

isco

veri

es Lead price ($ lb−1)

6

5

4

3

2

1

0

0.6

0.5

0.4

0.3

0.2

0.1

0

Price

1950 1960 1970 1980 1990

Num

ber

of d

isco

veri

es Nickel price ($ lb

−1)

4

3

2

1

0

7 6 5 4 3 2 1 0

Price

1950 1960 1970 1980 1990

(c) Lead-zinc (d) Nickel

FIG. 1.10 Discovery rate by commodity: (a) copper; (b) gold; (c) lead-zinc; (d) nickel. The metal prices (in US$)are uncorrected for inflation, compare them with Figures 1.4, 1.5 and 1.8 in which prices have been corrected.(From Blain 2000.)

Page 33: Introduction to mineral exploration

16 C.J. MOON & A.M. EVANS

1990s have been of gold deposits but the rateof discovery does not appear to have beensustained from more recent data (BHP Billiton2003).

The study by Mackenzie and Woodall (1988)of Australian and Canadian productivity isextremely penetrating and worthy of muchmore discussion than there is space for here. Itshould be emphasized that this analysis, andothers in the literature, is concerned almostexclusively with metallic deposits, and com-parable studies in the industrial mineral sectorstill wait to be made.

Mackenzie and Woodall studied basemetal exploration only and compared theperiod 1955–78 for Australia with 1946–77 forCanada. They drew some striking conclusions.Australian exploration was found to have beenuneconomic but Canadian financially veryfavorable. Although exploration expendituresin Canada were double those in Australia theresulting number of economic discoveries wereeight times greater. Finding an economic de-posit in Australia cost four times as much andtook four times longer to discover and assess asone in Canada. By contrast, the deposits foundin Australia were generally three times largerthat those found in Canada. Comparable ex-ploration expertise was used in the two coun-tries. Australia is either endowed with fewerand larger deposits, or the different surfaceblankets (glacial versus weathered lateritic)render exploration, particularly geophysical,more difficult in Australia, especially when itcomes to searching for small- and medium-sized deposits. The latter is the more probable

reason for the difference in productivity, whichsuggests that improved exploration methodsin the future may lead to the discovery of manymore small- and medium-sized deposits inAustralia.

Mackenzie and Dogget (in Woodall 1992)have shown that the average cost of finding andproving up economic metallic deposits inAustralia over the period 1955–86 was about$A51M or $A34M when discounted at the startof exploration. The average reward discountedin the same manner is $A35M. This is littlebetter than a breakeven situation. Dissectingthese data shows that the average explorationexpenditure per deposit is gold $A17M, nickel$A19M, and base metals $A219M. By contrast,the expected deposit value at the start of explo-ration for gold is $A24M and nickel $A42M.This provided the reasonable rates of return of21% and 19% on the capital invested. On aver-age $53M more has been invested in findingand proving up an economic base metal depositthan has been realized from its subsequentexploitation. Clearly gold and nickel have onaverage been wealth creating whilst base metalexploration has not.

A study of uranium exploration by Crowson(2003) compared exploration expenditure foruranium with resources added over the period1972–80 (Table 1.4). This suggests that, apartfrom the political requirement to have adomestic strategic supply, there was little jus-tification for exploring in France. By contrast,the exploration success in Canada is probablyunderstated, as there were major discoveriesafter 1980 (section 3.3).

TABLE 1.4 Exploration productivity for uranium in the 1970s. (Source: Crowson 2003.)

Total expenditure Resources Resources Spending1972–1980 in April 1970 January 1981 per tonneUS$ million (000 t uranium) (000 t uranium) of increased1982 terms resources

(US$)

Australia 300 34 602 530Canada 625 585 1018 1440France 340 73 121 6960USA 2730 920 1702 3490Brazil 210 2 200 1060India 60 3 57 1150Mexico 30 2 9 4000

Page 34: Introduction to mineral exploration

1: ORE, MINERAL ECONOMICS, AND MINERAL EXPLORATION 17

US

$ m

illio

n

6000

5000

4000

3000

2000

1000

01970 1975 1980 1985 1990 1995 2000

Total PDACTotal MEG estimateTotal MEG surveys

FIG. 1.11 Total global exploration spending in US$2001. Sources: MEG, Metals Economics Group;PDAC, Prospectors and Developers Association ofCanada. (Updated from Crowson 2003.)

1.6.3 Exploration expenditure

Some indication has been given in section 1.3.1of the cost of individual mineral explorationprograms. Here consideration is given to somestatistics on mineral exploration expenditureworldwide (Fig. 1.11). Accurate statistics arehard to obtain, as it is often not clear whetheroverheads are included, and coal and indus-trial mineral exploration is generally excluded(Crowson 2003). However statistics released bythe Metals Economics Group are generallycomprehensive. What is clear is the stronglycyclic nature of spending with distinct peaksin 1980, 1988, and particularly in 1996. IfAustralian and Canadian data are examined(Fig. 1.12) they show similar peaks, but the1996–97 Australian peak was higher than pre-vious peaks and superimposed on a generalincrease from 1970 to 1996, whereas Canadianexpenditure trended downwards from 1980.

In Fig. 1.13 the comparative total expend-itures on exploration for metallic and indus-trial minerals (including coal) in Australia andCanada are shown, as are the sums spent onindustrial minerals alone. Both countries areimportant producers of industrial mineralsdespite their considerably smaller expenditureon exploration for these commodities. Thegreater expenditure on exploration for metalsseems to arise from a number of factors: firstly,it is in general cheaper to prospect for indus-trial minerals and the success rate is higher;secondly, metal exploration, particularly forgold, attracts more high risk investment.

1968 1972 1976 1980 1984 1988 1992 1996 2000

Exp

endi

ture

(U

S$

mill

ion)

1600

1400

1200

1000

800

600

400

200

Canada Old Series

CanadaNew Series

Australia

FIG. 1.12 Australian and Canadian explorationspending in $US 20001. The “New” and “Old”series indicate slightly different methods ofcollecting spending. (Compiled from Crowson2003.)

FIG. 1.13 Comparative total exploration expenditurefor metals and industrial minerals (including coal)in Australia and Canada. The ordinate representscurrent Australian and Canadian dollars, i.e. nocorrections have been made for inflationary effects.Also shown are the graphs for industrial mineralexploration in these countries. Canadian data areshown with solid lines and the Australian data withpecked lines. Australian expenditure is forcompanies only; Canadian expenditure includesgovernment agencies. (Source Crowson 1988.)

1960 1974 1979

Exp

endi

ture

($C

, $A

mill

ions

)

700

600

500

400

300

200

100

0

Canada-Industrial MineralsCanada-MetalsAustralia-Industrial MineralsAustralia-Metals

Page 35: Introduction to mineral exploration

18 C.J. MOON & A.M. EVANS

Figures 1.12 and 1.13 show a general increasein exploration expenditure over recent decades,and Cook (1987) records that expenditure in thenoncommunist world rose from about $400Min 1960 to over $900M in 1980 (constant 1982US$). This increase is partially due to the useof more costly and sophisticated explorationmethods, for example it has been estimatedthat of deposits found in Canada before 1950,85% were found by conventional prospecting.The percentage then dropped as follows: 46%in 1951–55, 26% in 1956–65, 10% in 1966–75,and only 4% during 1971–75. A review bySillitoe (1995, updated in 2000) of discoveriesin the Circum-Pacific area showed that geo-logical work, particularly intimate familiaritywith the type of deposit being sought, was thekey to success. Blain (2000) concurs on the im-portance of geological work and his analysisshows that geology took over from prospectingas the dominant factor in exploration successin approximately 1970.

1.6.4 Economic influences

Nowadays the optimum target (section 1.2.3,“Metal and mineral prices”) in the metals sec-tor must be a high quality deposit of good ad-dress, and if the location is not optimal then anacceptable exploration target will have to beof exceptional quality. This implies that forbase metal exploration in a remote inland partof Australia, Africa, or South America a largemineral deposit must be sought with a hundredor so million tonnes of high grade resources oran even greater tonnage of near surface, lowergrade material. Only high unit value industrialminerals could conceivably be explored for insuch an environment. Anything less valuable

would probably not be mineable under existingtransportation, infrastructure, and productioncosts.

In existing mining districts smaller targetscan be selected, especially close to workingmines belonging to the same company, or acompany might purchase any new finds. Thesebrownfield finds can be particularly import-ant to a company operating a mine that hasonly a few years of reserves left and, in such acase with an expensive milling plant and per-haps a smelter to supply, an expensive satura-tion search may be launched to safeguard thefuture of the operation. Expenditure may thenbe much higher for a relatively small targetthan could be justified for exploration in virginterritory.

1.7 FURTHER READING

An excellent summary of the definitions andusages of the terms mineral resources and orereserves is to be found in the paper by Taylor(1989) and the more formal JORC (AustralasianJORC 2003). The two books by Crowson –Inside Mining (1998) and Astride Mining (2003)– provide stimulating reading on mineral mar-kets and more general aspects of the mineralindustry, including mineral exploration. Break-ing New Ground (MMSD 2002) provides a verydetailed background to the mineral industryand how environmental and social issues mightbe addressed.

Readers may wish to follow developments inmineral exploration using websites such asInfomine (2004) and Reflections (2004), as wellas material also available in printed form in theweekly Mining Journal and Northern Miner.

Page 36: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 19

2The Mineralogy ofEconomic Deposits

Anthony M. Evans

2.1 INTRODUCTION

Ore minerals are the minerals of economicinterest for which the explorationist is search-ing. They can be metallic or nonmetallic. Min-eralogy is used to understand the relationshipsbetween the ore mineral and the uneconomichost rock for their eventual separation.

Economic mineral deposits consist of everygradation from bulk materials or aggregates, inwhich most of the rock or mineral is of com-mercial value, to deposits of precious metals(gold, silver, PGM) from which only a few ppm(or ppb in the case of diamond deposits) areseparated and sold. The valuable mineral in onedeposit may be a gangue mineral in another,e.g. quartz is valuable in silica sands, but isa gangue mineral in auriferous quartz veins.Thus the presentation of lists of ore and gangueminerals without any provisos, as given insome textbooks, can be very misleading to thebeginner. This may lead to an erroneous ap-proach to the examination of mineral deposits,i.e. what is recovered and what is discarded?An alternative question is, how can we processeverything we are going to mine and marketthe products at a profit? There are few mineraloperations where everything is mined gain-fully. Fortunes can be made out of the wasteleft by previous mining and smelting opera-tions, but not usually by the company thatdumped it! A good example of a mine whereeverything is mined is at the King’s MountainOperation in the Tin–Spodumene Belt of NorthCarolina. It is in the world’s most import-ant lithium-producing area (Kunasz 1982).The spodumene occurs in micaceous granite–

pegmatites and in the mill the ore is processedto produce chemical grade spodumene andceramic spodumene concentrates, mica andfeldspar concentrates, and a quartz–feldsparmix marketed as sandspar. The amphibolitehost rock is crushed, sized, and sold as roadaggregate. Of course such comprehensive ex-ploitation of all the material mined is notpossible in isolated locations, but too oftenthe potential of waste material is overlooked.A comprehensive mineralogical examinationof a mineral deposit and its waste rocks maymean that additional valuable materials inthe deposit are identified and the presence ofdeleterious substances detected. This may addvalue if the project is ever brought to the pro-duction stage and will help to avoid embarrass-ing undervaluation.

Ore minerals may be native metals (ele-ments), of which gold and silver are examples,or compounds of metals with sulfur, arsenic,tellurium, etc., such as lead sulfide – the min-eral galena (PbS) – or they may be carbonates,silicates, borates, phosphates. There are fewcommon minerals that do not have an eco-nomic value in some mineralogical context orother. Some of the more important ore min-erals are listed in Table 2.1 and those whichare often classified as gangue minerals in Table2.2. Ore minerals may be classed as primary(hypogene) or secondary (supergene). Hypogeneminerals were deposited during the originalperiod of rock formation or mineralisation.Supergene minerals were formed during a laterperiod of mineralisation, usually associatedwith weathering and other near-surface proc-esses, leading to precipitation of the secondary

Page 37: Introduction to mineral exploration

20 A.M. EVANS

Com

mod

ity

and

prin

cipa

lor

e m

iner

als

Agg

rega

tes

An

tim

ony

stib

nit

ete

trah

edri

teja

mes

onit

eA

rsen

icar

sen

opyr

ite

enar

gite

löll

ingi

tete

nn

anti

teA

sbes

tos

chry

soti

leri

ebec

kit

eB

aryt

eB

auxi

tebo

ehm

ite

dias

pore

gibb

site

Ben

ton

ite

and

full

er’s

ear

thB

eryl

liu

mbe

rtra

ndi

tebe

ryl

Bis

mu

thbi

smu

thin

ite

Bor

ates

Ch

rom

ium

chro

mit

e

Coa

lC

obal

tca

rrol

lite

coba

ltif

erou

spy

rite

Com

mon

cla

y

TA

BLE

2.1

Som

e of

th

e m

ore

impo

rtan

t or

es a

nd

ore

min

eral

s.

Min

eral

form

ulae

Sb2S

3

(Cu

,Fe)

12Sb

4S13

Pb 4

FeSb

6S14

FeA

sSC

u3A

sS4

FeA

s 2(C

u,F

e)12

As 4

S 13

(Mg,

Fe) 3

Si2O

5(O

H) 4

Na 3

(Fe,

Mg)

4FeS

i 8O

22(O

H) 2

BaS

O4

AlO

(OH

)A

lO(O

H)

Al(

OH

) 3M

ontm

oril

lon

ite

grou

pm

iner

als

Be 4

Si2O

7(O

H) 2

Be 3

Al 2

Si6O

18

Bi 2

S 3

(Fe,

Mg)

(Cr,

Al)

2O4

C,O

&H

Cu

Co 2

S 4

(Fe,

Co)

S 2Se

e sh

ale

% M

etal

72 36 33 46 19 73 13 ≈39

81 68(C

r)

56 Var

iabl

e

Prim

ary

X X X X X X X X X X X X X X X X X X X X

Supe

rgen

e

X X X X X

Rem

arks

Man

y di

ffer

ent

rock

s an

dm

iner

als

use

dO

ften

a b

ypro

duct

of b

ase

met

al m

inin

g

Nor

mal

ly a

byp

rodu

ct o

fm

etal

min

ing

Dan

gero

us

mat

eria

ls,

usa

ge d

ecli

nin

g

Indu

stri

al m

iner

al u

ses

incr

easi

ng

Ber

tran

dite

is n

ow m

ore

impo

rtan

t th

an b

eryl

A b

ypro

duct

of v

ario

us

met

al m

inin

g op

erat

ion

sV

ario

us

un

usu

al m

iner

als

Byp

rodu

ct o

f som

e co

pper

min

es

Maj

or u

ses

Civ

il e

ngi

nee

rin

g,bu

ildi

ng

All

oys

Her

bici

de, a

lloy

s, w

ood

pres

erva

tive

Insu

lati

on

Dri

llin

g m

uds

, fill

erA

lum

inu

m p

rodu

ctio

n

Ble

ach

ing

clay

s, d

rill

ing

mu

ds, b

inde

rsA

lloy

s, e

lect

ron

ics

Ph

arm

aceu

tica

l in

dust

ry,

allo

ys

Gla

ss w

ool,

bor

osil

icat

egl

asse

s, d

eter

gen

ts,

agri

cult

ure

Dep

ends

on

Cr

v. A

lco

nte

nt.

Ste

el in

dust

ry,

refr

acto

ries

, ch

emic

als

En

ergy

sou

rce

Hig

h t

empe

ratu

re a

lloy

s,m

agn

ets,

too

l ste

els

Page 38: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 21

Cop

per

born

ite

chal

coci

tech

alco

pyri

teco

vell

ite

Dia

mon

d

Dia

tom

ite

Feld

spar

Flu

orsp

ar

Gol

dn

ativ

e go

ldca

lave

rite

sylv

anit

eG

raph

ite

Gyp

sum

Iron h

emat

ite

mag

net

ite

side

rite

Kao

lin

kao

lin

ite

Lim

esto

ne

Lit

hiu

mam

blyg

onit

ele

pido

lite

peta

lite

spod

um

ene

Lea

d gale

na

Mag

nes

ite

Cu

5FeS

4

Cu

2SC

uFe

S 2C

uS

C SiO

2.n

H2O

NaA

lSi 3

O8

KA

lSi 3

O8

CaF

2

Au

Au

Te 2

(Au

,Ag)

Te 2

C CaS

O4.

2H2O

Fe2O

3

Fe3O

4

FeC

O3

Al 4

Si4O

10(O

H) 8

CaC

O3

Li

K(L

iAl)

3(Si

,Al)

4O10

(F,O

H) 2

LiA

lSi 4

O10

LiA

lSi 2

O6

PbS

MgC

O3

63 80 34 66 90–1

0039 va

riab

le

72 70 48 86

X X X X X X X X X X X X X X X X X X X X X X

Cop

per

prod

uct

ion

Jew

elle

ry, c

utt

ing

and

grin

din

g to

ols

Filt

ers,

fill

er, a

bras

ive

Cer

amic

s, g

lass

mak

ing

Flu

x in

ste

el m

akin

g,fl

uor

och

emic

als

Jew

elle

ry, h

oard

ing,

den

tist

ry

Stee

l mak

ing,

ref

ract

orie

s,fo

un

drie

sP

last

erbo

ard,

insu

lati

onIr

on p

rodu

ctio

n

Pap

er m

anu

fact

ure

,co

atin

g cl

ay, fi

ller

,ex

ten

der

Con

stru

ctio

nal

,ag

ricu

ltu

ral,

ch

emic

al a

nd

met

allu

rgic

al in

dust

ries

Cer

amic

s, g

lass

, en

amel

s,L

i sal

ts, L

i ch

emic

als,

batt

erie

s

Lea

d pr

odu

ctio

n

Ref

ract

orie

s, a

nim

alfe

edst

uff

s, s

peci

alce

men

ts

X X X X X

Man

y ot

her

ore

s of

cop

per

Con

side

rabl

e pr

odu

ctio

nof

syn

thet

ic a

nd

indu

stri

al d

iam

onds

Flu

orsp

ar is

th

e ra

w o

re,

flu

orit

e th

e pu

re m

iner

alM

any

oth

er t

ellu

ride

sm

ay o

ccu

r in

gol

d or

es

A c

onst

itu

ent

of m

any

clay

s: b

all c

lay,

ref

ract

ory

clay

, etc

.O

ne

of t

he

wor

ld’s

mos

tw

idel

y u

sed

mat

eria

ls

Li b

rin

es a

re n

owim

port

ant

prod

uce

rs

Gal

ena

may

be

arge

nti

fero

us

Mar

ket

ed m

ain

ly a

sm

agn

esia

Page 39: Introduction to mineral exploration

22 A.M. EVANS

Man

gan

ese

pyro

lusi

teps

ilom

enan

ebr

aun

ite

man

gan

ite

Mer

cury

cin

nab

arM

ica

mu

scov

ite

phlo

gopi

te

Mol

ybde

nu

mm

olyb

den

ite

wu

lfen

ite

Nep

hel

ine–

syen

ite

and

shee

t

Nic

kel

pen

tlan

dite

garn

ieri

teP

erli

te

PG

M nat

. pla

tin

um

sper

ryli

tebr

aggi

tela

uri

teP

hos

phat

e ro

ckap

atit

e

Pyr

oph

ylli

te

Pot

ash

sylv

ite

carn

alli

tek

ain

ite

lan

gbei

nit

e

TA

BLE

2.1

(con

tin

ued

)

Com

mod

ity

Min

eral

form

ulae

% M

etal

Prim

ary

Supe

rgen

eR

emar

ksM

ajor

use

san

d pr

inci

pal

ore

min

eral

s

Mos

t im

port

ant

ferr

o-al

loy

met

al

Mar

ket

ed a

s sh

eet

or g

rou

nd

mic

a

Min

ed a

s th

e pr

inci

pal

met

al o

r as

a b

ypro

duct

Com

pose

d of

nep

hel

ine,

albi

te a

nd

mic

rocl

ine

A v

olca

nic

gla

ss t

hat

expa

nds

on

hea

tin

gSe

vera

l oth

er P

GM

min

eral

s m

ay b

e pr

esen

tin

ore

s

Stee

l mak

ing.

Ove

r 1

Mt

p.a.

is u

sed

for

indu

stri

alm

iner

al p

urp

oses

Mer

cury

pro

duct

ion

Ele

ctri

cal i

nsu

lato

r,fu

rnac

e w

indo

ws,

wal

lpap

er, p

ain

ts,

plas

terb

oard

Mol

ybde

nu

m p

rodu

ctio

n

Con

tain

er g

lass

,w

hit

ewar

es, g

laze

s

Nic

kel

pro

duct

ion

Insu

lati

on b

oard

, pla

ster

,co

ncr

ete

Cat

alys

ts, e

lect

rica

lin

dust

ry, j

ewel

lery

Fert

iliz

er (9

0%),

dete

rgen

ts, a

nim

alfe

edst

uff

sL

adle

lin

ings

in s

teel

mil

ls, c

eram

ics,

inse

ctic

ides

Abo

ut

95%

goe

s in

tofe

rtil

izer

man

ufa

ctu

re,

rest

for

soap

s, g

lass

,ce

ram

ics,

etc

.

Mn

O2

(BaM

n)M

n4O

8(O

H) 2

Mn

7SiO

12

Mn

O.O

H

HgS

KA

l 2(A

lSi 3

O10

)(OH

) 2K

Mg 3

(AlS

i 3O

10)(F

,OH

) 2

MoS

2

PbM

oO4

(Fe,

Ni)

9S8

(Ni,

Mg)

3Si 2

O5(

OH

) 4R

hyo

liti

c co

mpo

siti

on

Pt

PtA

s 2(P

t,P

d,N

i)S

(Ru

,Ir,

Os)

S 2

Ca 5

(PO

4)3(

F,O

H)

Al 2

Si4O

11.H

2O

KC

lK

Cl.

MgC

l 2.6

H2O

4KC

l.4M

gSO

4.11

H2O

K2S

O4.

2MgS

O4

63 50 64 62 86 60 26 28(m

ax)

<20

≈100

57 vari

able

vari

able

P=

18

X X X X X X X X X X X X X X X X X X X X

X X X X X X X X

Page 40: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 23

RE

E bast

näs

ite

pari

site

mon

azit

eSa

lt hal

ite

Shal

e an

dti

les,

sew

erco

mm

on c

lay

Sili

ca s

and

Sill

iman

ite

min

eral

san

dalu

site

kya

nit

esi

llim

anit

eSi

lver

nat

. sil

ver

acan

thit

ear

gen

tife

rou

ste

trah

edri

tece

rarg

yrit

ean

d m

any

oth

erm

iner

als

Sodi

um

car

bon

ate

tron

a

Stro

nti

um

min

eral

sce

lest

ite

Sulp

hu

rn

at. s

ulp

hu

rpy

rite

Tal

c talc

Tin

, cas

site

rite

stan

nit

e

Man

y m

ore

RE

E in

th

ese

min

eral

s th

an s

how

n in

form

ula

e

60–8

0% o

f all

“cl

ay”

min

ed fa

lls

into

th

isca

tego

rySa

nd

and

grav

el w

ork

ing

tak

en w

orld

wid

ere

pres

ents

on

e of

th

em

ost

impo

rtan

t m

inin

gin

dust

ries

Byp

rodu

ct o

f man

y ba

sem

etal

min

es p

arti

cula

rly

Pb-

Zn

tet

rah

edri

te a

nd

man

y ot

her

min

eral

s

Bu

lk o

f sod

a as

h is

(tro

na)

prod

uce

d sy

nth

etic

ally

inSo

lvay

pla

nts

Cel

esti

te is

th

e on

lyec

onom

ical

ly im

port

ant

Sr m

iner

al.

Use

d in

indu

stry

as

SrC

O3

Mai

n s

ourc

e is

cru

de o

il.

Nat

ive

sulf

ur

30%

, pyr

ite

16%

Has

ver

y m

any

impo

rtan

tu

ses

Cas

site

rite

is t

he

mai

nor

e m

iner

al

(Ce,

La)

(CO

3)F

(Ce,

La)

2Ca(

CO

3)3F

2

(Ce,

La,

Nd,

Th

)PO

4

NaC

l

Cla

y an

d m

ica

min

eral

s,qu

artz

SiO

2

Al 2

SiO

5

Al 2

SiO

5

Al 2

SiO

5

Ag

(Cu

,Fe,

Ag)

12Sb

4S13

AgC

l

Na 2

CO

3.N

aHC

O3.

2H2O

SrSO

4

S FeS 2

Mg 3

Si4O

10(O

H) 2

SnO

2

Cu

2FeS

nS 4

100

Ag 2

S

75 100

78 27

X X X X X X X X 87 X X X X X X X

X X X X

Cat

alys

ts, g

lass

, cer

amic

s,te

levi

sion

tu

bes,

perm

anen

t m

agn

ets,

etc

.

Inn

um

erab

le! O

ver

hal

f of

all p

rodu

ctio

n u

sed

in t

he

chem

ical

indu

stry

, dei

cin

gag

ent,

food

pre

serv

ativ

eB

rick

s, p

ipes

, cem

ent,

ligh

twei

ght

aggr

egat

e

Bu

ildi

ng,

civ

ilen

gin

eeri

ng,

gla

ssm

anu

fact

ure

Ref

ract

orie

s ac

cou

nt

for

90%

of p

rodu

ctio

n

Ph

otog

raph

y, e

lect

rica

lan

d el

ectr

onic

indu

stri

es,

ster

lin

g w

are,

jew

elle

ry,

etc.

Pro

duct

ion

of s

oda

ash

,N

a 2C

O3,

for

glas

sm

anu

fact

ure

, sod

ium

chem

ical

s, p

aper

prod

uct

ion

, etc

.P

yrot

ech

nic

s, c

olor

TV

tu

bes,

per

man

ent

mag

net

s, g

reas

es, s

oaps

Sulf

uri

c aci

dpr

odu

ctio

n,

fert

iliz

ers,

etc

.

Fill

er in

pai

nts

, pla

stic

s,pa

per,

ru

bber

, por

cela

in,

tile

s, c

osm

etic

sSo

lder

, tin

-pla

te, a

lloy

s,et

c.

Page 41: Introduction to mineral exploration

24 A.M. EVANS

Tit

ania

ilm

enit

eru

tile

Tu

ngs

ten

sch

eeli

tew

olfr

amit

eU

ran

ium

ura

nin

ite-

pitc

hbl

ende

ura

nop

han

eca

rnot

ite

torb

ern

ite

Van

adiu

mca

rnot

ite

tyu

yam

un

ite

mon

tros

eite

rosc

oeli

tem

agn

etit

eV

erm

icu

lite

Wol

last

onit

eZ

eoli

tes

Zin

c hem

imor

phit

esm

ith

son

ite

sph

aler

ite

Zir

con

ium

badd

eley

ite

zirc

on

TA

BLE

2.1

(con

tin

ued

)

Com

mod

ity

Min

eral

form

ulae

% M

etal

Prim

ary

Supe

rgen

eR

emar

ksM

ajor

use

san

d pr

inci

pal

ore

min

eral

s

FeT

iO3

TiO

2

CaW

O4

(Fe,

Mn

)WO

4

UO

2

Ca(

UO

2)2S

i 2O

7.6H

2OK

2(U

O2)

2(V

O4)

2.3H

2OC

u(U

O2)

2(P

O4)

2.8H

2O

K2(

UO

2)2(

VO

4)2.

3H2O

Ca(

UO

2)(V

O4)

2.5-

8H2O

VO

(OH

)va

nad

ium

mic

a(F

e,T

i,V

) 2O

4

(Mg,

Fe,A

l)3(

Al,

Si) 4

O10

(OH

) 2.4

H2O

CaS

iO3

Zn

4(O

H) 2

Si2O

7.H

2OZ

nC

O3

(Zn

,Fe)

S

ZrO

2

ZrS

iO4

TiO

2=

45–6

5T

iO2

=90

–98

64 61 88 56 55 51 13 15 61 V=

1.5

–2.1

54 52 up

to 6

7

74 50

X X X X X X X X X X X X X X X X X

X X X X X X X X X

Bot

h a

n im

port

ant

met

alan

d in

dust

rial

mat

eria

l in

form

TiO

2

Man

y ot

her

ore

min

eral

s

Th

e m

ain

pro

duce

ris

now

th

e R

SA fr

omT

i–V

–mag

net

ites

in t

he

Bu

shve

ld C

ompl

ex

Exp

ands

on

hea

tin

g

Use

s m

ult

iply

ing

rapi

dly,

mu

ch o

f th

e de

man

dsu

ppli

ed b

y sy

nth

etic

zeol

ites

Mos

t zi

rcon

is c

onve

rted

to Z

rO2 f

or in

dust

rial

use

Wh

ite

pigm

ent

in p

ain

t,pl

asti

cs, r

ubb

er, p

aper

,et

c.C

utt

ing

mat

eria

ls, s

teel

s,el

ectr

ic li

ght

bulb

s

Nu

clea

r-po

wer

edge

ner

ator

s, s

peci

al s

teel

s,m

ilit

ary

purp

oses

Van

adiu

m s

teel

s, s

upe

ral

loys

Aco

ust

ical

an

d th

erm

alin

sula

tor,

car

rier

for

fert

iliz

er a

nd

agri

cult

ura

lch

emic

als

Cer

amic

s, fi

ller

, ext

ende

rC

emen

ts, p

aper

fill

er, i

onex

chan

ge r

esin

s, d

ieta

rysu

pple

men

t fo

r an

imal

s,m

olec

ula

r “s

ieve

”,ca

taly

st

All

oy c

asti

ngs

,ga

lvan

izin

g ir

on a

nd

stee

l,co

pper

-bas

ed a

lloy

s, e

.g.

bras

sR

efra

ctor

ies,

fou

ndr

ysa

nds

, nu

clea

r re

acto

rs

Page 42: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 25

Aluminum is of course abundant in manysilicate rocks, but it must be usually in the formof hydrated aluminum oxides, the rock calledbauxite, for economic recovery. The mineral-ogy of the ore mineral will also place limits onthe maximum possible grade of the concen-trate. For example, in a mineral deposit con-taining native copper it is theoretically possibleto produce a concentrate containing 100% Cubut, if the ore mineral chalcopyrite (CuFeS2) isthe principal source of copper, then the bestconcentrate would only contain 34.5% Cu.

Undesirable substances

Deleterious elements may be associated withboth ore and gangue minerals. For example,tennantite (Cu12As4S13) in copper ores canintroduce unwanted arsenic and sometimesmercury into copper concentrates. These, likephosphorus in iron concentrates and arsenicin nickel concentrates, will lead to customsmelters imposing financial penalties or refus-ing the shipment. The ways in which gangueminerals may lower the value of an ore arevery varied. For example, an acid leach isnormally employed to extract uranium fromthe crushed ore, but if the carbonate, calcite(CaCO3), is present, there will be excessive acidconsumption and the less effective alkali leach

TABLE 2.2 List of common gangue minerals.

Name Composition Primary Supergene

Quartz SiO2 XChert SiO2 XLimonite Fe2O3.nH2O XCalcite CaCO3 X XDolomite CaMg(CO3)2 X XAnkerite Ca(Mg,Fe)(CO3)2 X XBaryte BaSO4 XGypsum CaSO4.2H2O X XFeldspar All types XFluorite CaF2 XGarnet Andradite most common XChlorite Several varieties XClay minerals Various X XPyrite FeS2 XMarcasite FeS2 XPyrrhotite Fe1–xS XArsenopyrite FeAsS X

minerals from descending solutions. Whensecondary mineralisation is superposed onprimary mineralisation the grade increasesand this is termed supergene enrichment.

2.2 MINERALOGICAL INVESTIGATIONS

Before looking at some of the many methodsthat may be used, the economic importance ofthese investigations will be emphasized by dis-cussing briefly the importance of mineralogicalform and undesirable constituents.

Mineralogical form

The properties of a mineral govern the easewith which existing technology can extractand refine certain metals and this may affectthe cut-off grade (see section 10.4.2). Thusnickel is far more readily recovered fromsulfide than from silicate minerals and sulfideminerals can be extracted down to about 0.5%,whereas silicate minerals must assay about1.5% to be economic.

Tin may occur in a variety of silicate mineralssuch as stanniferous andradite (Ca3Fe2Si3O12)and axinite ((Ca,Fe,Mn)Al2BSi4O15OH), fromwhich it is not recoverable, as well as in itsmain ore mineral form, cassiterite (SnO2).

Page 43: Introduction to mineral exploration

26 A.M. EVANS

method may have to be used. Some primary tindeposits contain appreciable amounts of topazwhich, because of its hardness, increases theabrasion of crushing and grinding equipment,thus raising the operating costs.

To summarize, the information that isrequired from a sample includes some, or all, ofthe following: (i) the grade of the economicminerals; (ii) the bulk chemical composition;(iii) the minerals present; (iv) the proportionsof each of these and their chemical composi-tions; (v) their grain size; (vi) their textures andmineral locking patterns; (vii) any changes inthese features from one part of an orebody toanother.

2.2.1 Sampling

Mineralogical investigations will lose much oftheir value if they are not based on systematicand adequate sampling of all the material thatmight go through the processing plant, i.e. min-eralized material and host rock. The basics ofsound sampling procedures are discussed inChapter 10. The material on which the miner-alogist will have to work can vary from solid,coherent rock through rock fragments andchips with accompanying fines to loose sand.Where there is considerable variation in thesize of particles in the sample it is advant-ageous to screen (sieve) the sample to obtainparticles of roughly the same size, as thesescreened fractions are much easier to samplethan the unsized material.

The mineralogist will normally subsamplethe primary samples obtained by geologistsfrom the prospect to produce a secondarysample, and this in turn may be further reducedin bulk to provide the working sample usingtechniques discussed in Jones (1987) and recenttechnological innovations.

2.2.2 Mineral identification

Initial investigations should be made usingthe naked eye, the hand lens and a stereobino-cular microscope to: (i) determine the ore typespresent and (ii) select representative specimensfor thin and polished section preparation. Atthis stage uncommon minerals may be iden-tified in the hand specimen by using the deter-minative charts in mineralogical textbooks

such as Berry et al. (1983) or the more compre-hensive method in Jones (1987).

The techniques of identifying minerals inthin section are taught to all geologists andin polished sections to most, and will not bedescribed here. For polished section work thereader is referred to Craig and Vaughan (1994)and Ineson (1989), as well as the online manualof Ixer and Duller (1998). Modern optical micro-scopes have significantly increased resolutionand oil immersion is not often used in com-mercial laboratories. Simple microscope andscanning electron microscope (SEM) methodsare usually all that is required to effectivelyidentify all the minerals in the samples. SEMand other methods requiring sophisticatedequipment are discussed below.

X-ray diffraction

X-ray diffraction is used to identify clay min-eral structure and properties, and for mineralanalysis and mineral abundance measurementsthrough spectroscopic sensing. Modern X-raydiffractometers can work well on solid speci-mens, compacted powder pellets representingwhole rocks, or on a few grains on a smearmount. Multiple mounts can be automaticallyfed into the diffractometer.

The rock sample is normally powdered andpacked into an aluminum holder. It is thenplaced in the diffractometer and bombardedwith X-rays. The diffracted rays are collectedby a detector and the information relayed to acomputer where it is converted to d-valuesof specific intensities. This information canthen be shown graphically in the form ofa diffraction pattern or “diffractogram.” Thediffractograms from the unknown sample arethen matched against a database of 70,000recorded phases for mineral identification. Thelatest instruments allow for rapid recognitionof the entire spectrum of the sample in minutesusing a computer to match patterns and iden-tify the minerals present.

Electron and ion probe microanalyzers

With this equipment a beam of high energyelectrons is focused on to about 1–2 µm2 of thesurface of a polished section or a polished thinsection. Some of the electrons are reflected and

Page 44: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 27

provide a photographic image of the surface.Other electrons penetrate to depths of 1–2 µmand excite the atoms of the mineral causingthem to give off characteristic X-radiationwhich can be used to identify the elementspresent and measure their amounts. With thischemical information and knowledge of someoptical properties reasonable inferences can bemade concerning the identity of minute grainsand inclusions. In addition important elementratios such as Fe:Ni in pentlandite ((Fe,Ni)9S8)and Sb:As in tetrahedrite ((Cu,Fe,Ag,Zn)12Sb4S13)–tennantite ((Cu, Ag,Fe,Zn)12As4S13), and smallamounts of possible byproducts and theirmineralogical location can be determined. Use-ful references are Goldstein et al. (1981), Reed(1993), and Zussman (1977).

The ion microprobe uses an ion rather anelectron source and measures the mass ofsecondary ions from the sample rather than X-rays. This technique allows the measurementof much lower concentrations of heavy ele-ments than the electron microprobe and hasbeen widely used in the search for the locationof gold and PGM in metallurgical testing. Asummary of the technique can be found inLarocque and Cabri (1998).

Scanning electron microscopy (SEM)

SEM is of great value in the three-dimensionalexamination of surfaces at magnifications from×20 to 100,000. Textures and porosity can bestudied and, with an analytical facility, indi-vidual grains can be analyzed and identified insitu. Excellent microphotographs can be takenand, with a tilting specimen stage, stereo-graphic pairs can be produced. This equipmentis particularly valuable in the study of lime-stones, sandstones, shales, clays, and placermaterials. The method is discussed in Gold-stein et al. (1981) and Tucker (1988). Recentdevelopments such as the QEM (QuantitativeEvaluation of Minerals)*SEM allows the auto-matic quantification of mineral compositionand size in a similar manner to point countand image analysis for optical microscopes,although the QEM*SEM is rapidly being over-taken by the Mineral Liberation Analyser(MLA) which has particular importance inapplied mineralogy and metallurgical pro-cessing. Mineral Liberation Analyser data are

fundamental parameters used in the designand optimization of processing plants. Gu(2002) explained that the MLA system consistsof a specially developed software package and astandard modern SEM fitted with an energydispersive spectrum (EDS) analyzer. Theon-line program of the MLA software pack-age automatically controls the SEM, capturessample images, performs necessary image ana-lysis (see Chapter 6), and acquires EDS X-rayspectra. Typically, 40–100 images (containing4000–10,000 grains) are acquired for eachsample block and a dozen blocks (of 30 mmdiameter) are measured overnight. The MLAoff-line processing program transforms the rawimage into quality sample images, from whichmost important minerals can be differentiatedusing modern image analysis methods.

Differential thermal analysis

This method is principally used for clay andclay-like minerals which undergo dehydrationand other changes on heating. Measurementof the differences in temperature between anunknown specimen and reference materialduring heating allow for the determination ofthe position and intensity of exothermic andendothermic reactions. Comparison with thebehavior of known materials aids in the iden-tification of extremely fine-grained particlesthat are difficult to identify by other methods(Hutchison 1974).

Autoradiography

Radioactive minerals emit alpha and betaparticles which can be recorded on photo-graphic film or emulsions in contact with theminerals, thus revealing their location in a rockor ore. In ores this technique often shows upthe presence of ultra-fine-grained radioactivematerial whose presence might otherwise gounrecorded. The technique is simple and cheapand suitable for use on hand specimens andthin and polished sections (Robinson 1952,Zussman 1977).

Cathodoluminescence

Luminescence (fluorescence and phosphores-cence) is common in the mineral kingdom.

Page 45: Introduction to mineral exploration

28 A.M. EVANS

In cathodoluminescence the exciting radiationis a beam of electrons and a helpful supple-ment to this technique is ultraviolet fluores-cence microscopy. Both techniques are usedon the microscopic scale to study transparentminerals. Minerals with closely similar opticalproperties or which are very fine-grained can bereadily differentiated by their different lumi-nescent colors, e.g. calcite v. dolomite, feldsparv. quartz, halite v. sylvite. Features not seenin thin sections using white light may appear,thin veins, fractures, authigenic overgrowths,growth zones in grains, etc. A good descriptionof the apparatus required and the method itselfis given in Tucker (1988).

2.2.3 Quantitative analysis

Grain size and shape

The recovery is the percentage of the totalmetal or industrial mineral contained in theore that is recovered in the concentrate; arecovery of 90% means that 90% of the metalin the ore passes into the concentrate and 10%is lost in the tailings. It might be thought that ifone were to grind ores to a sufficiently finegrain size then complete separation of mineralphases might occur to make 100% recoverypossible. In the present state of technology thisis not the case, as most mineral processingtechniques fail in the ultra-fine size range.Small mineral grains and grains finely inter-grown with other minerals are difficult orimpossible to recover in the processing plant,and recovery may be poor. Recoveries fromprimary (bedrock) tin deposits are tradition-ally poor, ranging over 40–80% with an aver-age around 65%, whereas recoveries fromcopper ores usually lie in the range 80–90%.Sometimes fine grain size and/or complexintergrowths may preclude a mining opera-tion. The McArthur River deposit in the North-ern Territory of Australia contains 200 Mtgrading 10% zinc, 4% lead, 0.2% copper, and45 ppm silver with high grade sections run-ning up to 24% zinc and 12% lead. Thisenormous deposit of base metals remainedunworked from its discovery in 1956 until1995 because of the ultra-fine grain size anddespite years of mineral processing researchon the ore.

FIG. 2.1 A random section through a solid consistingof a framework of spheres of equal size.

Grain size measurement methods for loosematerials, e.g. gravels and sands, placer de-posits, or clays, vary, according to grain size,from calipers on the coarsest fragments,through sieving and techniques using settlingvelocities, to those dependent upon changes inelectrical resistance as particles are passedthrough small electrolyte-filled orifices. Thesemethods are described in Tucker (1988) andother books on sedimentary petrography.

Less direct methods have to be employedwith solid specimens because a polished orthin section will only show random profilesthrough the grains (Fig. 2.1). Neither grain size,nor shape, nor sorting can be measured dir-ectly from a polished or thin section andwhen the actual grains are of different sizesmicroscopic measurements using micrometeroculars (Hutchison 1974) or other techniquesinvariably overestimate the proportion ofsmall grains present. This bias can be removedby stereological methods if the grains are ofsimple regular shapes (cubes, spheres, parall-elepipeds, etc.). Otherwise stereological trans-formation is not possible and great care mustbe taken in using grain size measurementstaken from sectioned specimens.

An indication of grain shape can be obtainedby measuring a large number of intercepts(lengths of randomly chosen grain diameters)and analyzing these, e.g. the presence of a sub-stantial number of very small intercepts wouldindicate that the grains were angular and the

Page 46: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 29

opposite would show that the grains lackededges and were smooth and convex (Jones1987).

Modal analysis

Modal analysis produces an accurate represen-tation of the distribution and volume percentof a given mineral in a thin or polished section.Two methods of analysis are normally used,namely:1 Area percentage. The surface area of mineralgrains of the same mineral are measured rela-tive to the total surface area of the thin section,giving the areal proportions of each mineraltype. Since volumes in this situation are dir-ectly proportional to areas, these are also thevolume percentages.2 Point count. Each mineral occurrence alonga series of traverse line across a given thin sec-tion is counted. At least 2000 individual pointsmust be counted for a statistically valid result.

The number of grains counted, the spacingbetween points, and successive traverse linesis dependent on the mean grain size of thesample.

Modal analysis can be used to comparerocks from different areas if there are only thinsections. No chemical analysis is required.The work can be achieved manually usinga petrographic microscope. Modern opticalmicroscopes and SEMs use image analysis (seeChapter 6) to count mineral grains and to cal-culate areal proportions automatically (Jones1987, Sprigg 1987). Statistically representativenumbers of points are achieved routinely usingimage analysis.

However, care must be taken with foliated,banded rocks which should only be sampled atright angles to the banding (Hutchison 1974).Experience shows that porphyritic rocks aredifficult to count. Similarly, care must be takento ensure that the total area of the sample islarger than the maximum diameter of thesmallest grain size. Very coarse-grained rockssuch as pegmatites can be measured with agrid drawn on transparent material and placedon outcrop, joint, or mine surfaces. A similartechnique can be used to visually estimate thegrade of a mineral deposit where the sampler issure that no ore mineral will be missed, e.g.tungsten deposits where the scheelite is the

only ore mineral and can be picked out usingultraviolet light.

The volume percentage of ore and/ordeleterious minerals can be of crucial import-ance in mineral processing and a knowledgeof whether a wanted metal is present in oneor several minerals. If the latter is the case,their relative proportions may also be of greatimportance. Some examples of this are:1 Gold ores – is all the gold present as nativegold (free-milling gold) or is some in the form oftellurides or enclosed by sulfides (refractorygold)? Native gold is readily leached frommilled ores by cyanide solutions, but refractorygold resists leaching and has to be roasted (afterconcentration) before cyaniding or leachedunder pressure, thereby increasing the cost ofthe treatment and of course decreasing thevalue of the ore.2 Titanium ores in anorthosites will havesignificant amounts of titanium locked up intitaniferous magnetite, sphene, and augite fromwhich it is not recoverable.3 The skarn iron orebody at Marmoraton,Ontario assayed on average 50% Fe, but only37.5% (in magnetite) was recoverable, the restwas locked in silicates. These and similardevaluing features are readily detected andquantified by microscopic investigations.Valuations based on assays alone may begrossly exaggerated.

If the chemical compositions of the mineralsare known, dividing volume percentages bymineral densities (and converting to percent)provides the weight percentages of the min-erals, and by using Table 2.1 (or by calcula-tion) gives us an entirely independent way ofobtaining an estimate of the grade. This maybe of value as a check on chemical or X-rayfluorescence assays.

The mineral explorationist should trainhim or herself to make visual modal estimatesin the field using hand specimens and naturalexposures. By estimating the volume percent-age of a metallic mineral such as chalcopyrite(often the only or principal copper mineral in amineral deposit) and looking up the copper con-tent in Table 2.1, a visual assay can be made.When the first laboratory assays become avail-able for a prospect under investigation explora-tionists will have reference material withwhich they can compare their estimates and

Page 47: Introduction to mineral exploration

30 A.M. EVANS

improve their accuracy for that particular min-eral deposit type. The American GeologicalInstitute Data Sheets (sheet 15.1) and variousbooks (Spock 1953, Thorpe & Brown 1985,Tucker 1988, Barnes & Lisle 2003 and others)have comparison charts to help the field geo-logist in estimating percentage compositionsin hand specimens.

2.2.4 Economic significance of textures

Mineral interlocking

Ores are crushed during milling to liberatethe various minerals from each other (section2.2.3) and for concentration a valuable mineralhas to be reduced to less than its liberation sizein order to separate it from its surroundinggangue. Crushing and grinding of rock is expen-sive and if the grain size of a mineral is belowabout 0.05 mm the cost may well be higherthan the value of the liberated constituents.In addition there are lower limits to the degreeof milling possible dictated by the separationprocesses to be employed because these aremost effective over certain grain size ranges:e.g. magnetic separation, 0.02–2.5 mm; frothflotation, 0.01–0.3 mm; electrostatic separa-tion, 0.12–1.4 mm.

In Fig. 2.2 a number of intergrowth patternsare illustrated. Further crushing of the granulartextured grains in (a) will give good separationof ore (black) from gangue – this is an ideal tex-ture from the processing point of view. In (b)further crushing of the tiny pyrite grain veinedby chalcopyrite (black) is out of the questionand this copper will be lost to the tailings.The chalcopyrite (black) occurring as spheroidsin sphalerite grains (c) is too small to be liber-ated and will go as a copper loss into the zincconcentrate. The grain of pyrite coated withsupergene chalcocite (black) in (d) will, duringfroth flotation, carry the pyrite as a dilutingimpurity into the copper concentrate. Thegrain is too small for separation of the twominerals by crushing. It must be noted thatthe market price for a metal does not applyfully or directly to concentrates. The purchaseterms quoted by a custom smelter are usuallybased on a nominal concentrate grade andlower concentrate grades are penalized accord-

(a) (b)

(c)(d)

(e) (f)

ing to the amount by which they fall belowthe contracted grade. Exsolution textures com-monly devalue ores by locking up ore mineralsand by introducing impurities. In (e) the tinyflame-shaped exsolution bodies of pentlandite(black) in the pyrrhotite grain will go with thepyrrhotite into the tailings and the ilmenitebodies (black) in magnetite (f) are likewise toosmall to be liberated by further grinding andwill contaminate the magnetite concentrate. Ifthis magnetite is from an ilmenite orebodythen these interlocked ilmenite bodies will be atitanium loss.

FIG. 2.2 (a)–(d) Grains from a mineral dressing plant.(a) Granular texture; black represents an oremineral, unornamented represents gangue (×0.6). (b)A pyrite grain veined by chalcopyrite (black) (×177).(c) A sphalerite grain containing small roundedinclusions of chalcopyrite (black) (×133).(d) Pyrite grain coated with supergene chalcocite(black) (×233). (e) Grains of pyrrhotite with exsolvedgranular pentlandite in the interstices and flameexsolution bodies within the pyrrhotite (×57).(f) Exsolution blades of ilmenite in a magnetite grain (×163).

Page 48: Introduction to mineral exploration

2: MINERALOGY OF ECONOMIC DEPOSITS 31

erals mention can be made of the presenceof limonite coatings on quartz grains in sandrequired for glass making. Extra cost in ex-ploitation will arise from the need for anacid leach, or other method, to remove thecoatings. Coal fragments in gravel will renderit valueless as gravel processing plants do notinclude equipment for eliminating the coal.Lastly it has become apparent recently thatthe alkaline-silica reaction, which produces“concrete cancer,” is due in some cases to thepresence of certain types of opaline silica inthe aggregate.

2.2.6 Miscellaneous examples of the use ofmicroscopy in ore evaluation andmineral processing

Chromite ores

Chromite is never pure FeCr2O4 and iron(III)may substitute for chromium, particularlyalong grain boundaries and fractures, producing“off color grains” which can be detected inpolished sections by the experienced observer.The magnetite rims will introduce mineralprocessing difficulties, as will badly fracturedchromite grains present in some podiformdeposits; these may disintegrate rapidly into avery fine-grained powder on grinding.

Nickel sulfide ores

The normal opaque mineralogy is magnetite–pyrrhotite–pentlandite–chalcopyrite. Themagnetite and pyrrhotite are separated mag-netically and normally become waste. Nickelpresent in flame-like exsolution bodies (Fig.2.2e) and in solid solution in the pyrrhotite willbe lost. In an ore investigated by Stephens(1972) use of the electron probe microanalyzerrevealed that 20% of the nickel would be lostin this way. Minute grains of PGM mineralsare often included in the pyrrhotite, again lead-ing to losses if this has not been noted. Also insuch ores the chalcopyrite may contain largeexsolved bodies of cubanite, a magnetic min-eral, which could mean a substantial copperloss in the pyrrhotite concentrate. In such acase the ore must be roasted (at extra expense)to destroy the magnetism of the cubanite.

Sieve size % Sn

–y1 fraction+ y2

F

S

60

40

–y2 fraction+ y3

F

S

40

60

–y3 fraction+ y4

F

S

15

85

–y4 fraction+ y5

F

S

9

91

–y5 fraction+ y6

F

S

8

92

–y6 fraction+ y7

F

S

7.5

92.5

FIG. 2.3 Mineral liberation size investigation of apossible tin ore. F, float; S, sink; y1 to y7, grain (sieve)sizes. For discussion see text.

Mineral liberation size

This is usually investigated as follows. Supposewe have a simple example such as a tin oreconsisting of cassiterite (ρ = 6.99) and quartz(ρ < of the minerals are separated, using heavyliquids such as (ρ = 3.2) sodium-polytunstateand lithium-tungstate. The float and sink frac-tions are analyzed for tin grade. Figure 2.3shows a possible result, where little improve-ment in separation would be obtained by grind-ing beyond the y5–y6 fraction.

2.2.5 Deleterious substances

These have been discussed in under “Undesir-able substances” in section 2.2 with referenceto arsenic, mercury, phosphorus, calcite, andtopaz in metallic ores. Among industrial min-

Page 49: Introduction to mineral exploration

32 A.M. EVANS

Tin ores

At mines where separate concentrates ofcassiterite and copper–zinc–arsenic sulfides(often as a minor byproduct) are produced,cassiterite coated with stannite (Cu2FeSnS4)will pass into the sulfide concentrate.

Zinc loss

Lead–zinc ore from a new orebody was testedby being processed in a mill at a nearby mine.Despite being apparently identical to the ore atthe mine, substantial zinc losses into the tail-ings occurred. It was discovered with the useof an electron probe microanalyzer that thesiderite in the gangue carried 8–21% Zn insolid solution.

Mercury impurity

In 1976 Noranda Mines Ltd cut its copper–gold–silver concentrate purchases from Con-solidated Rambler Mines Ltd, Newfoundlandby nearly 50% because of “relatively highimpurities” (Anon 1977). The major impuritywas mercury. An electron probe investigationshowed that this occurred in solid solution(1–2%) in the minor sphalerite in the ore. Bydepressing the zinc in the flotation circuit themercury content of the concentrate was virtu-ally eliminated.

Probably the first mine to recover mercuryfrom copper concentrate was Rudnany inCzechoslovakia where it occurs in tetrahedrite.At the former Gortdrum Mine in Ireland thepresence of cinnabar in the copper–silver ore,

and of mercury in solid solution in the ten-nantite, remained unknown for several yearsand the smelting of concentrates from thismine at a custom smelter in Belgium presum-ably produced a marked mercury anomaly overa substantial part of western Europe. A mer-cury separation plant was then installed atGortdrum and a record production of 1334flasks was reached in 1973.

Sulfur in coal

Of the three sulfur types in coal (organic,sulfate, and sulfide) the sulfide is usually pres-ent as pyrite and/or marcasite. If it is coarse-grained then much can be removed duringwashing, but many coals, e.g. British ones, havesuch fine-grained pyrite that little can be doneto reduce the sulfur content. Such coals are nolonger easily marketable in this time of con-cern about acid rain.

2.3 FURTHER READING

General techniques in applied mineralogy arewell discussed in Jones’ Applied Mineralogy –A Quantitative Approach (1987). Hutchison’sLaboratory Handbook of Petrographic Tech-niques (1974) is also an invaluable book, as isZussman’s (1977) Physical Methods in Deter-minative Mineralogy. Those working on pol-ished sections should turn to Ore Microscopyand Ore Petrography by Craig and Vaughan(1994) and the online atlas of Ixer and Duller(1998). References to techniques not covered bythese books are given in the text of this chapter.

Page 50: Introduction to mineral exploration

3Mineral Deposit

Geology and ModelsAnthony M. Evans and Charles J. Moon

A detailed understanding of the geology ofmineral deposits is required to explore effec-tively for them. As this is beyond the scopeof this volume, only some key features arediscussed here as well as how a more advancedappreciation of mineral deposit geology maybe used in exploration programs. For furtherdetailed consideration of mineral deposit geo-logy the reader is referred to the companionvolume in this series by Robb (2004) and refer-ences listed in section 3.4.

3.1 NATURE AND MORPHOLOGY OFOREBODIES

3.1.1 Size and shape of ore deposits

The size, shape, and nature of ore depositsaffects the workable grade. Large, low gradedeposits which occur at the surface can beworked by cheap open pit methods, whilst thintabular vein deposits will necessitate moreexpensive underground methods of extraction.Open pitting, aided by the savings from bulkhandling of large daily tonnages (say >30 kt),has led to a trend towards the large scale miningof low grade orebodies. As far as shape is con-cerned, orebodies of regular shape can generallybe mined more cheaply than those of irregularshape, particularly when they include barrenzones. For an open pit mine the shape andattitude of an orebody will also determine howmuch waste has to be removed during mining.The waste will often include not only over-burden (waste rock above the orebody) but alsowaste rock around and in the orebody, which

has to be cut back to maintain a safe overallslope to the sides of the pit (see section 11.2.1).Before discussing the nature of ore bodies wemust learn some of the terms used in describ-ing them.

If an orebody viewed in plan is longer in onedirection than the other we can designate thislong dimension as its strike (Fig. 3.1). Theinclination of the orebody perpendicular to thestrike will be its dip and the longest dimensionof the orebody its axis. The plunge of the axisis measured in the vertical plane ABC but itspitch or rake can be measured in any otherplane, the usual choice being the plane contain-ing the strike, although if the orebody is faultcontrolled then the pitch may be measured inthe fault plane. The meanings of other termsare self-evident from the figure.

It is possible to classify orebodies in thesame way as we divide up igneous intrusionsaccording to whether they are discordant orconcordant with the lithological banding (oftenbedding) in the enclosing rocks. Consideringdiscordant orebodies first, this large class can besubdivided into those orebodies which have anapproximately regular shape and those whichare thoroughly irregular in their outlines.

Discordant orebodies

Regularly shaped bodies

Tabular orebodies. These bodies are extensivein two dimensions, but have a restricted devel-opment in their third dimension. In this classwe have veins (sometimes called fissure-veins)

Page 51: Introduction to mineral exploration

34 A.M. EVANS & C.J. MOON

FIG. 3.1 Illustrations of terms used in the descriptionof orebodies.

Flat Thick impervious shale

Limestone

Shale

Limestone

Shale

Sandstone

Hangingwall

Footwall

20 m

FIG. 3.2 Vein occupying a normal fault andexhibiting pinch-and-swell structure, giving rise toribbon ore shoots. The development of a flat beneathimpervious cover is shown also.

The infilling of veins may consist of onemineral but more usually it consists of anintergrowth of ore and gangue minerals. Theboundaries of vein orebodies may be the veinwalls or they may be assay boundaries withinthe veins.

Tubular orebodies. These bodies are relat-ively short in two dimensions but extensive inthe third. When vertical or subvertical theyare called pipes or chimneys, when horizontalor subhorizontal, “mantos.” The Spanish wordmanto is inappropriate in this context for itsliteral translation is blanket; it is, however,firmly entrenched in the English geologicalliterature. The word has been and is employedby some workers for flat-lying tabular bodies,but the perfectly acceptable word “flat” (Fig.3.2) is available for these; therefore the readermust look carefully at the context when he orshe encounters the term “manto.” Mantos andpipes may branch and anastomose and pipesfrequently act as feeders to mantos.

In eastern Australia, along a 2400 km beltfrom Queensland to New South Wales, thereare hundreds of pipes in and close to graniteintrusions. Most have quartz fillings and someare mineralized with bismuth, molybdenum,tungsten, and tin; an example is shown inFig. 3.3. Pipes may be of various types andorigins (Mitcham 1974). Infillings of min-eralized breccia are particularly common, agood example being the copper-bearing breccia

and lodes (Fig. 3.2). These are essentially thesame and only the term vein is now normallyused. Veins are often inclined, and in suchcases, as with faults, we can speak of the hang-ing wall and the footwall. Veins frequentlypinch and swell out as they are followed up ordown a stratigraphical sequence (Fig. 3.2). Thispinch-and-swell structure can create difficult-ies during both exploration and mining oftenbecause only the swells are workable. If theseare imagined in a section at right angles to thatin Fig. 3.2, it can be seen that they form ribbonore shoots. Veins are usually developed in frac-ture systems and therefore show regularities intheir orientation throughout the orefield inwhich they occur.

Strike of orebodyD

A

E

C

B

Pitch orrakeDip

Plunge

Axis of o

rebody

AB and CB lie in the samevertical plane.DB, AB and EB are in thesame horizontal plane andEB is perpendicular to DB

Longitudinal section of an orebody Cross sectionof the sameorebody

Surface

Width orthicknessLevels

Surface

Shaft

Breadth

Plunge lengthStope or level length

Page 52: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 35

FIG. 3.3 Diagram of the Vulcan pipe, Herberton,Queensland. The average grade was 4.5% tin.(After Mason 1953.)

FIG. 3.4 Stockwork of molybdenite-bearing quartzveinlets in granite that has undergone phyllicalteration. Run of the mill ore, Climax, Colorado.

pipes of Messina in South Africa (Jacobsen& McCarthy 1976).

Irregularly shaped bodies

Disseminated deposits. In these deposits, oreminerals are peppered throughout the body ofthe host rock in the same way as accessoryminerals are disseminated through an igneousrock; in fact, they often are accessory minerals.A good example is that of diamonds in kim-berlites. In other deposits, the disseminationsmay be wholly or mainly along close-spacedveinlets cutting the host rock and formingan interlacing network called a stockwork(Fig. 3.4), or the economic minerals may be dis-seminated through the host rock along veinlets.Whatever the mode of occurrence, mineralisa-tion of this type generally fades gradually out-wards into subeconomic mineralisation andthe boundaries of the orebody are assay limits.They are, therefore, often irregular in formand may cut across geological boundaries. The

overall shapes of some are cylindrical, othersare caplike, whilst the mercury-bearing stock-works of Dubnik in Slovakia are sometimespear-shaped.

Stockworks most commonly occur in por-phyritic acid to intermediate plutonic igneousintrusions, but they may cut across the contactinto the country rocks, and a few are whollyor mainly in the country rocks. Disseminateddeposits produce most of the world’s copperand molybdenum (porphyry coppers and dis-seminated molybdenums) and they are also ofsome importance in the production of tin, gold,silver (see Chapter 16), mercury, and uranium.Porphyry coppers form some of the world’smonster orebodies. Grades are generally 0.4–1.5% Cu and tonnages 50–5000 Mt.

Irregular replacement deposits. Many oredeposits have been formed by the replacementof pre-existing rocks, particularly carbonate-rich sediments, e.g. magnesite deposits. Thesereplacement processes often occurred at hightemperatures, at contacts with medium-sizedto large igneous intrusions. Such deposits havetherefore been called contact metamorphicor pyrometasomatic; however, skarn is now

Page 53: Introduction to mineral exploration

36 A.M. EVANS & C.J. MOON

tion of the orebodies. The principal materialsproduced from skarn deposits are iron, copper,tungsten, graphite, zinc, lead, molybdenum,tin, uranium, and talc.

Concordant orebodies

Sedimentary host rocksConcordant orebodies in sediments are veryimportant producers of many different metals,being particularly important for base metalsand iron, and are of course concordant withthe bedding. They may be an integral part of thestratigraphical sequence, as is the case withPhanerozoic ironstones – or they may beepigenetic infillings of pore spaces or replace-ment orebodies. Usually these orebodies showa considerable development in two dimen-sions, i.e. parallel to the bedding and a limiteddevelopment perpendicular to it (Fig. 3.6), andfor this reason such deposits are referred to asstratiform. This term must not be confusedwith strata-bound, which refers to any type ortypes of orebody, concordant or discordant,which are restricted to a particular part of the

Replacement bodyof Fe ore 100 m

Fault

lgneous intrusionLimestone

Shale

Sandstone

FIG. 3.5 Skarn deposit at Iron Springs, Utah. (AfterGilluly et al. 1959.)

FIG. 3.6 Cross-section through the ore zone, Sullivan Mine, British Columbia. (After Sangster & Scott 1976.)

Sulfide ore

Quartzite beds

Footwallconglomerate

Granophyre

T

Pyrrhotite

Tourmalinization

Footwall breccia

Albitization

SURFACE

UPPERQUARTZITES

Chloritization

Diorites

T

T

TT

T

0 150 m

the preferred and more popular term. Theorebodies are characterized by the develop-ment of calc-silicate minerals such as diopside,wollastonite, andradite garnet, and actinolite.These deposits are extremely irregular in shape(Fig. 3.5); tongues of ore may project along anyavailable planar structure – bedding, joints,faults, etc., and the distribution within thecontact aureole is often apparently capricious.Structural changes may cause abrupt termina-

Page 54: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 37

stratigraphical column. Thus the veins, pipes,and flats of the Southern Pennine orefield ofEngland can be designated as strata-bound, asthey are virtually restricted to the Carbonifer-ous limestone of that region. A number ofexamples of concordant deposits which occurin different types of sedimentary rocks will beconsidered.

Limestone hosts. Limestones are very com-mon host rocks for base metal sulfide deposits.In a dominantly carbonate sequence ore is of-ten developed in a small number of preferredbeds or at certain sedimentary interfaces.These are often zones in which the permeabil-ity has been increased by dolomitization orfracturing. When they form only a minor partof the stratigraphical succession, limestones,because of their solubility and reactivity, canbecome favorable horizons for mineralisation.For example the lead–zinc ores of Bingham,Utah, occur in limestones which make up 10%of a 2300 m succession mainly composed ofquartzites.

Argillaceous hosts. Shales, mudstones,argillites, and slates are important host rocksfor concordant orebodies which are oftenremarkably continuous and extensive. In Ger-many, the Kupferschiefer of the Upper Permianis a prime example. This is a copper-bearingshale a meter or so thick which, at Mansfeld,occurred in orebodies which had plan dimen-sions of 8, 16, 36 and 130 km2. Mineralisationoccurs at exactly the same horizon in Poland,where it is being worked extensively, andacross the North Sea in north-eastern England,where it is subeconomic.

The world’s largest, single lead–zinc orebodyoccurs at Sullivan, British Columbia. The hostrocks are late Precambrian argillites. Abovethe main orebody (Fig. 3.6) there are a numberof other mineralized horizons with concordantmineralisation. This deposit appears to besyngenetic and the lead, zinc and other metalsulfides form an integral part of the rocks inwhich they occur. The orebody occurs in asingle, generally conformable zone 60–90 mthick and runs 6.6% Pb and 5.9% Zn. Othermetals recovered are silver, tin, cadmium, anti-mony, bismuth, copper, and gold. This orebodyoriginally contained at least 155 Mt of ore.

Other good examples of concordant depositsin argillaceous rocks, or slightly metamor-phosed equivalents, are the lead–zinc depositsof Mount Isa, Queensland, many of theZambian Copperbelt deposits, and the coppershales of the White Pine Mine, Michigan.

Arenaceous hosts. Not all the ZambianCopperbelt deposits occur in shales andmetashales. Some bodies occur in alteredfeldspathic sandstones such as Mufulira, whichconsists of three extensive lenticular orebodiesstacked one above the other and where theore reserves in 1974 stood at 282 Mt assaying3.47% Cu. The largest orebody has a strikelength of 5.8 km and extends severalkilometers down dip. Many other concordantsandstone-hosted orebodies occur around theworld, such as those in desert sands (red bedcoppers), which are very important in Chinawhere they make up nearly 21% of thestratiform copper reserves of that country(Chen 1988).

Many mechanical accumulations of highdensity minerals such as magnetite, ilmenite,rutile, and zircon occur in arenaceous hosts,usually taking the form of layers rich in heavyminerals in Pleistocene and Holocene sands.As the sands are usually unlithified, the de-posits are easily worked and no costly crushingof the ore is required. These orebodies belongto the group called placer deposits. Beach sandplacers supply much of the world’s titanium,zirconium, thorium, cerium, and yttrium. Theyoccur along present-day beaches or ancientbeaches where longshore drift is well devel-oped and frequent storms occur. Economicgrades can be very low and sands running aslittle as 0.6% heavy minerals are worked alongAustralia’s eastern coast.

Rudaceous hosts. Alluvial gravels and con-glomerates also form important recent andancient placer deposits. Alluvial gold depositsare often marked by “white runs” of vein quartzpebbles as in the White Channels of the Yukon,the White Bars of California, and the WhiteLeads of Australia. Such deposits form one ofthe few types of economic placer depositsin fully lithified rocks, and indeed the majorityof the world’s gold is won from Precambriandeposits of this type in South Africa (see

Page 55: Introduction to mineral exploration

38 A.M. EVANS & C.J. MOON

Chapter 14). Figure 3.7 shows the distributionof the gold orebodies in the East Rand Basinwhere the vein quartz pebble conglomeratesoccur in quartzites of the WitwatersrandSupergroup. Their fan-shaped distributionstrongly suggests that they occupy distributarychannels. Uranium is recovered as a byproductof the working of the Witwatersrand goldfields.In the very similar Blind River area of Ontariouranium is the only metal produced.

Chemical sediments. Sedimentary iron andmanganese formations and evaporites occurscattered through the stratigraphical columnwhere they form very extensive beds conform-able with the stratigraphy.

Igneous host rocks

Volcanic hosts. The most important deposittype in volcanic rocks is the volcanic-associated massive sulfide (see Chapter 15) oroxide type. The sulfide variety often consists ofover 90% iron sulfide usually as pyrite. Theyare generally stratiform bodies, lenticular tosheetlike (Fig. 3.8), developed at the interfacesbetween volcanic units or at volcanic–sedimentary interfaces. With increasingmagnetite content, these sulfide ores gradeinto massive oxide ores of magnetite and/orhematite such as Savage River in Tasmania,Fosdalen in Norway, and Kiruna in Sweden

0 5 km

Pay streaks

FIG. 3.7 Distribution of pay streaks (gold orebodies)in the Main Leader Reef in the East Rand Basin ofthe Witwatersrand Goldfield of South Africa. Thearrows indicate the direction of dip at the outcrop orsuboutcrop. For the location of this ore district seeFig. 14.2. (After Du Toit 1954.)

FIG. 3.8 Schematic section through an idealized volcanic-associated massive sulfide deposit showing theunderlying feeder stockwork and typical mineralogy. Py, pyrite; sp, sphalerite; ga, galena; cp, chalcopyrite.

Stockwork

AndesiteBedded chert, oftenferruginous

Rhyolite

Sulfidic tuff

Massive sulfidesPy - sp - ga - cp (+ Ag, Au)

Py-cp low insp, ga, Ag, Au

Page 56: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 39

(Solomon 1976). They can be divided into threeclasses of deposit: (a) zinc–lead–copper, (b)zinc–copper, and (c) copper. Typical tonnagesand copper grades are 0.5–60 Mt and 1–5%,but these are commonly polymetallic depositsoften carrying other base metals and significantprecious metal values which make them plumtargets for exploration, e.g. Neves-Corvo (seesection 1.2.3, “Metal and mineral prices”).

The most important host rock is rhyolite andlead-bearing ores are only associated with thisrock type. The copper class is usually, but notinvariably, associated with mafic volcanics.Massive sulfide deposits commonly occur ingroups and in any one area they are found atone or a restricted number of horizons withinthe succession (see section 15.2.5). Thesehorizons may represent changes in com-position of the volcanic rocks, a change fromvolcanism to sedimentation, or simply a pausein volcanism. There is a close association withvolcaniclastic rocks and many orebodies over-lie the explosive products of rhyolite domes.These ore deposits are usually underlain by astockwork that may itself be ore grade andwhich appears to have been the feeder channelup which mineralizing fluids penetrated toform the overlying massive sulfide deposit. Allthese relationships are of great importance inthe search for this orebody type.

Plutonic hosts. Many plutonic igneous intru-sions possess rhythmic layering and this isparticularly well developed in some basic in-trusions. Usually the layering takes the form ofalternating bands of mafic and felsic minerals,but sometimes minerals of economic interestsuch as chromite, magnetite, and ilmenite mayform discrete mineable seams within suchlayered complexes. These seams are naturallystratiform and may extend over many kilo-meters, as is the case with the chromite seamsin the Bushveld Complex of South Africa andthe Great Dyke of Zimbabwe.

Another form of orthomagmatic deposit isthe nickel–copper sulfide orebody formed bythe sinking of an immiscible sulfide liquidto the bottom of a magma chamber containingultrabasic or basic magma. These are known asliquation deposits and they may be formed inthe bottom of lava flows as well as in plutonicintrusions. The sulfide usually accumulates in

hollows in the base of the igneous body andgenerally forms sheets or irregular lenses con-formable with the overlying silicate rock. Fromthe base upwards, massive sulfide gives waythrough disseminated sulfides in a silicategangue to lightly mineralized and then barrenrock (Fig. 3.9).

Metamorphic host rocksApart from some deposits of metamorphic ori-gin such as the irregular replacement depositsalready described and deposits generated in con-tact metamorphic aureoles – e.g. wollastonite,andalusite, garnet, graphite – metamorphicrocks are important for the metamorphosedequivalents of deposits that originated in sedi-mentary and igneous rocks and which havebeen discussed above.

Residual depositsThese are deposits formed by the removal ofnonore material from protore (rock in whichan initial but uneconomic concentration ofminerals is present that may by further naturalprocesses be upgraded to form ore). For ex-ample, the leaching of silica and alkalis from anepheline–syenite may leave behind a surfacecapping of hydrous aluminum oxides (bauxite).Some residual bauxites occur at the presentsurface, others have been buried under youngersediments to which they form conformablebasal beds. The weathering of feldspathic rocks(granites, arkoses) can produce important kao-lin deposits which, in the Cornish granites ofEngland, form funnel or trough-shaped bodiesextending downwards from the surface for asmuch as 230 m.

Other examples of residual deposits includesome laterites sufficiently high in iron to beworked and nickeliferous laterites formed bythe weathering of peridotites.

3.2 WALL ROCK ALTERATION

Many ore deposits, particularly the epigeneticones, may have beside or around them a zone orzones of wall rock alteration. This alteration ofthe host rock is marked by color, textural,mineralogical or chemical changes or any com-bination of these. The areal extent of the altera-tion can vary considerably, sometimes being

Page 57: Introduction to mineral exploration

40 A.M. EVANS & C.J. MOON

FIG. 3.9 Generalized section through the Creighton ore zone, Sudbury, Ontario, looking west. (After Souchet al. 1969.)

Surface

Norite

0 0.5 km

Footwall rocks

Interstitial sulfide in norite

Ragged disseminated sulfide

Gabbro-peridotite inclusionsulfide and massive sulfide

limited to a few centimeters on either side ofa vein, at other times forming a thick haloaround an orebody and then, since it widensthe drilling target, it may be of considerableexploration value. Hoeve (1984) estimated thatthe drilling targets in the uranium field of theAthabasca Basin in Saskatchewan are enlargedby a factor of 10–20 times by the wall rockalteration. The Atlas of Alteration by Thomp-son and Thompson (1996) is a very well illus-trated place to start your study.

3.3 GEOLOGICAL MODELS OF MINERALDEPOSITS

One of the aims of the planning stage (seeChapter 4) is to identify areas for reconnais-sance and to do this we must have some idea ofhow the materials sought relate to geologicalfactors including geophysics and geochemistry.This is best achieved by setting up a model ormodels of the type of deposit sought. But whatis a model? The term has been defined in vari-ous ways but a useful one is that of “functionalidealization of a real world situation used to aidin the analysis of a problem.” As such it is a

synthesis of available data and should includethe most informative and reliable characteris-tics of a deposit type, identified on a variety ofscales and including definition of the averageand range of each characteristic (Adams 1985).It is therefore subject to uncertainty andchange; each new discovery of an example of adeposit type should be added to the data base.

In mineral deposit models there are two maintypes which are often combined; the empiricalmodel based on deposit descriptions and agenetic model which explains deposits in termof causative geological processes. The geneticmodel is necessarily more subjective but can bemore powerful, as it can predict deposits notcontained in the descriptive data base. Anothertype of model which is extremely useful forpreliminary economic evaluations is a grade-tonnage model. This accumulates grade andtonnage data for known deposits and from thisit is possible to estimate the size and grade of anaverage or large deposit and the cash flow if onewere found. Examples of this type of modelingare given by Gorman (1994) for South Americangold and copper deposits. Cost curves can alsobe calculated for differing deposit types. Exam-ples for copper (Fig. 3.10) show the low cost of

Page 58: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 41

200 10

30

100

Cumulative production (billion lb)

90

80

70

60

50

40

987654321

Porphyrycopper deposits

Sedimentaryhosted deposits

Other types of deposits

Volcanic hostedmassive sulfide deposits

C1

prod

uctio

n co

st 1

994

(US

/¢lb

–1)

FIG. 3.10 Cost curves for copper production.The C1 cost is the cash operating cost. Note thepredominance of porphyry copper production.(After Moore 1998.)

producing from porphyry copper deposits andthe economies of scale from large productionrelative to sedimentary copper and volcanic-associated massive sulfide deposits.

To understand how geological models areconstructed we will examine the deposit modelfor unconformity-related uranium deposits.These deposits are currently the main source ofhigh grade uranium in the Western world andmainly occur in two basins; the AthabascaBasin of Northern Saskatchewan and theAlligator River area of the Northern Territoryin Australia. Good accounts of most depositmodels can be found in two publications byNorth American Geological Surveys, Cox andSinger (1986) and Eckstrand (1984), later re-vised as Eckstrand et al. (1995). Cox and Singeradopt a strictly descriptive approach, based ontheir classification of mineral deposits (see sec-tion 3.6), but also give grade-tonnage curvesand geophysical signatures whereas Eckstrandis more succinct but gives brief genetic models.Some major deposit types are also dealt with indepth by Roberts and Sheahan (1988), Kirkhamet al. (1993), and a special edition of the Aus-tralian Geological Survey Organisation journal(AGSO 1998). Internet versions of the USGSdeposit models and a more extensive classifica-tion by the British Columbia Geological Surveyare available at their web sites (BCGS 2004,USGS 2004).

Unconformity

Glacial till

Basementrocks

Limit ofhydrothermal

alteration

Carbon -bearing rocks,schists, marbles, etc.

Fault or shear zone

Sandstone●

● ●

●●

●●

● ●●

Massive ore

Fracturecontrolled ore Approximate scale

0 100 m

FIG. 3.11 Generalized diagram of an unconformity-associated uranium deposit. (After Clark et al. 1982.)

Table 3.1 shows the elements that wereused by Eckstrand and Cox and Singer intheir unconformity-related deposit models.Both accounts agree that the deposits occur ator near the unconformable contact betweenregionally metamorphosed Archaean to LowerProterozoic basement and Lower to MiddleProterozoic (1900–1200 Ma) continental clasticsediments, as shown in Fig. 3.11. The detailsof the controls on mineralisation are more sub-jective; the Canadian deposits occur both in theoverlying sandstones and basement whereasthe Australian deposits are almost exclusivelyin the basement. Most deposits in Canada arespatially associated very closely with graphiticschists whereas the Australian deposits areoften in carbonates, although many are car-bonaceous. Both models agree that the keyalteration is chloritization together with seri-citization, kaolinization, and hematitizationalong the intersection of faults and the un-conformity. Generalization on the mineralogyof the deposits is difficult but the key mineralis pitchblende with lesser amounts of the

Page 59: Introduction to mineral exploration

42 A.M. EVANS & C.J. MOON

TABLE 3.1 Summary of elements used in the models for unconformity-related uranium deposits by Grauch andMoiser in Cox and Singer (1986) and Tremblay and Ruzicka in Eckstrand (1984).

Element

Synonym:Commodities(Eckstrand)

Description

Geological environmentRock types

Textures

Age range

Depositionalenvironment:Associated rocks(Eckstrand)

Tectonic setting(Cox & Singer)Geological setting(Eckstrand)

Associated deposittypes

Form of deposit

Deposit descriptionMineralogy

Cox & Singer

Vein-like type U

U mineralisation in fracturesand breccia fills in metapelites,metapsammites, and quartzarenites below, across, and abovean unconformity separating earlyand middle Proterozoic rocks

Regionally metamorphosedcarbonaceous pelites, psammites,and carbonates. Youngerunmetamorphosed quartz arenites

Metamorphic foliation and laterbrecciation

Early and middle Proterozoicaffected by Proterozoic regionalmetamorphism

Host rocks are shelf deposits andoverlying continental sandstone

Intracratonic sedimentary basinson the flanks of Archaean domes.Tectonically stable since theProterozoic

Gold- and nickel-rich depositsmay occur

Pitchblende + uraninite ± coffinite± pyrite ± galena ± sphalerite ±arsenopyrite ± niccolite. Chlorite +quartz + calcite + dolomite + hematite+ siderite + sericite. Late veins containnative gold, uranium, and tellurides.Latest quartz–calcite veins containpyrite, chalcopyrite, and bituminousmatter

Eckstrand

U ( Ni, Co, As, Se, Ag, Au, Mo)

Ores occur in clay sericite and chloritemasses at the unconformity and alongintersecting faults and in clay-alteredbasement rocks and kaolinized cover rocks

Most basement rocks are Aphebian, somemay be Archaean. Helikian cover rocks.Ore: Helikian (1.28 ± 0.11 Ga)

Basement: graphitic schist and gneiss,coarse-grained granitoids, calc silicatemetasediments. Cover rocks: sandstoneand shale

Relatively undeformed, intracratonic,Helikian sedimentary basin restingunconformably on intensely deformedArchaean and Aphebian basement.Deposits are associated with unconformitywhere intersected by faults. Palaeoregolithpresent in Saskatchewan

Flattened cigar shaped, high grade bodiesoriented and distributed along theunconformity and faults, especiallylocalized at unconformity–faultintersections. High grade bodies gradeoutwards into lower grade

Pitchblende coffinite, minor uraniumoxides. Ni and Co arsenides and sulfides,native selenium and selenides, nativegold and gold tellurides, galena, minormolybdenite, Cu and Fe sulfides, clayminerals, chlorite, quartz, graphite,carbonate

Page 60: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 43

TABLE 3.1 (continued)

Element Cox & Singer Eckstrand

Texture and structure

Alteration

Ore controls

Weathering

Geochemical andgeophysicalSignature (Cox &Singer)

Genetic model

Examples

Importance

Typical gradeand tonnage

Breccias, veins, and disseminations.Uranium minerals coarse andcolloform. Latest veins have openspace filling textures

Multistage chloritization dominant.Local sercitization, hematization,kaolinization, and dolomitization.Vein silicification in alterationenvelope. Alteration enriched inMg, F, REE, and various metals.Alkalis depleted

Fracture porosity controlled oredistribution in metamorphics.Unconformity acted as disruption influid flow but not necessarily ore locus

Various secondary U minerals

Increase in U, Mg, P, and locally Ni,Cu, Pb, Zn, Co, As; decrease in SiO2.Locally Au with Ag, Te, Ni, Pd, Re,Mo, Hg, REE, and Rb. Anomalousradioactivity. Graphitic schists insome deposits are strong EMconductors

Rabbit Lake, Saskatchewan CluffLake Saskatchewan

see Fig. 3.12

1 At or near unconformity betweenHelikian sandstone and Archaeanbasement2 Intersection of unconformity withreactivated basement faults3 Associated with graphitic basementrocks and gray and/or multicolored shaleand sandstone cover4 Intense clay, chlorite, and sericitealteration of basement and cover rocks5 Basement rocks with higher thanaverage U content

Combinations of:1 Preconcentration of U during depositionof Aphebian sediments and their anatexis2 Concentration in lateritic regolith(Helikian)3 Mobilization by heated oxidizedsolutions and precipitation in reducingenvironment at unconformity and faultlocus4 Additional cycles of mobilization andprecipitation leading to redistribution ofuranium

Key Lake, Rabbit Lake, Cluff Lake,Canada. Jabiluka I and II, Ranger, N.T.,Australia

Canada: 35% of current U production but50% of reservesWorld: 15% of reservesCanada: small to 5 Mt of 0.3–3% U.Australia: maximum 200,000 t ofcontained U but grade lower than Canada

Page 61: Introduction to mineral exploration

44 A.M. EVANS & C.J. MOON

uranium silicate, coffinite. The more contenti-ous issue is the associated minerals; in Canadahigh grade nickel arsenides are common, but byno means ubiquitous, whereas the mineralogyof the Australian deposits is simpler. However,gold is more common and may be sufficientlyabundant to change the deposit model tounconformity-related uranium–gold. Selenidesand tellurides are present in some deposits,although Cox and Singer omit the former butmention enrichment in palladium.

The genetic model for these deposits in-cludes elements of the following:1 Preconcentration of uranium and associatedelements in basement sedimentary rocks.2 Concentration during the weathering of thebasement prior to the deposition of the over-lying sediments.3 Mobilization of uranium and the associatedelements by oxidizing fluids and precipitationin a reducing environment at fault–unconformity intersections.4 Additional cycles of oxidation and mobiliza-tion.

As an example of the differences between thedescriptive and genetic models, the descriptivemodel would suggest exploration for graphiticconductors whereas the genetic model wouldbroaden the possible depositional sites to anyreducing environment. Descriptive and geneticmodels can be combined with the grade-tonnage curves of Cox and Singer (Fig. 3.12)to suggest the probable economic benefits. Itshould be noted that high tonnage does notnecessarily mean low grade, and large depositssuch as Cigar Lake (1.47 Mt of 11.9% U3O8) inSaskatchewan have grade and tonnage thatrank in the top 10% of all deposits.

The history of exploration in the two areasclearly illustrates the uses and limitations ofmodels. Unconformity vein deposits were firstdiscovered as a result of exploration in knownuraniferous areas of Northern Canada (Fig. 3.13)and Australia in the late 1960s. The initialCanadian discoveries were made by a Frenchcompany which was exploring for uraniumveins around the Beaverlodge deposit. Thesehave no obvious relation to the mid Proterozoicsandstones and it was only the company’s pre-vious experience in Gabon and minor knownoccurrences near the unconformity whichsuggested that the sandstones might be miner-

alized (Tona et al. 1985). On this premise thecompany (Amok) conducted an airborne radio-metric survey which led them to bouldersof mineralized sandstone and massive pitch-blende close to the source deposit (Cluff Lake

0.1

0.0

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Pro

port

ion

of d

epos

its

0.03

2

0.05

6

0.1

0.18

0.32

0.56 1.

0

1.8

3.2

5.6 10

Uranium grade in percent U3O8

0.18 0.49 2.0

n =36

0.1

0.0

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Pro

port

ion

of d

epos

its

0.00

04

0.00

16

0.00

63

0.02

5

0.1

0.4

1.6

6.3 25 100

400

Million tonnes

0.0055 0.26 9.6

n =361.1(a)

(b)

FIG. 3.12 (a,b) Grade–tonnage curves forunconformity associated deposits. (From Cox &Singer 1986.)

Page 62: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 45

exports, which inhibited uranium exploration.The importance of recognizing the elementsassociated with this type of deposit becameobvious in the late 1980s with the discoveryof the Coronation Hill gold–palladium depositin the Northern Territory and similar prospectsin Saskatchewan. Unfortunately for the com-pany concerned Coronation Hill is on the edgeof the Kakadu National Park and will not bemined in the foreseeable future.

The application of a particular deposit modelwill depend on the quality of the database andshould not be regarded as a panacea for the ex-ploration geologist. Some deposit types such asplacer gold are easy to understand and havewell-developed models, whereas others such asBesshi-style massive sulfides or the OlympicDam models are not well developed and may berepresented by a single deposit on which infor-mation is difficult to obtain.

The major pitfalls of using models are ablyillustrated in cartoon form by Hodgson (1989).He identifies a number of problems which helikens to religious cults:1 The cult of the fad or fashion. An obsessionwith being up to date and in possession of thenewest model.2 The cult of the panacea. The attitude thatone model is the ultimate and will end allcontroversy.3 The cult of the classicists. All new ideas arerejected as they have been generated in the hothouse research environment.4 The cult of the corporate iconoclasts. Onlymodels generated within an organization arevalid, all outside models are wrong.5 The cult of the specialist. In which only oneaspect of the model is tested and usually not inthe field.

The onus is thus on the exploration geologistto ensure that a model is correctly applied andnot to exclude deviations from the norm. Otherexamples of the application of models are givenin Chapters 12–17.

3.4 FURTHER READING

Longer discussions of the subjects forming themain sections of this chapter can be found inRobb Introduction to Ore-Forming Processes(2004). Kesler’s Mineral Resources, Economics,

RabbitLake

AthabascaFormation

Cluff Lake

LakeAthabasca

Unconformity

McCleanLake

Key Lake

300 m depth

0 100 km

Cigar Lake

McArthurRiver

FIG. 3.13 Outline of the Athabasca Basin anddistribution of the associated uranium deposits.(Modified after Clark et al. 1982.)

A). At roughly the same time a small Canadiancompany explored the eastern edge of the basinknowing that they had an aircraft available andthat the sandstones might be prospective bycomparison with the much younger WesternUS deposits (Reeves & Beck 1982). They alsodiscovered glacial boulders which led themback to the deposit (Rabbit Lake). At roughlythe same time (1969) in Australia airborne sur-veys in the East Alligator Valley detected majoranomalies at Ranger, Koongarra, and Narbalek.The area selection was initially based on thesimilarity of the geology with known produc-ing areas in South Alligator Valley, with acidvolcanics as the postulated source rock, andby a re-interpretation of granites, that were pre-viously thought to be intrusive, as basementgneiss domes (Dunn et al. 1990b). When drill-ing started it became clear that the miner-alisation was closely associated with theunconformity. Thus the overlying sandstones,rather than being responsible for the erosion ofthe deposits or being a barren cover, became atarget. In Canada the Key Lake and MidwestLake deposits were discovered as a result of thischange in model (Gatzweiler et al. 1981, Scott1981). The exploration model has been furtherdeveloped in Canada, to the point at whichdeposits with no surface expression (blinddeposits), such as Cigar Lake and McArthurRiver, have been found by deep drilling of elec-tromagnetic conductors caused by graphiticschists (McMullan et al. 1989). Similar devel-opments could have been anticipated in Aus-tralia but for government control on uranium

Page 63: Introduction to mineral exploration

46 A.M. EVANS & C.J. MOON

and the Environment (1994) provides a goodintroduction and a link with economics.Sawkins’ Metal Deposits in Relation to PlateTectonics (second edition 1990, SpringerVerlag) gives an overview of the major struc-tural controls on deposits.

3.5 SUMMARY OF CHAPTERS 1, 2, AND 3

The general nature of ore, industrial minerals,and orebodies are discussed in sections 1.1 and1.2 and it is emphasized that these depositsmust contain valuable constituents that can beeconomically recovered using suitable treat-ment. (Chapter 11 is largely devoted to a de-tailed coverage of this principle.) In the past toomuch interest in many mining circles has beenplaced on the more glamorous metallic de-posits and the general importance of industrialminerals neglected (see section 1.2.2). Thevalue of both is governed by demand andsupply (see section 1.2.3) and many factors in-cluding government action, recycling, and sub-stitution play a part in determining theirmarket prices. In section 1.3 the principal stepsinvolved in the exploration for, and develop-ment of, a mineral deposit are summarizedand these are the subjects that are covered inmore detail in the rest of this book. A soundknowledge of these is necessary in choosingexploration areas (see section 1.5.) leading tothe development of a rationale of mineral ex-ploration (see section 1.6).

Economic mineral deposits are extremelyvariable in their mineralogy and grade and the

geologist must be able to identify, or have iden-tified for him or her, every mineral in a possibleorebody (see Chapter 2). This will help them toassess the full economic potential of any mate-rial to be mined from it and prevent them over-looking the presence of additional valuableconstituents or deleterious substances thatmay render the deposit unworkable. There aremany techniques now available for compre-hensive mineralogical examination of mineralsamples and these are discussed in section 2.2.However, such investigations must be quant-itative as well as qualitative; grain size andshape, relative mineral amounts, and the man-ner of interlocking must be determined.

The nature and morphology of mineral de-posits are very varied and only a restricted cov-erage of these subjects can be given in this book(see section 3.1). The tyro is therefore stronglyrecommended to acquire a broad knowledge ofthese subjects from extended reading so thatwhen he or she detects signs of mineralisationthey may soon develop a working hypothesis ofthe nature of the particular beast they havecome upon.

Having recognized the tectonic setting ofan exploration region, models of the depositslikely to be present, and particularly of thosebeing sought, should be set up. Both empiricaland genetic models are used, often in combina-tion, to show how the materials sought relateto geological factors including geophysics andgeochemistry. To understand how models areconstructed and used, unconformity-relateduranium deposits are discussed in some detail(see section 3.2.3).

3.6 APPENDIX: MINERAL DEPOSIT MODEL CLASSIFICATION OF COX AND SINGER (1986). ONLINELINKS CAN BE FOUND AT USGS (2004).

Mafic and ultramafic intrusions

A. Tectonically stable area; stratiform complexes

Stratiform depositsBasal zone

Stillwater Ni-Cu 1Intermediate zone

Bushveld chromitite 2aMerensky Reef PGE 2b

Upper zoneBushveld Fe-Ti-V 3

Page 64: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 47

Pipe-like depositsCu-Ni pipes 4aPGE pipes 4b

B. Tectonically unstable area

Intrusions same age as volcanic rocksRift environment

Duluth Cu-Ni-PGE 5aNoril’sk Cu-Ni-PGE 5b

Greenstone belt in which lowermost rocks of sequence contain ultramafic rocksKomatiitic Ni-Cu 6aDunitic Ni-Cu 6b

Intrusions emplaced during orogenesisSynorogenic in volcanic terrane

Synorogenic-synvolcanic Ni-Cu 7aSynorogenic intrusions in nonvolcanic terrane

Anorthosite-Ti 7bOphiolite

Podiform chromite 8aMajor podiform chromite 8b(Lateritic Ni) (38a)(Placer Au-PGE) (39a)

SerpentineLimassol Forest Co-Ni 8cSerpentine-hosted asbestos 8d(Silica-carbonate Hg) (27c)(Low-sulfide Au-quartz vein) (36a)

Cross-cutting intrusions (concentrically zoned)Alaskan PGE 9(Placer PGE-Au) (39b)

C. Alkaline intrusions in stable areas

Carbonatite 10Alkaline complexes 11Diamond pipes 12

Felsic intrusions

D. Mainly phanerocrystalline textures

PegmatiticBe-Li pegmatites 13aSn-Nb-Ta pegmatites 13b

Granitic intrusionsWallrocks are calcareous

W skarn 14aSn skarn 14bReplacement Sn 14c

Other wallrocksW veins 15aSn veins 15b

Page 65: Introduction to mineral exploration

48 A.M. EVANS & C.J. MOON

Sn greisen 15c(Low-sulfide Au-quartz vein) (36a)(Homestake Au) (36c)

Anorthosite intrusions(Anorthosite Ti) (7b)

E. Porphyroaphanitic intrusions present

High-silica granites and rhyolitesClimax Mo 16(Fluorspar deposits) (26b)

Other felsic and mafic rocks including alkalicPorphyry Cu 17Wallrocks are calcareous

Deposits near contactPorphyry Cu, skarn-related 18a

Cu skarn 18bZn-Pb skarn 18cFe skarn 18dCarbonate-hosted asbestos 18e

Deposits far from contactPolymetallic replacement 19aReplacement Mn 19b(Carbonate-hosted Au) (26a)

Wallrocks are coeval volcanic rocksIn granitic rocks in felsic volcanics

Porphyry Sn 20aSn-polymetallic veins 20b

In calcalkalic or alkalic rocksPorphyry Cu-Au 20c(Epithermal Mn) (25g)

Wallrocks are older igneous and sedimentary rocksDeposits within intrusions

Porphyry Cu-Mo 21aPorphyry Mo, low-F 21bPorphyry W 21c

Deposits within wallrocksVolcanic hosted Cu-As-Sb 22aAu-Ag-Te veins 22bPolymetallic veins 22c(Epithermal quartz-alunite Au) (25e)(Low-sulfide Au-quartz vein) (36a)

Extrusive rocks

F. Mafic extrusive rocks

Continental or rifted cratonBasaltic Cu 23(Sediment-hosted Cu) (30b)

Marine, including ophiolite-relatedCyprus massive sulfide 24a

Page 66: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 49

Besshi massive sulfide 24bVolcanogenic Mn 24cBlackbird Co-Cu 24d(Komatiitic Ni-Cu) (6a)

G. Felsic-mafic extrusive rocks

SubaerialDeposits mainly within volcanic rocks

Hot-spring Au-Ag 25aCreede epithermal vein 25bComstock epithermal vein 25cSado epithermal vein 25dEpithermal quartz-alunite Au 25eVolcanogenic U 25fEpithermal Mn 25gRhyolite-hosted Sn 25hVolcanic-hosted magnetite 25i(Sn polymetallic veins) (20b)

Deposits in older calcareous rocksCarbonate-hosted Au-Ag 26aFluorspar deposits 26b

Deposits in older elastic sedimentary rocksHot-spring Hg 27aAlmaden Hg 27bSilica-carbonate Hg 27cSimple Sb 27d

MarineKuroko massive sulfide 28aAlgoma Fe 28b

(Volcanogenic Mn) (24c)(Volcanogenic U) (25f)(Low-sulfide Au-quartz vein) (36a)(Homestake Au) (36b)(Volcanogenic U) (25f)

Sedimentary rocks

H. Clastic sedimentary rocks

Conglomerate and sedimentary brecciaQuartz pebble conglomerate Au-U 29aOlympic Dam Cu-U-Au 29b(Sandstone U) (30c)(Basaltic Cu) (23)

SandstoneSandstone-hosted Pb-Zn 30aSediment-hosted Cu 30bSandstone U 30c(Basaltic Cu) (23)(Kipushi Cu-Pb-Zn) (32c)(Unconformity U-Au) (37a)

Page 67: Introduction to mineral exploration

50 A.M. EVANS & C.J. MOON

Shale-siltstoneSedimentary exhalative Zn-Pb 31aBedded barite 31bEmerald veins 31c(Basaltic Cu) (23)(Carbonate-hosted Au-Ag) (26a)(Sediment-hosted Cu) (30b)

I. Carbonate rocks

No associated igneous rocksSoutheast Missouri Pb-Zn 32aAppalachian Zn 32bKipushi Cu-Pb-Zn 32c(Replacement Sn) (14c)(Sedimentary exhalative Zn-Pb) (31a)(Karst bauxite) (38c)

Igneous heat sources present(Polymetallic replacement) (19a)(Replacement Mn) (19b)(Carbonate-hosted Au-Ag) (26a)(Fluorspar deposits) (26b)

J. Chemical sediments

OceanicMn nodules 33aMn crusts 33b

ShelfSuperior Fe 34aSedimentary Mn 34bPhosphate, upwelling type 34cPhosphate, warm-current type 34d

Restricted basinMarine evaporate 35aPlaya evaporate 35b(Sedimentary exhalative Zn-Pb) (31a)(Sedimentary Mn) (34b)

REGIONALLY METAMORPHOSED ROCKS

K. Derived mainly from eugeosynclinal rocks

Low-sulfide Au-quartz vein 36aHomestake Au 36b(Serpentine-hosted asbestos) (8d)(Gold on flat faults) (37b)

L. Derived mainly from pelitic and other sedimentary rocks

Unconformity U-Au 37aGold on flat faults 37b

Page 68: Introduction to mineral exploration

3: MINERAL DEPOSIT GEOLOGY AND MODELS 51

Surficial and unconformity-related

M. Residual

Lateritic Ni 38aBauxite, laterite type 38bBauxite, karst type 38c(Unconformity U-Au) (37a)

N. Depositional

Placer Au-PGE 39aPlacer PGE-Au 39bShoreline placer Ti 39cDiamond placers 39dStream placer Sn 39e(Quartz pebble conglomerate Au-U) (29a)

Page 69: Introduction to mineral exploration

52 C.J. MOON & M.K.G. WHATELEY

4Reconnaissance

ExplorationCharles J. Moon and Michael K.G. Whateley

FIG. 4.1 Stages of an exploration project. (Modified from Eimon 1988.)

MiningDecommiss-ioning and

rehabilitation

EX

PE

ND

ITU

RE

– L

OG

AR

ITH

MIC

SC

ALE

AC

TIV

ITY

Conceptualplanning

Detailedplanning

Reconn-aissance

exploration

Targetappraisal

Explorationdrilling

Assessmentdrilling

Minedevelop.

Test miningand

metallurgicalstudies

Productionshaft,

decline oroverburden

removal

Key decision node

Rate of expenditure

HIGH RISK

DECREASINGRISK

RegionalSelection

AreaSelection

PrefeasibilityStudy

FeasibilityStudy

Mineproductioninitiation

Literaturereview

Literaturereview

field visit

Remotesensing,regional

geochem,airborne

geophysics

Geochem,ground

geophysics,mapping,surveying

Pilotmetallurgical

studies

Exploration can be divided into a number of in-terlinked and sequential stages which involveincreasing expenditure and decreasing risk (Fig.4.1). The terminology used to describe thesestages is highly varied. The widely acceptedterms used for the early stages of explorationare planning and reconnaissance phases. These

phases cover the stages leading to the selectionof an area for detailed ground work; this is usu-ally the point at which land is acquired. Theplanning stage covers the selection of commod-ity, type of deposit, exploration methods, andthe setting up of an exploration organization.The process of selecting drill targets within

Page 70: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 53

license blocks we term target selection and thatof drilling, target testing (discussed in Chap-ter 5). The deposit is then at the stage of pre-development followed by a feasibility study(Chapter 11). Before we consider how explora-tion is planned we should discuss who explores.

The exploration players

Private sectorMost mineral exploration in developed coun-tries is conducted by companies with a sub-stantial capital base generated either fromexisting mineral production or from investorson stock markets. The size of the companycan range from major multinational miningcompanies, such as Rio Tinto plc or AngloAmerican Corporation with operations on sev-eral continents, to small venture capital com-panies, usually known as a junior company,with one or two geologists. Exploration byindividual prospectors has been an importantfactor in countries with large unexplored areasand liberal land tenure laws, e.g. Canada,Australia and Brazil. Here you can still meetthe grizzled prospector or garimpeiro. Althoughthey usually lack the sophisticated training ofthe corporate geologist, this can be compen-sated for by a keen eye and the willingness toexpend a little boot leather.

State organizationsIn more centrally directed economies, mostexploration is carried out by state run com-panies, geological surveys, and often in the caseof developing countries, international aidorganizations. The role of a geological surveyusually includes some provision of informationon mineral exploration to government and theprivate sector. Usually this takes the form ofreconnaissance work. By contrast, the SovietMinistry of Geology had until 1991 exclusiveprospecting rights in the former USSR, al-though exploration was carried out by a widevariety of organizations at the Union (federal)and republic levels. In China exploration isundertaken by a number of state and provincialgroups, including the army.

These two groups essentially explore insimilar ways although state enterprises aremore constrained by political considerationsand need not necessarily make a profit. The

remainder of the chapter will be devoted toprivate sector organizations although muchmay be applicable to state organizations.

4.1 EXPLORATION PLANNING

Mineral exploration is a long-term commit-ment and there must be careful planning of acompany’s long-term objectives. This shouldtake particular regard of the company’sresources and the changing environment inwhich it operates (Riddler 1989). The key fac-tors are:1 Location of demand for products. This willdepend on the areas of growth in demand. Formetals the most obvious areas are the industri-alizing countries of the Pacific Rim, and whoare resource deficient. China has assumed par-ticular importance since the late 1990s for awide range of commodities.2 Metal prices. Price cycles should be esti-mated as far as possible and supply and demandforecast (see section 1.2.3).3 Host country factors. The choice of countryfor operation is important in an industry whichhas seen substantial nationalization, such ascopper in The Democratic Republic of Congoand Zambia and coal in the UK. The standingof foreign investment, the degree of controlpermitted, percentage of profits remittable tothe home country, and most of all governmentstability and attitude are important. Otherfactors are availability of land, security oftenure, and supply of services and skilled labor(see section 1.4).4 The structure of the mining industry.Barriers to the entry of new producers are com-petition from existing producers and the needfor capital to achieve economies of scale.

These factors combined to encourage explo-ration companies in the 1980s to concentrateon precious metals in Australia, Canada, andthe USA (see section 1.2.3 “Metal and mineralprices”). In the 1990s the net was more widelyspread with access possible to the countries ofthe former USSR and exploration undertakenin countries with high political risk. The late1990s saw a concentration of exploration activ-ity in countries with proactive mineral policiesin South America and selected parts of Africaand Asia. There was a revival in interest in base

Page 71: Introduction to mineral exploration

54 C.J. MOON & M.K.G. WHATELEY

2 Sound basis of operations. The organizationworks within corporate guidelines towardsobjectives.3 Creative and productive atmosphere. Thegroup encourages independent creative andinnovative thinking in an environment freefrom bureaucratic disruption.4 High standard of performance, integrity, andethics.5 Entrepreneurial acumen. Innovation is fos-tered in a high risk, high reward environment.6 Morale and team spirit. High morale, enthu-siasm, and a “can do” attitude.7 The quality of communication is high. The“top brass” are aware of the ideas of geologists.8 Pre-development group. Successful organiza-tions are more likely to have a specialist groupresponsible for the transition of a deposit fromexploration to development.

All these points make it clear that themanagement must consist of flexible indi-viduals with considerable experience ofexploration.

If these are the optimum characteristicsof an exploration group, is there an optimumsize and what structure should it have? Stud-ies such as those of Holmes (1977) show thatthe most effective size is in the range sevento ten geologists; larger organizations tend tobecome too formalized and bureaucratic, lead-ing to inefficiency, whereas small groupslack the budgets and the manpower to mounta successful program. For the large mininggroup which wishes to remain competitivewhile spending a large budget, the solutionis to divide its explorationists into semi-autonomous groups.

Exploration groups can be organized on thebasis of geographical location or of deposittype. The advantage of having deposit special-ists is that in-depth expertise is accumulated;however, the more usual arrangement is toorganize by location with geologists in eachregion forming specialist subgroups. At thereconnaissance stage most work will be car-ried out in offices located in the national headoffice or state office, but as exploration focusesin more detail smaller district offices can beopened. A typical arrangement for a large com-pany is shown in Fig. 4.2. In a company withexisting production there may be close liaisonwith mines and engineering divisions.

and ferrous metals in the early 2000s due to theinitially lower gold price and boom in demandfrom China.

After the initial corporate planning, usuallyby senior executives, an exploration strategymust be chosen, a budget allocated, and desir-able deposit type(s) defined.

The choice of exploration strategy variesconsiderably between companies and dependson the objects of the company and its willing-ness to take risks. For new entrants into acountry the choice is between exploration byacquisition of existing prospects or grass roots(i.e. from scratch) exploration. Acquisitionrequires the larger outlay of capital but carrieslower risk and has, potentially, a shorter leadtime to production. Acquisitions of potentialsmall producers are particularly attractive tothe smaller company with limited cash flowfrom existing production. Potential large pro-ducers interest larger companies which havethe capital necessary to finance a large project.This has been particularly marked in the earlyyears of the twenty-first century in the con-solidation of the gold industry into a few largeproducers that have devoured the mediumsized producers. Larger companies tend to ex-plore both by acquisition and by grass rootsmethods and often find that exploration pre-sence in an area will bring offers of properties(“submittals”). Existing producers have theadditional choice of exploring in the immedi-ate vicinity of their mines, where it is likelythat they will have substantial advantages incost saving by using existing facilities. Mostof the following section refers to grass rootsexploration, although evaluation of potentialacquisitions could run in parallel.

4.1.1 Organization

The key to exploration organization is to havethe best available staff and adequate finance inorder to create confidence throughout the or-ganization (Woodall 1984, Sillitoe 1995, 2000).A number of factors that characterize a suc-cessful exploration team have been recognizedby Snow and Mackenzie (1981), Regan (1971),and discussed in detail by White (1997).1 High quality staff and orientation towardspeople. Successful organizations tend to pro-vide more in-house training.

Page 72: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 55

President

Central Engineering

Vice - President, Mining

Mines Managers

Vice - President, Exploration

Manager, Exploration Manager, Geophysics

Region Managers GeophysicalServices

GeophysicalResearch

ManagerGeologicalResearch

GeochemistryLabs.

GeologyLabs.

Admin.Services

ChiefGeologist

SpecialProjects

Chief MineGeologist

ManagerStaff

Functions

ManagerMines

Geology

ProjectGeologists

StaffGeologists

Project Managers

FIG. 4.2 Organization of exploration in a large mining group with producing mines.

4.1.2 Budgets

Exploration costs are considered in two ways:(i) as an expenditure within an organization and(ii) within the context of a project. It is usuallythe exploration manager that considers theformer, but it is as a geologist on a specificexploration project that one becomes involvedin the latter.

Corporate exploration expenditure

Finance for corporate exploration is derivedfrom two main sources, revenue from existingproduction and by selling shares on the stockmarket. In the first case the company sets asidea percentage of its before-tax profits for explora-tion. The decision as to the percentage set asideis based upon how much the company wishesto keep as capital, for their running costs, asdividend for their shareholders, and for taxes.Exploration costs may range from 1% to 20%of the annual corporate cash flow but for largediversified companies this is an average 2.5%of sales and 6% for gold companies (Crowson2003). This may be anything between US$0.5

million to $100 million per annum dependingupon the size of the company and the sizeof the deposit for which they are searching(Table 4.1). For producing companies mostexploration can be written off against tax. Thesmaller company is at a disadvantage in thatmoney must be raised from shareholders. Thisis easy in times of buoyant share prices. Forexample, about $C2 billion were raised in thelate 1980s for gold exploration in Canada,largely on the Vancouver Stock Exchange. Thiswas aided by flow-through schemes whichenabled share purchasers to offset this againsttax liabilities. In a similar way, but without taxbreaks, about US$100 million was raised onthe Irish Stock Exchange from 1983 to 1989(Gardiner 1989). At the peak of the 1996 boom$1.1 billion was raised in one year on theVancouver Stock Exchange. This was throughinitial public offerings of companies and privateplacements to individuals and investmentfunds (Hefferman 1998). Although major min-ing groups should be able to maintain con-sistent budgets and avoid business cyclicity,there seems little evidence of this. In a study ofmajor groups Eggert (1988) demonstrated that

Page 73: Introduction to mineral exploration

56 C.J. MOON & M.K.G. WHATELEY

TABLE 4.1 Corporate Exploration Spending in 2003(Source: Corporate websites and Mining Journal,March 2004.)

Company US$ million

De Beers 140Rio Tinto 140Barrick Gold 110Newmont 86Companhia Vale do Rio Doce 81BHP Billiton 68Anglogold 63Anglo American (including 61

platinum)Placer Dome 60Phelps Dodge (including 50

research)Noranda-Falconbridge 35Teck-Cominco 30Inco 27Gold Fields 23WMC Resources 20Newcrest 19

Total spending ~2400

–200

Con

stan

t 19

82 U

S$

(mill

ions

)

US

$ (m

illio

ns)

–100

0

100

200

300

400

500550 60

50

40

30

20

10

0

Expenditure (E)

Income (I)Inco

Netincome

Explorationexpenditures

1975 19801970

–100

0

100

200

300

400450

40

30

20

10

0

Noranda

FIG. 4.3 Relation between exploration expenditureand income for two Canadian mining companies.(After Eggert 1988.)

spending was linked to income, and thereforemetal prices, but lagged about 18 monthsbehind the changes in income (Fig. 4.3).

The overall budget is then subdivided. Forexample, a multinational corporation may haveseveral regional exploration centers, each ofwhich may have anything from one or twopersons to a fully equipped office with up to50 people employed. The budget for the lattermust include the salaries, equipment, officerentals, and vehicular leases before a singlegeologist sets foot in the field. Each projectwould be given a percentage of the regionaloffice’s budget, e.g. exploration for coal 20%,base metals 20%, uranium 15%, gold 25%, andindustrial minerals 20%. Particular explora-tion projects then have to compete for funds.If a particular project is successful then it willattract additional spending. This distributionof funds is carried out on a regular basis andis usually coupled with a technical review ofexploration projects in which the geologist incharge of a project has to account for moneyspent and put forward a bid for further funding.

Careful control of funds is essential and usu-ally involves the nomination of budget holdersand authorization levels. For example, the

exploration manager may be able to authorizeexpenditure of $100,000 but a project geologistonly $2000, so all drilling accounts will be sentto head office, whereas the geologist will dealwith vehicle hire and field expenses. In thiscase written authorization from the explorationmanager would also be required before drillingstarts. Summary accounts will be kept in headoffice under a qualified accountant or book-keeper and will be subject to regular audit byexternal accountants. For most companiesglobal expenditure on exploration will be in-cluded in annual accounts and will be writtenoff against the year’s income.

Project basis

In order to contain costs and provide a basis forfuture budgeting, costs will be calculated foreach exploration project: firstly to monitorspending accurately, and secondly so that theexpenses can be written off against production,if exploration is successful, or if the project issold to another company. Budgets are normally

Page 74: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 57

TABLE 4.2 Example of a budget worksheet used to estimate the annual budget requirement for a project.

Project Name:Project No:

Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total

Salaries

Wages

Drilling

Transport

Office Lease

Administration

Analytical Costs

Field Expenses

Options – New

Options – Renewals

Surveys

Other Contracts

MONTHTOTALS

calculated on a yearly basis within whichprojects can be allocated funds on a monthlybasis. At the desk study stage the main costsare salaries of staff and these can be calculatedfrom a nominal staff cost per month, the aver-age of the salaries is usually multiplied by twoto cover pensions, housing costs, and secretarialsupport. Other support costs must be includedsuch as vehicles and helicopter transport inremote areas. Direct costs of exploration areeasier to estimate and include geochemistry,remote sensing, and geophysical costs. Thesecan be easily obtained by asking for quotationsfrom contractors. Perhaps a more contentiousmatter is the allocation of overhead costs toeach project; these cover management timeand will include the time of head office staffwho review projects for presentation to boardlevel management. Often these can be frighten-ingly large, e.g. the costs of keeping a seniorgeologist in a metropolitan center will be muchlarger than in a small town. Table 4.2 shows a

worksheet for the calculation of budget costsand Table 4.3 gives an example of a completedbudget and typical costs for reconnaissanceprograms.

4.2 DESK STUDIES AND RECONNAISSANCE

4.2.1 Desk studies

Once the exploration organization is in place,initial finance budgeted, and target deposittype selected, then desk studies can start andareas can be selected for reconnaissance – buton what basis?

The first stage in a totally new program is toacquire information about the areas selected.Besides background information on geology,data on the occurrence of currently producingmines and prospects and their economic statusare essential. This information will normally bebased largely on published material but could

Page 75: Introduction to mineral exploration

58 C.J. MOON & M.K.G. WHATELEY

less developed countries, in spite of the avail-ability of satellite information. Besides surfacemapping familiar to most geologists, a numberof other sources of information on regional geo-logy are widely used in exploration.

An invaluable addition to surface regionalgeology is the use of regional geophysics (seeChapter 7). Airborne magnetics, radiometrics,and regional gravity data are available for muchof the developed world and help in refininggeological interpretation and, particularly, inmapping deep structures. For example, a belt ofthe Superior Province in northern Manitobawhich hosts a number of major nickel deposits,including the large deposit at Thompson, canbe clearly followed under Palaeozoic cover tothe south (Fig. 4.4). Regional seismic data arehelpful but are usually only available if oilcompanies donate them to the public domain.Specially commissioned seismic surveys havegreatly helped in deciphering the subsurfacegeology of the Witwatersrand Basin (see sec-tion 14.5.4). Subsurface interpretations ofgeophysical information can be checked bylinking them with information from any avail-able deep drill hole logs.

Regional geochemical surveys (see section8.4) also provide much information in areas ofpoor outcrop and have defined major litho-logical provinces covered by boulder clay inFinland.

The sources of information for mineraloccurrence localities are similar to those forregional geology. Geological surveys usuallyhave the most comprehensive data base withina country and much of this is normallypublished (e.g. the summary of UK mineralpotential of Colman 2000). Many surveys havecollated all the mineral occurrences withintheir country and the results are available asmaps, reports, or even on computerized data-bases. Two useful types of maps are mineraloccurrence maps and metallogenic maps, manyof which are now available in digital format.The former type merely shows the location ofthe occurrences whereas the latter attempts toshow the form of the deposit and associatedelements overlain on background geology usingGIS (see section 9.2). Overlays of the mineralprospects with geology will provide clues to re-gional controls on mineralisation. At this pointsome economic input is required as it is often

TABLE 4.3 Examples of budget forecasting for alow cost project and a project requiring drilling.Land costs are excluded.

Project Cost in US$

Low cost project costTime: 6 days @ US$200 per day 1200Hotels and meals for 2 nights 350Transport (vehicle hire) 200Assays 200Office staff time 100TOTAL 2050

Medium cost project20 boreholes @ 200 m

each = 4000 m4000 m at US$50 per meter 200,000Overhead costs 200,000TOTAL 400,000

also include open-file material from geologicalsurveys and departments of mines, data fromcolleagues and from consultants with particu-lar expertise in the area concerned.

Background geological information is avail-able for most areas in the world although itsscale and quality vary considerably. In someparts of Europe geological maps are publishedat 1:50,000 and manuscript field sheets at1:10,000 are available. In less populated partsof the world, such as Canada or Australia, thebase cover is 1:250,000 with more detailedareas at 1:100,000. For the mineral explorationgeologist the published geological data willonly be a beginning and he or she will interpretthe geology using the geological featuresdefined in the deposit model. This is bestachieved by starting with a synoptic view ofthe geology from satellite imagery. Unless thearea is extensively vegetated this is likely to befrom Landsat or SPOT imagery (see Chapter 6).Landsat has the added advantage of highlight-ing areas of hydrothermal alteration within aridareas if processed correctly. In areas of densevegetation side-looking radar can be used. Forexample, most of Amazonian Brazil has beenmapped in this way. For smaller areas airphotography provides better resolution oftenat considerably less cost, although air photo-graphs are still regarded as top secret in some

Page 76: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 59

FIG. 4.4 Aeromagnetic map of western Canada showing the continuity of anomalies underneath sedimentarycover. The edge of the sedimentary cover is the bold dashed line striking NW–SE . The belt containing theThompson nickel deposit runs NNE from 102°E, 52°N. Hudson Bay forms the northeast corner of the map.(Reproduced with permission from Ross et al. 1991.)

the case that significant prospects have differ-ent controls from weakly mineralized occur-rences. The sort of economic information thatis useful are the grades and tonnages mined,recovery methods, and the reasons for stoppingmining. This can be obtained from journals orfrom company reports.

An example of a desk study is the recognitionof target areas for epithermal gold depositsin western Turkey. Here mineral explorationincreased significantly after 1985 due to thereform of Turkish mining laws and the abilityof non-Turkish mining companies to obtain amajority shareholding in any discovery. West-ern Turkey is of interest as its geological settingis similar to eastern Nevada (see section 16.4),with extension and graben formation in the

Miocene accompanied by extensive volcanismand much current hot spring activity. In addi-tion a cursory glance at the Metallogenic Mapof Europe shows that a number of gold occur-rences, as well as other elements, such asarsenic, antimony, and mercury which areoften associated with epithermal activity, arepresent (Fig. 4.5). The mineral map of Turkey(Erseçen 1989) provides further information onreserves at several of these localities. The com-piler must test this information against theepithermal model in which gold is likely to beassociated with graben bounding faults orvolcanics. Most of the gold occurrences shownin western Turkey are within high grade meta-morphic rocks and represent metamorphosedquartz veins of little interest. However the gold

Page 77: Introduction to mineral exploration

60 C.J. MOON & M.K.G. WHATELEY

FIG. 4.5 Simplified metallogenic map of Turkey. The original sheet of the Metallogenic Map of Europe showsdifferent deposit types and commodities in different colors as well as underlying geology. (From Lafitte 1970,Sheet 8.)

Istanbul

Fe

Mn

CuFe

Mn

Mn

sep

B

Mg

B

W

MoCrB

S

B

FeSb

Fe

FeCu

Al

Au

aluHg

Hg

SbAu

SbHg

Hg

Al

Al

Cr

CrMn

S

S Mg

AuAsAuAs Al asb

Cu

B Hg

Sb

Al

Al

Al

AuAs

FeFe

Fe

CrCrCr

CrCr

Cr

CrMn

CrCrCr

Cu

S

AuAs

AuAsFeAl

AuAg

SbB

Sb

PbCu

WW

Fe

FePbBa

ZnFeCu

PbFe

PbZnFe Sb

PbZnFe

Pb Cu

Fe

CuAu

CrCr

Mg

As

Mo

PbZnMn

FeSbW

Mg Cr

FeAs

AsMn

Cu

CrFe

PbZn

Ba

aluFe

PbAs

AuA

F

BC D E

Deposits associated withmedial stage volcanism

Deposits assosiated withlate orogenic volcanism

Hydrothermal deposits: Late orogenic

Residual deposits

Sepiolitesep

Alunitealu

Districts

Sedimentary deposits

Volcanic and volcano sedimentarydeposits or early magmatic

Large Small

Occurrences

VeinsStratiformdepositsStockworks

Size Morphology

Genetic TypeSurficial alterationVolcanogenicMesothermalMagmaticMetamorphic and metasomatic

G

0 100 km

37°

38°

39°

40°

41°

28° 29° 30° 31° 32°

Black Sea

Marmara Sea

BULGARIA

GREECE

Page 78: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 61

occurrences within volcanics to the south ofÇanakkale and to the northwest of Izmir(points A and B, Figs 4.5, 4.6) remain of interest,as do the mercury and antimony occurrencesalong the graben bounding faults (points C, D,E). The area south of Çanakkale was mined bya British company until the outbreak of the1914–18 World War and has lain nearly dor-mant since then. The first company to appreci-ate the potential of this area has defined aresource of 9 Mt at 1.25 g t−1 Au within one ofthe areas of brecciated volcanics and a numberof other targets have been identified in thevicinity. The area near Izmir has been minedsince Roman times but the gold mineralisationis largely confined to a vein structure. Recentexploration by the Turkish geological survey(MTA) who hold the license has defined afurther resource of 1.4 Mt at 1.3 g t−1 Au. Themercury and antimony deposits are largely theconcessions of the state mining group, Etibank,but at least one exploration company was ableto agree a joint venture and found significant,although subeconomic, gold grades at theEmirli Antimony Mine (point C, Figs 4.5, 4.6).The most significant gold discovery in the area,Ovacik (point F), is not shown on either map,although a number of veins in the area had, atleast in part, been mined by ancient workers.The Ovacik deposit was discovered during aregional reconnaissance of favorable volcanicareas using stream sediment geochemistry andprospecting. Larson (1989) provides a furtheroverview of the potential of the area includingmuch on logistics. Since the Ovacik discoveryexploration has been much reduced as thecompany (originally Eurogold, a joint venturebeteeen Australian and Euopean interests) wasunable to get permission to mine even thoughit had constructed all the facilities. The minewas, however, brought into production in 2001after local opposition had been overcome. Thelimited exploration in west Turkey since 1990has also identified a major porphyry sytem atKisladag (point G) which contains a reserveof 5M ounces of gold at 1.1 g t−1 Au (EldoradoGold 2003).

4.2.2 Reconnaissance

The aim of reconnaissance is to evaluate areasof interest highlighted in the desk study rapidly

and to generate other, previously unknown,targets preferably without taking out licenses.

In areas of reasonable outcrop the first stagewill be to check the geology by driving alongroads or from the air if ground access is impos-sible. A considerable number of disseminatedgold prospects have been found by helicopterfollow-up of Landsat data processed to high-light argillic alteration and siliceous caps inarid areas, particularly in Chile. Airborne geo-logy can be extremely effective in solvingmajor questions quickly, but it is expensive,US$500 per hour, and is best undertaken byexperienced geologists with specific problemsto solve. In any case a preliminary visit to thefield should be made as soon as practicable tocheck the accessibility of the area, examine onthe ground some of the data emerging from thedesk study, and check on competitor activity.

Reconnaissance techniques can be dividedinto those which enable a geologist to locatemineral prospects directly and those whichprovide background information to reduce thesearch area.

Airborne geophysics and stream or lake sedi-ment geochemistry are the principal tools fordirectly detecting mineralisation. Perhaps themost successful of these has been the use ofairborne electromagnetic techniques in thesearch for massive sulfide deposits within theglaciated areas of the Canadian Shield. This hasled to the discovery of a number of deposits,such as Kidd Creek (see section 15.3.2), bydrilling the airborne anomalies after very littleground work. However, each method has draw-backs. In the case of airborne electromagnetictechniques it is the inability to distinguishbase-metal-rich sulfides from pyrite and graph-ite and interferences from conductive over-burden. Airborne radiometrics have been verysuccessful in finding uranium deposits innonglaciated areas and led to the discovery ofdeposits such as Ranger 1 and Yeelirrie inAustralia (Dunn et al. 1990a,b). In glaciatedareas radiometric surveys have found a numberof boulder trains which have led indirectly todeposits. In areas of residual overburden andactive weathering stream sediment samplinghas directly located a number of deposits, suchas Bougainville in Papua New Guinea, but itis more likely to highlight areas for groundfollow-up and licensing.

Page 79: Introduction to mineral exploration

62 C.J. MOON & M.K.G. WHATELEY

FIG. 4.6 Part of the Mineral Map of Turkey. (Redrawn from Erseçen 1989.)

●● ●

● ●

1

3

2

2

2

2 2

●● ●

● ●

GypsumAluniteAsbestosBariteBentoniteBoron mineralsCement raw materialsDistheneDiatomiteDolomiteFeldsparFluorsparGraphiteKaoliniteClayLimestone

Industrial raw materials

Emery

Copper + Lead + Zinc1

2

3

Metallic mineralsGoldGold + SilverAluminum (Diasporite)AntimonyCopper + Pyrite

Copper + ZincMercuryIronChromiteLeadLead + ZincManganeseNickelTitaniumTungsten (Wolfram)

Town

Antalya

Eskesehir

Izmir

Manisa

Balikesir

Çanakkale

Istanbul

F

A

Energy raw materialsBituminous schistLignitePetroleumNatural vaporUranium

Industrial sandsSulfurMagnesiteMarbleMicaPerliteSepiolite (Meerschaum)TalcTravertineBrick-tile raw materials

B

C D E

G

0 100 km

37°

38°

39°

40°

41°

28° 29° 30° 31° 32°

Black Sea

Marmara Sea

BULGARIA

GREECE

Page 80: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 63

Present limits of Karoo Basin in South Africa

Outcrop of Beaufort Group,Molteno –Drakensberg volcanics

Uranium province in main Karoo basin

Palaeocurrents (Theron 1973)

Johannesburg

SpringbokFlats

Qwa Qwa

Ficksburg

Edenburg

Richmond

Graff ReinetBeaufort WestSutherland

Fraserburg

Laingsburg

Cape Town

SOUTHAFRICA

Cape Town

Johannesburg

● ●

FIG. 4.7 The Karoo Basin in South Africa. (From Moon & Whateley 1989.)

An example of an exploration strategy canbe seen in the search for sandstone-hosted,uranium deposits in the Karoo Basin of SouthAfrica (Fig. 4.7) during the 1970s and early1980s. The initial reconnaissance was con-ducted by Union Carbide in the late 1960s.They became interested in the basin because ofthe similarity of the Karoo sediments to thoseof the Uravan area in Colorado, where the com-pany had uranium mines, and because of thediscovery of uranium anomalies in Karoo sedi-ments in Botswana and Zimbabwe. As little ofthe basin, which is about 600,000 km2, hadbeen mapped, the company chose a carborne

radiometric survey as its principal technique.In this technique a scintillometer is carried in avehicle and gamma radiation for the area sur-rounding the road is measured. Although a verylimited area around the road is measured it hasthe advantages that the driver can note thegeology through which he is passing, can stopimmediately if an anomaly is detected, and it ischeap. In this case a driver was chosen withexperience in Uravan geology and he was ableto use that experience to compare the geologytraversed with favorable rocks in the westernUSA. After a good deal of traversing a signific-ant anomaly (Fig. 4.8) within sandstones was

Page 81: Introduction to mineral exploration

64 C.J. MOON & M.K.G. WHATELEY

100

80

60

40

Cou

nts

Paardefontein Rietkuil

0 10 20 30 40 50

Distance (km)

FIG. 4.8 Trace of a carborne radiometric survey similar to that which discovered the first major radiometricanomaly in the Karoo Basin. (After Moon 1973.)

discovered to the west of the small town ofBeaufort West (Fig. 4.7). Once the initial dis-covery was made the higher cost reconnais-sance technique of fixed wing airborneradiometrics was used as this provided muchfuller coverage of the area, about 30% at 30 mflying height. This immediately indicated toUnion Carbide a number of areas for follow upwork and drilling.

It became rapidly apparent that the depositsare similar to those of the Uravan area. In boththe uranium is in the fluviatile sandstones inan alternating sequence of sandstone, siltstone,and mudstone; in the South African case, in theUpper Permian Beaufort Group of the KarooSupergroup (Turner 1985). These sedimentswere deposited by braided to meanderingstreams which resulted in interfingering andoverlapping of sandstone–mudstone transi-tions and made stratigraphical correlation diffi-cult. Significant mineralisation is confined tothicker multistorey sandstones, usually >14 mthick, and is epigenetic (Gableman & Conel1985). The source of the uranium is unknown,possibly it was leached from intrabasinal vol-canic detritus but certainly mobilized as uranylcarbonate complexes by oxygen-rich, alkalineground waters which then migrated into theporous sandstone (Turner 1985, Sabins 1987).The migrating ground water then encounteredreducing conditions in the presence of organicmaterial, bacterially generated hydrogensulfide, and pyrite. This change from oxic toanoxic conditions caused the uranium to pre-cipitate as minerals (coffinite and uraninite)

that coated sand grains and filled pore spaces.Diagenesis rendered the deposits impermeableto further remobilization. The outcrop charac-teristics of the deposits are either calcareous,manganese-rich sandstone, or bleached altera-tion stained with limonite and hematite, thatlocally impart yellow, red, and brown colora-tion. Typical sizes of deposits, which occur as anumber of pods, are 500 by 100 m. Averagegrades are about 0.1% U3O8 and thicknessesvary from 0.1 to 6 m but are typically 1 m.

From the airborne radiometrics and follow-up drilling Union Carbide defined one area ofpotential economic interest and inevitablyword of the discovery leaked out. In additionthe price of uranium rose following the 1973Arab–Israeli war leading to interest on the partof a large number of other companies. Thereconnaissance approach of competitor com-panies was however markedly different in thatthey knew that the basin hosted significantmineralisation and that Union Carbide hadimportant areas under option to purchase. Ingeneral these companies selected at least onearea around the known Union Carbide dis-covery and a number of areas elsewhere withinthe basin. These were chosen on the basis ofcomparison with deposits in the western USAwhere the deposits are preferentially hostedat specific horizons and within thicker fluvialor lacustrine sandstones. Mosaics of the, then(1973) newly available, Landsat MSS images,as well as government-flown air photography,were used to derive a regional stratigraphyand to highlight thicker sandstone packages.

Page 82: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 65

Landsat imagery offered a hitherto unavailablesynoptic view and the ability to use the spectralcharacteristics of the different bands to dis-criminate alteration and rock types. On falsecolor images (see sections 6.2.3 & 6.2.6) sand-stone and mudstone rocks could be differenti-ated, but it was not possible to distinguish thealtered sandstone from sandstone that had anatural desert varnish of limonite. To enhancethe appearance of the altered rocks, color ratioimages were produced by combining the ratiosof bands 4/7, 6/4, and 7/4 in blue, green, and redrespectively. The altered sandstones were iden-tified by light yellow patches, which were usedby field geologists as targets for ground follow-up surveys. This was less successful than flyingradiometric surveys over areas of favorablesandstones. In more rugged areas helicopterswere used to trace the sandstones and check forradiometric anomalies (Fig. 4.9).

Although airborne radiometric surveys wereextremely successful in areas of good outcrop,they were ineffective in locating deposits con-cealed under inclined sedimentary rock cover.The most successful exploration methods inthe search for subsurface mineralisation weredetailed surface mapping coupled with hand-held surface radiometric surveys and down thehole logging (see section 7.13) of all availableboreholes drilled for farm water supplies (Moon& Whateley 1989).

4.2.3 Land acquisition

The term land is deliberately used as the actuallegal requirements for exploration and miningvaries from country to country. What the ex-plorer needs to acquire is the right (preferablyexclusive) to explore and to mine a deposit ifthe exploration is successful. Normally a com-pany will obtain the right to explore the pro-perty for a particular period of time and theoption to convert this into a right to mine, ifdesired, in return for an annual payment andin some cases the agreement to expend a min-imum amount on exploration and to report allresults. Unfortunately most legal systems areextremely complicated and the explorer maynot be able to obtain the exact right that heor she requires, for example gold or energyminerals may be excluded and the right to thesurface of the land (surface rights) may be sep-arate from the right to mine (mineral rights).Two end-member legal systems can be distin-guished as far as mineral rights are concerned:the first in which all mineral rights are ownedby the state and the explorer can mine with noregard to the current occupier of the surfacerights, and the second in which all mineral andsurface rights are privately owned. The firstnormally results from governmental decree orrevolution whereas the second is typical ofmany former British colonies.

FIG. 4.9 View from a helicoptertraversing in the central KarooBasin, northwest of Beaufort West.

Page 83: Introduction to mineral exploration

66 C.J. MOON & M.K.G. WHATELEY

Private model

In the private situation, for example in Britain,the company’s lawyer will negotiate with aprivate mineral rights owner and obtain anexploration or option agreement under whichthe company will be able to explore for a min-imum period, normally 3 years, and then beable to renew the option; or to buy the mineralrights for a fixed sum, normally in excess of thefree market value. In exchange, the mineralrights owner will receive a fixed annual sumoption payment and compensation for anydamage to the surface if he or she owns thesurface rights. If the surface rights are separ-ately owned then an agreement must be madewith that owner. In the case of Britain therights to gold are owned by the Crown andmust be covered in a further separate agree-ment. There is no legal obligation to report theresults to government although summarydrillhole results must be reported to the BritishGeological Survey. Most physical explorationof any significance and drilling of greater than28 days duration requires consent from thelocal planning authority.

State model

In the case of state ownership the state willnormally own the mineral rights and be able togrant access to the surface. In this case applica-tion for an exploration license will usually bemade to the department of mines. The size ofthe exploration area may be fixed. For examplethe law in Western Australia limits a 2-yearprospecting license to 200 ha although a 5-yearexploration license can cover between 10 and200 km2. Annual work commitments on theformer are $A40 ha−1 and $A300 km−2 on thelatter. Mining leases are granted for a renew-able period of 21 years with a maximum area of1000 ha. Normally in areas with state owner-ship a full report of exploration results must befiled to the mines department every year and atthe termination of the lease. Such reports oftenprovide information for future exploration aswell as data for government decision making.

New mineral laws

An example of a new mineral law designedto encourage exploration is the 1985 law of

Turkey. Under this exploration licenses aregranted for 30 months; these can be convertedinto pre-operation licenses for 3 years and theninto an operating or mining license. The cost ofan exploration license in 1992 was 4000 Turk-ish Lira (US$0.10) ha−1, which is refundablewhen the license is relinquished. The rights ofsmall miners are protected by a right of denun-ciation under which any Turkish citizen whocan demonstrate a previous discovery in thearea can claim 3% of the gross profit. Thewhole operation is policed by a unit in the geo-logical survey, which uses a computerizedsystem to monitor license areas.

Problem countries

In some countries the rule of law is less secure.For example, following the collapse of theSoviet Union it was not clear who was respon-sible for, or owned, mineral rights in the newlycreated countries or in Russia. Deposits hadbeen explored by state-financed organizations,which were left without funds and effectivelyprivatized. Although these organizations werekeen to sell rights to the deposits it was byno means certain that national governmentsrecognized their title. Even when title could beagreed, some governments tore up agreements,without compensation, when they realised thevalue of the deposits. An example of this con-fusion was the Grib diamond pipe in northernRussia, which was discovered by a junior northAmerican company in 1996 and worth morethan $5 billion in situ (see section 17.1.6). Thejunior was in a 40% joint venture with aRussian expedition which held the license. Sub-sequent to the discovery the junior companywas unable to get the license transferred to anew joint venture company between them andthe Russian expedition as had been previouslyunderstood. This was probably partly becausethe assets of the expedition had in the meantime been taken over by a major Russian oilcompany and a Russian entrepreneur. A fewjoint ventures have been successful under theseconditions, notably those in which govern-ments have a significant stake. A company 67%owned by the government of Kyrgyzstan and33% by Cameco Corporation of Canada (nowCenterra Gold) commissioned, and operates,the large Kumtor gold mine in central Asia.

Page 84: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 67

Legal advice is a virtual necessity in landacquisition and most exploration groups havea lawyer or land person on the staff or on aretainer. If legal tenure is insecure then allexploration effort may be wasted and allrevenue from the property go elsewhere. Thelawyer will also be responsible for checkingthe laws that need to be observed, e.g. informa-tion from drillholes may need to be reportedto government. Corporate administrationshould be checked so that there is no insiderdealing or any activities that could be con-sidered unethical. A famous recent casewas that in which Lac Minerals lost controlof part of the large and profitable Hemlodeposit in Ontario. It was alleged that LacMinerals had bought claims from the widowof a prospector after receiving confidentialinformation from another company, CoronaCorp., and in spite of a verbal agreement notto stake that ground. Besides losing controlof the deposit, Lac was faced with a large legalbill.

A more recent example was the dispute be-tween Western Mining Corporation and SavageExploration Pty Ltd over control of the ErnestHenry deposit in northern Queensland (West-ern Mining 1993). Leases in the area, includinga tenement named ML2671, had originallybeen pegged in 1974 by Savage Exploration foriron ore on the basis of the results of a govern-ment aeromagnetic survey. In 1989 WesternMining decided to explore the area for basemetals and obtained information in the publicdomain including the government airbornegeophysics. When Western Mining selectedareas for further work they found that much ofthe land was held by Hunter Resources Ltd andtherefore entered into a joint venture withHunter but operated by Western Mining.

One of the areas targeted for further workby the joint venture was in the general area ofML2671 as described in the Mines Departmentfiles. A baseline was pegged in July 1990 at100 m intervals and crossed the described posi-tion of ML2671 although no pegs were seen onthe ground to indicate the position of the lease.A magnetic survey was undertaken includingreadings within the described area of ML2671and an initial TEM survey in August 1990.Savage Exploration was approached in March1991 and agreement made over the terms of an

option by which the Western Mining–Hunterjoint venture could acquire a number of leases,including ML2671. In May 1991 further TEMsurveys were made and a formal option signedin October 1991. In late October 1991 the firsthole was drilled and this intersected strongmineralisation.

After further drilling the discovery of theErnest Henry deposit was announced and Sav-age Exploration were advised in June 1992 thatthe joint venture wished to exercise its optionand buy the lease of ML2671. Savage Explora-tion then commenced court proceedings alleg-ing Western Mining had misrepresented thesituation when agreeing an option on ML2671and that Western Mining had trespassed onML2671 during the ground surveys. In theensuing court case it transpired that the pegsfor ML2671 were not in the place indicated onthe Mines Dept files and were 850 m north ofthe stated location and the lease was rotatedseveral degrees anticlockwise from its plottedposition. The case was settled out of court withWestern Mining agreeing to give up any claimto ML2671 and paying Hunter $A17 millionand certain legal costs.

4.3 SUSTAINABLE DEVELOPMENT

One of the major changes in exploration inthe last 10 years has been the difficultyin obtaining a permit to develop a new mine.As discussed in section 1.5.2 much researchhas been undertaken on the socioeconomicaspects of opening and operating a new mine(see section 11.2.7). The major componentsof sustainable development in addition to thetechnical and economic components are: (i)environmental; (ii) social; (iii) governmental.Most major companies issue annual reports ontheir progress in regard to these three aspectsof their operations and these are available ontheir websites. These form a good place todelve more deeply into specific examples andpolicies in different countries.

Environmental aspects

Obtaining an environmental permit to operatea mine has become a vital part of the feasibilityprocess but also needs to be addressed during

Page 85: Introduction to mineral exploration

68 C.J. MOON & M.K.G. WHATELEY

Diavik mines (see section 17.2) and modernzinc mines in Ireland are good examples ofwhat can, and should be, achieved.

Social aspects

Another aspect of environmental studies ispublic relations, particularly keeping the localpopulation informed of progress and obtainingtheir active approval for any development pro-ject. The past few years are littered with ex-amples of projects that are technically excellentbut have failed to obtain permission for devel-opment, and others that have been significantlydelayed causing them to become economicallyunviable. It is at the exploration stage that thelocal population form impressions of the natureof the exploration group and whether they wishto be involved in its activities. Establishinggood relation and communications with thelocal community is the first step in gainingtheir backing for future mining – “a sociallicense to operate” (MMSD 2002). A good sum-mary of the problems is freely available fromthe PDAC website (E3 2004). Initial concerns ofthe local community are the transient natureof exploration and the lack of knowledge of thelocal population to the techniques used andtheir scale. In addition, the local populationmay have economic expectations and the arrivalof exploration from another part of, or fromoutside the, country may cause cultural stress.It is always advisable to obtain local advice. Forexample, in Australia, sacred Aboriginal siteswill be known to the local population but notto most geologists. Their unintentional des-ecration has caused the development of intenseopposition to further exploration.

The initial contact with the local populationshould be carefully planned and, if possible, beenabled by an intermediary, such as a local offi-cial trusted by both parties and probably afterconsulting someone with well-developed skillsin dealing with local government and commu-nity leaders in the area. The process of explora-tion and possible outcomes should be carefullyexplained so that unrealistic expectations arenot raised. Local labor and purchasing shouldbe used wherever possible and training shouldbe provided. Major mining companies are nowaware of the problems and provide training

the exploration phases. Obtaining a permitinvolves the preparation of an environmentalimpact statement (EIS) describing the problemsthat mining will cause and the rehabilitationprogram that will be followed once mining iscomplete (Hinde 1993). Such an impact state-ment requires that the condition of the envi-ronment in the potential mining area beforedevelopment began is recorded (a baselinesurvey). Thus it is essential to collect data dur-ing the exploration stage for use in these EISs.Initial data might include surface descriptionsand photographs, and geochemical analysesindicating background levels of metals andacidity as well as water levels and flows. It is ofcourse essential to minimize damage duringexploration and to set a high standard for envi-ronmental management during any exploita-tion. Trenches and pits should be filled andany damage by tracked vehicles should beminimized and if possible made good. It maybe that a more expensive method of access, e.g.helicopter-supported drilling, may be neces-sary to minimize impact during exploration.Checklists for various exploration activitiesand discussion of best practices are available atthe E3 website (E3 2004). Data collection andbaseline surveys become more intense as aprospect becomes more advanced and the pros-pect of an EIS looms.

One example of a major environmental prob-lem in mining that has had impact in explora-tion has been the use of cyanide in goldextraction. This has been controversial particu-larly in Europe where cyanide extraction hasbeen prohibited in the Czech Republic since2000, causing the abandonment of goldexploration including the advanced prospectat Kaperske Hory. In other areas, the use ofcyanidation will be restricted, particularlyafter two spills, in 1998 in Kyrgyzstan from adelivery truck and in 2000 in Romania from atailings dam. Strong objections to cyanidationwere raised at Ovacik in Turkey and this wasone of the reasons for the long delay in com-missioning that mine after development(section 4.2.1).

Although mining has had a poor record in en-vironmental impact, modern mines are capableof being designed to minimize these impacts.The very limited impact of the Ekati and

Page 86: Introduction to mineral exploration

4: RECONNAISSANCE EXPLORATION 69

for their field geologists, emphasizing theimportance of contacts with local communit-ies in the exploration process.

Governmental aspects

Relations with governments can be problem-atic particularly with major projects that willgenerate a large part of a developing country’sexport earnings. If large amounts of money areinvolved in countries where public servants arepoorly paid there is a tendency for significantamounts to be appropriated, either in the formof taxes that do not reach the governmenttreasury or bribes. Governmental relationsneed to be handled carefully at a senior leveland care should be taken that they are notin conflict with relations at a community level(see sections 11.2.6 & 11.2.8). There are anumber of recent (2004) examples, such as theEsquel gold deposit in Argentina, where thenational government is keen on the develop-ment but local or provincial bodies are opposed.

4.3.1 Health and safety

A key aspect of a company’s reputation, bothas an employer and with the local community,is the health and safety of its staff. Althoughproblems at the exploration stage are less severethan during mining, serious injuries and deathshave occurred. Assessments of hazards shouldbe made so that high risk activities can be rec-ognized and mitigated. First aid training shouldalso be provided. At least one major mininggroup has linked staff pay to safety record andclaim that this is the only way to effectivelyimprove their safety record. All contractorsshould behave in a similar way to companystaff and safety record should be a significantfactor in choosing contractors. Probably themajor source of serious accidents in the au-thors’ experiences is road transport, especiallyin remote areas. Staff should be provided withtraining in driving on poor road surfaces.

4.4 SUMMARY

Mineral exploration is conducted by both theprivate and public sectors in both market and

most centrally planned economies. Groupsfrom both sectors must be clear about whatcommodity and type of deposit they areseeking before setting up an exploration organ-ization. Mineral exploration is a long-termcommitment and there must be careful plan-ning of the participant’s long-term objectivesto ensure viable budgeting. With the objectivedecided upon an exploration organization mustbe set up and its success will be enhanced bydeveloping all the factors listed in section 4.1.1.

Budgets must be carefully evaluated and notbe figures drawn out of the blue. An under-funded project will in most cases be a failure.Careful control of funds is essential and it pro-vides a basis for future budgeting.

With the organization in place and the targetdeposit type selected, desk studies can start inearnest and areas for reconnaissance be chosen.Relevant information must be acquired, as-sessed, and selected using published works andopen-file material from government institu-tions (section 4.2.1). The aim of reconnaissanceis to evaluate rapidly areas highlighted in thedesk study and to identify targets for follow-upwork and drilling. If this is successful then theexclusive rights to explore and to mine any de-posits found in the target area must be acquired(section 4.2.3).

4.5 FURTHER READING

Literature on the design and execution ofexploration programs is sparse. Peters (1987)provides a good introduction. Management ofMineral Exploration by White (1997) is wellworth reading with a wealth of practical experi-ence. Papers edited by Huchinson and Grauch(1991) discuss the evolution of genetic conceptsas well as giving some case histories of majordiscoveries and are updated in the volume ofGoldfarb and Nielsen (2002). The two volumesof Sillitoe (1995, 2000) provide succinct casehistories of discoveries in the Pacific Rim area.

An enthusiastic view from industry, tem-pered by economic reality, is provided byWoodall (1984, 1992). Much data on theeconomics of exploration can be found in thevolume edited by Tilton et al. (1988), updatedin Crowson (2003).

Page 87: Introduction to mineral exploration

70 C.J. MOON & M.K.G. WHATELEY

5From Prospect to

PrefeasibilityCharles J. Moon and Michael K.G. Whateley

Once land has been acquired the geologist mustdirect his or her efforts to proving whether ornot a mineral prospect is worthy of commercialevaluation. Proving a discovery to be of suffici-ent size and quality inevitably involves a sub-surface investigation and the geologist usuallyfaces the task of generating a target for drilling.In exceptional cases, such as very shallow min-eralisation, a resource may be proved by dig-ging pits or trenches or, in mountainous, areasby driving adits into the mountainside.

Whatever the method used the key require-ment is to explore the area at the lowest costwithout missing significant targets. Fulfillingthis requirement is not easy and the mineraldeposit models, such as those discussed inChapter 3, must be modified to include eco-nomic considerations. There should be a clearidea of the size of the deposit sought, the max-imum depth of interest, and whether under-ground mining is acceptable. Finding a drillingtarget normally involves the commissioningof a number of different surveys, such as ageophysical survey, to locate the target andindicate its probable subsurface extension. Therole of the geologist and the exploration man-agement is to decide which are necessary andto integrate the surveys to maximum effect.

5.1 FINDING A DRILLING TARGET

5.1.1 Organization and budgeting

Once land has been acquired, an organizationand budget must be set up to explore it andbring the exploration to a successful conclu-

sion. The scale and speed of exploration willdepend on the land acquisition agreements (seesection 4.2.3) and the overall budget. If a pur-chase decision is required in 3 years, then thebudget and organization must be geared to this.Usually the exploration of a particular pieceof land is given project status and allocated aseparate budget under the responsibility of ageologist, normally called the project geologist.This geologist is then allocated a support teamand he or she proceeds to plan the various sur-veys, usually in collaboration with in-houseexperts, and reports his or her recommenda-tions to the exploration management. Typic-ally, reporting of progress is carried out inmonthly reports by all staff and in 6-monthlyreviews of progress with senior explorationmanagement. These reviews are often also oralpresentations (“show and tell”) sessions andlinked to budget proposals for the next fin-ancial period.

Reconnaissance projects are normally dir-ected from an existing exploration office, butonce a project has been established serious con-sideration should be given (in inhabited areas)to setting up an office nearer the project loca-tion. Initially this may be an abandoned farm-house or caravan but for large projects it will bea formal office in the nearest town with goodcommunications and supplies. A small officeof this type is becoming much easier to organ-ize following the improvement of communica-tion and computing facilities during the lastdecade. Exploration data can be transferred bymodem, fax, and satellite communications toeven the remotest location. If the project growsinto one with a major drilling commitment

Page 88: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 71

then it is probable that married staff will workmore effectively if their families are moved tothis town. In general, the town should be lessthan 1 hour’s commuting time so that a visitto a drilling rig is not a chore. In the remotestareas staff will be housed in field camps andwill commute by air to their home base. If pos-sible exploration staff should not be expectedto stay for long periods in field camps as this isbad for morale and efficiency declines.

Budgeting for the project is more detailedthan at the reconnaissance stage and will takeaccount of the more expensive aspects of ex-ploration, notably drilling. A typical budgetsheet is shown in Table 4.2. Usually the majorexpenditure will be on labor and on the varioussurveys. If the area is remote then transportcosts, particularly helicopter charter, can be-come significant. Labor costs are normally cal-culated on the basis of man months allocatedand should include an overhead component tocover office rental, secretarial and drafting sup-port. Commonly overheads equal salary costs.For geophysical, and sometimes geochemicalsurveys, contractors are often hired as theircosts can be accurately estimated and compa-nies are not then faced with the possibility ofhaving to generate work for staff. Estimates arenormally made on the basis of cost per line-kilometer or sample. In remote areas carefulconsideration should be given to contractoravailability. For example in northern Canadathe field seasons are often very short and anumber of companies will be competing forcontractors. For other types of survey and fordrilling, estimates can be obtained from reput-able operators.

5.1.2 Topographical surveys

The accurate location of exploration surveysrelative to each other is of crucial importanceand requires that the explorationist knows thebasics of topographical surveying. The effortput into the survey varies with the successand importance of the project. At an early stagein a remote area a rough survey with the ac-curacy of a few meters will be adequate. Thiscan be achieved using aerial photographs andhandheld Global Positioning Satellite (GPS) re-ceivers (Ritchie et al. 1977, Sabins 1987) whereonly a few survey points are needed. For a major

drilling program surveying to a few millimeterswill be required for accurate borehole location.

The usual practice when starting work ona prospect is to define a local grid for the pro-spect using GPS to relate the local grid to thenational or international grid systems (see sec-tion 9.1.7). Some convenient point such as awind pump or large rock is normally taken asthe origin. The orientation of the grid will beparallel to the regional strike if steeply dippingmineralisation is suspected (from geophysics)but otherwise should be N–S or E–W.

Surveying methods

Geologists generally use simple and cheaptechniques in contrast to those used by profes-sional surveyors. Surveying has recently beentransformed by the advent of GPS, based ona network of satellites installed by the USDepartment of Defense. These enable a fix ofapproximately 5 m accuracy to be made any-where on earth, with an inexpensive (currentlyapproximately $US200) receiver, where threesatellites can be viewed. The major problemsare in forested areas.

One of the simplest and most widely usedtechniques, before the advent of GPS, is thetape and compass survey. This type of surveystarts from a fixed point with directions meas-ured with a prismatic or Brunton compassand distances measured with a tape or chain.Closed traverses, i.e. traverses returning to theinitial point, are often used to minimize theerrors in this method. Errors of distance andorientation may be distributed through thetraverse (section 5.1.4). Grids are often laid outusing this technique with baselines measuredalong a compass bearing. Distances are bestmeasured with a chain which is less vulnerableto wear and more accurate than a tape. Longerbaselines in flat country can be measured usinga bicycle wheel with a cyclometer attached.Lightweight hip-mounted chains are com-monly used in remote areas. Straight grid linesare usually best laid out on flat ground by backand forward sighting along lines of pegs orsticks. Sturdy wood or metal pegs should beused, the grid locations marked with metallictags and flagged with colored tape. The tapeshould be animal proof; goats have a particularfondness for colored tape.

Page 89: Introduction to mineral exploration

72 C.J. MOON & M.K.G. WHATELEY

Leveling is the most accurate method ofobtaining height differences between stationsand is used for example to obtain the eleva-tion of each station when undertaking grav-ity surveys. This method measures the heightdifference between a pair of stations using asurveyor’s level. Leveling is required in anunderground mine to determine the minimumslope required to drain an adit or drive and in asurface mine the maximum gradient up whichload, haul, dump (LHD) trucks can climb whenfully laden (usually <10%).

More exact ground surveys use a theodoliteas a substitute for a compass. It is simpleenough for geologists to learn to use. Numer-ical triangulation is carried out using anglesmeasured by theodolite to calculate x and ycoordinates. By also recording the verticalangles, the heights of points can be computed(Ritchie et al. 1977). Professional surveyors arereadily available in most parts of the world andthey should be contracted for more exact sur-veys. They will use a theodolite and electronicdistance measuring (EDM) equipment, oftencombined in one instrument, and can producean immediate printout of the grid location ofpoints measured.

5.1.3 Geological mapping

One of the key elements during the explorationof a prospect is the preparation of a geologicalmap. Its quality and scale will vary with theimportance of the program and the financeavailable. Initial investigation of a prospectmay only require sketch mapping on an aerialphotograph, whereas detailed investigationsprior to drilling may necessitate mapping everyexposure. Mapping at the prospect scale is gen-erally undertaken at 1:10,000 to 1:2500. Fordetailed, accurate mapping a telescopic alidadeand plane table or differential GPS may beused. The principle of the alidade is the same asfor a theodolite, except that the vertical andhorizontal distances between each point arecalculated in the field. The base of the alidadeis used to plot the position of the next point onthe waterproof drafting film covering the planetable.

The process of geological mapping of mineralprospects is similar to that of general geolo-gical mapping, but is more focused, and is well

described in the book by Majoribanks (1997).Although the regional environment of the pro-spect is important, particular attention will fallon known mineralisation or any discoveredduring the survey. The geological relationshipof the mineralisation must be assessed indetail, in particular whether it has any of thefeatures of the geological model sought. Forexample, if the target is a volcanic-associatedmassive sulfide deposit then any sulfidesshould be carefully mapped to determine ifthey are concordant or cross cutting. If thesulfides are cross cutting it should be estab-lished whether the sulfides are in a stockworkor a vein. Particular attention should be paidto mapping hydrothermal alteration, which isdescribed in detail by Pirajno (1992).

Detailed guidance on geological mapping isbeyond the scope of this book but can be foundin a series of handbooks published by the Geo-logical Society of London (Fry 1991, McClay1991, Thorpe & Brown 1993, Tucker 2003), par-ticularly in the summary volume of Barnes andLisle (2003).

One of the key elements of mapping is itsfinal presentation. Conventionally this was inthe form of a map drafted by Indian ink pen onto transparent film. From film, multiple copieseither on film or on paper can be made usingthe dyeline process and the map can be overlaidon other maps of the same scale, allowing easycomparison of features and selection of targets.

The conventional pen and paper approachhas been superseded by computerized draftingthat allows the storage of information in digitalform. Computer packages, such as AUTOCAD,a computer-aided drafting package, are widelyused in industry (see section 9.2). Maps andplans can be produced to scale and differentfeatures of the overall data set, held on differ-ent layers in the computer, can be selected forviewing on the computer screen or printed asa hard copy. The data can also be transmittedto more sophisticated Geographical Informa-tion Systems (GIS), for example ArcGIS orMapInfo, that allow the inquisition of data (seesection 9.2).

5.1.4 Mapping and sampling of old mines

Many prospects contain or are based aroundold mines. They may become attractive as

Page 90: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 73

exploration targets because of rising commod-ity prices, cheaper mining and processing costs,the development of new technology whichmay improve recovery, or the development ofa new geological model which could lead toundiscovered mineralisation. The presence of amining district indicates mineral potential,which must reduce the exploration risk. How-ever, there will be a premium to pay, as theproperty will probably already be under optionto, or owned by, a rival company.

The type of examination warranted by anold mine will depend on its antiquity, size, andknown history. In Europe and west Asia oldmines may be over 2000 years in age and be theresult of Roman or earlier activity. In this casethere are few, if any, records and the targetcommodity can only be guessed at. In suchcases small areas of disturbed ground, indicat-ing the presence of old prospecting pits ortrenches, can best be found from aerial photo-graphy. Field checking and grab sampling willconfirm the presence and indicate the possibletype of mineralisation.

Nineteenth or twentieth century mines arelikely to be larger and have more extensiverecords. Available records should be obtainedbut should be treated with some caution, asmany reports are unreliable and plans likelyto be incomplete. An aim of this type of inves-tigation is to check any records carefully byusing systematic underground sampling abovethe water level, as old mines are frequentlyflooded. Evaluation of extensive undergroundworkings requires considerable planning andwill be more expensive than surface explora-tion because equipment and labor for develop-ment and securing old underground workingsare costly. A key consideration is safety andaccess to the old workings must be made safebefore any sampling program is established. Itmay be necessary to undertake trenching andpitting in areas adjacent to the old workings,and eventually drilling may be used to examinethe deeper parts of the inaccessible mineralisa-tion. Guidelines on safe working practices inold workings can be found in Peters (1987) andBerkman (2001), and in the UK in publicationsof the National Association of Mining HistoryOrganizations (NAMHO 1985) and the Institu-tion of Geologists (now the Geological Societyof London) (IG 1989).

Before a sampling program can be put intooperation, a map of the old workings will berequired, if none is available from archives. Theexploration geologist is often the first personat the site and it is up to him or her to producea plan and section of the old workings. Thiswould be done using a tape and compass survey(Ritchie et al. 1977, Reedman 1979, Peters1987, Majoribanks 1997) (Table 5.1, Figs 5.1 &5.2). Once the layout of the old workingsis known, the mapping and complementarysampling program can begin. The survey pegsestablished during the surveying will be usedto locate the sample points and guide map-ping. With the tape held between the pegs, thesample points are marked on the drive or cross-cut walls and the distance from one peg tothe sample point recorded in the field notebook along with the sample number. The samenumber is written on a sample ticket andincluded with the sample in the sample bag.The samples are normally collected at regularintervals from channels cut normal to the dipof the mineralized rock (Fig. 5.3). The sampleinterval varies depending upon the type of min-eralisation. A vein gold deposit may well besampled at 1 m intervals along every drive,while a copper deposit may only be sampledevery 5 or 10 m.

If old records are available and reliable, thentheir data should be evaluated in conjunctionwith new sample data acquired during theremapping and resampling exercises. Geolog-ical controls on mineralisation should be estab-lished using isopach and structure contourmaps as discussed in section 5.2.2. It may benecessary to apply a cut-off value below whichthe mineralisation is not considered mineable.In Table 5.2 two cut-off parameters (Lane 1988)are used, in one a direct cut-off and in thesecond a weighted average value is used.

In the first case the upper sample cut-off istaken where the individual grade falls below1.5%. Some assays within this sample are alsobelow 1.5%, but they are surrounded by highervalues, which, when averaged out locally, havea mean greater than 1.5%. The samples be-tween the two lines are then averaged using thesample thickness as the weighting function,giving an average of 2.26% Zn over 2.10 m.In the second case, the samples are averagedfrom the base upwards using thickness as the

Page 91: Introduction to mineral exploration

74 C.J. MOON & M.K.G. WHATELEY

TA

BLE

5.1

An

exa

mpl

e of

bal

anci

ng

a cl

osed

tra

vers

e. (M

odifi

ed fr

om R

eedm

an 1

979.

)

Line

Bea

ring

Qua

dran

tD

ista

nce

∆E∆N

Una

djus

ted

coor

dsFi

nal c

oord

s

Ø(D

) m±s

inØ

±cos

رD

.sin

رD

.cos

ØE

NE

N

AB

561

82+0

.829

0+0

.559

2+6

7.98

11+4

5.85

3810

,000

10,0

0010

,000

10,0

00A

BC

331

474

−0.4

848

+0.8

746

−35.

8759

+64.

7219

10,0

67.9

810

,045

.76

10,0

69.7

610

,045

.53

BC

D28

24

91−0

.978

1+0

.207

9−8

9.01

14+1

8.92

0010

,032

.10

10,1

10.5

810

,035

.49

10,1

09.9

7C

DE

172

210

0+0

.139

2−0

.990

3+1

3.91

73−9

9.02

6899

43.0

910

,129

.50

9948

.46

10,1

28.5

3D

EA

130

245

+0.

7660

−0.6

428

+34.

4720

−28.

9254

9957

.01

10,0

30.4

799

64.5

410

,029

.11

E99

91.4

810

,001

.54

10,0

00.0

010

,000

.00

A′

TO

TA

L39

2−8

.52

+1.5

4

Cor

rect

ion

s fo

r E

asti

ngs

Cor

rect

ion

for

B82

/392

×8.

52=

1.78

Cor

rect

ion

for

C74

/392

×8.

52+

B=

1.61

+1.

78=

3.39

Cor

rect

ion

for

D91

/392

×8.

52+

C=

1.98

+3.

39=

5.37

Cor

rect

ion

for

E10

0/39

8.52

+D

=2.

17+

5.37

=7.

54C

orre

ctio

n fo

r A

45/3

92×

8.52

+E

=0.

98+

7.54

=8.

52

Qu

adra

nts

N4

1si

n−v

esi

n+v

eco

s+v

eco

s+v

e

sin

−ve

sin

+ve

cos

−ve

cos

−ve

32

Page 92: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 75

FIG. 5.2 Plot of a typical compass traverse. Theerrors in closure of the traverse can be correctedusing the calculations shown in Table 5.1. (Modifiedfrom Reedman 1979.)

0 25 m

AA′ Closure errorCorrected traverse

Uncorrected traverse

D

C

B

E

A′ A

N

(d)

(a)

(c)

(b)

1 2 3 4 51 2 3 4 5

34

5

21

FIG. 5.3 Sampling strategies across a mineralizedvein. (a) Grab samples across the vein. (b) Channelsamples normal to the vein. (c) Channel samples atthe same height above the floor. (d) Channel samplesat the same height but with individual samplesoriented normal to the strike of the vein. In all casesthe vein is shown in vertical section.

Met

ers

Met

ers

Distance (meters)

ShaftWinze

Joint sets

Vein strike and dipFractures

Survey station

Hematite veins

Contact between granitealteration types

Adit entrance

10Stope

11

219

55°

8

12

13

14 15 16

76

205

19

18

4

Cubby

17

3 1 Gate

PLAN OF LEVELS ONE AND TWO, NORTH LODEGREAT ROCK MINE, DEVON

10,020

10,010

10,000

99,99099,980 99,990 10,000 10,010 10,020 10,030 10,040 10,050 10,060 10,070 10,080 10,09099,970

1025

1020

1015

1010

1005

1000

Stope

2

Winze

LONGITUDINAL SECTION OFLEVELS ONE AND TWO, NORTH LODE

GREAT ROCK MINE, DEVON

21 10 9

11

13 14 12

8 7 6 5

15 16

19 4 18 17

Peg

3 2 1

Ventilationshaft

Travelingway

Stope

Stope

Peg

0 10m

FIG. 5.1 Typical underground survey of a small vein mine using tape and compass.

Page 93: Introduction to mineral exploration

76 C.J. MOON & M.K.G. WHATELEY

weighting function. The weighted average iscalculated until the average falls above the cut-off, which in this case makes 1.59% Zn over3.40 m. The inclusion of just one more samplewould bring the average down below 1.5%. Theoverall weighted average is 1.35% Zn over4.00 m.

5.1.5 Laboratory program

Once the samples have been collected, bagged,and labeled, they must be sent to the laboratoryfor analysis. Not only must the elements ofinterest be specified, but the type of analyt-ical procedure should be discussed with thelaboratory. Cost should not be the overridingfactor when choosing a laboratory. Accuracy,precision, and an efficient procedure are alsoneeded. An efficient procedure within one’sown office is also required. Sample numbersand the analyses requested should be noteddown rigorously on a sample control form(Box 5.1).

An alternative practice is to state clearly theinstructions for test work on the sample sheet

(Box 5.2). In coal analytical work there areseveral ways in which the proximate analyses(moisture, volatile, ash, and fixed carbon con-tents) can be reported; e.g. on an as received,moisture free or dry, ash free basis (Stach 1982,Speight 1983, Ward 1984, Thomas 1992).

A similar form may well be utilized for sandand gravel or crushed rock analyses where thegeologist requires special test work on his orher samples, such as size analysis, aggregatecrushing value (ACV), or polished stone value(PSV) to name but a few.

In today’s computerized era, results are oftenreturned to the company either on a floppy diskor direct to the company’s computer from thelaboratory’s computer via a modem and a tele-phone link. Care must be taken when enteringthe results thus obtained into the company’sdatabase, that the columns of data in the labor-atory’s results correspond with the columnsin your own data base (discussed in detail insection 9.1). Gold values of several percent andcopper values in the ppb range should alerteven the most unsuspecting operator to anerror!

Assay Zn% Average grade Grade using a 1.5%width at a 1.5% weighted average

cut-off cut-off

0.30 0.010.30 0.050.30 0.130.20 0.170.20 0.250.20 0.640.20 0.920.20 1.100.20 2.300.20 1.200.20 2.10 2.26% 1.59%0.20 5.30 over over0.23 1.40 2.10 m 3.40 m0.15 2.000.20 1.900.20 1.800.20 1.300.20 2.600.12 3.504.00 m 1.35% (overall average)

TABLE 5.2 Example of the use ofa grade cut-off to establish themining width of a Pb-Zn veindeposit.

Page 94: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 77

BOX 5.1 An example of a sample control form used to ensure correct analyses are requested.

Submitted by: Submitted to:Sample control file: Date submitted:Returned to: Job File:

Job no. /Area: Date returned:Personnel file:

Channel or Sample no. From (m) To (m) Thickness Description, analyses andborehole no. (m) elements requested

BOX 5.2 An example of a form used to give instructions to the laboratory for detailed coal test work.

To: Date:Order no:

Please analyse the following samples as per the attached instructions

SAMPLE DETERMINATION REQUIREDIDENTIFICATION

On individual samples marked B or M undertake the following:1. Specific gravity of sample2. Crush to 4.75 mm and screen out the <0.5 mm fraction3. Cut out a representative sample of the raw coal and undertake:4. Proximate analysis5. Ultimate analysis (H, N, C, & O) on a moisture free basis6. Analyze for inorganic and organic forms of S and Cl7. Calorific value on a moisture free basis8. Free silica %9. Combined silica %

Submitted by: Signature:

Page 95: Introduction to mineral exploration

78 C.J. MOON & M.K.G. WHATELEY

of experience of the recognition of weatheredoutcrops and the use of a number of simplefield tests.

Many deposits which crop out have beenrecognized because they have a very differentappearance from the surrounding rocks andform distinct hills or depressions. A classicexample of this is the Ertsberg copper depositin Irian Jaya, Indonesia. It was recognizedbecause its green-stained top stood out throughthe surrounding jungle. Its presence was firstnoted by two oil exploration geologists on amountaineering holiday in 1936 and reportedin a Dutch university geological journal (Dozyet al. 1939). A literature search by geologistsworking for a Freeport Sulfur–East Borneo com-pany joint venture found the report and invest-igated the discovery resulting in one of thelargest gold deposits outside South Africa. TheCarajas iron deposits in the Brazilian Amazonwere also recognized because they protrudedthrough rain forest (Machamer et al. 1991). Ona smaller scale silicification is characteristicof many hydrothermal deposits, resulting inslight topographical ridges. It is said that thesilicification of disseminated gold deposits isso characteristic that one deposit in Nevadawas discovered by the crunchy sound of a geo-logist’s boots walking across an altered zoneafter dark. Karst-hosted deposits in limestoneterrain often form distinct depressions, as domany kimberlite pipes, which often occurunder lakes in glaciated areas (see section 17.2).

Besides forming topographical features mostoutcropping mineral deposits have a charac-teristic color anomaly at the surface. The mostcommon of these is the development of a red,yellow or black color over iron-rich rocks,particularly those containing sulfides. Thesealtered iron-rich rocks are known generic-ally as ironstones, and iron oxides overlyingmetallic sulfide deposits as gossans or ironhats (an example is shown in Fig. 5.4). Theserelic ironstones can be found in most areas ofthe world, with the exception of alpine moun-tains and polar regions, and result from theinstability of iron sulfides, particularly pyrite.Weathering releases SO4

2− ions, leaving relic rediron oxides (hematite) or yellow-brown oxy-hydroxides (limonite) that are easily recognizedin the field. Other metallic sulfides weatherto form even more distinctively colored oxides

Element selection

Before samples are submitted to the laboratory,discussions between the project manager, chiefgeologist, and the field staff should take placeto ensure that all the elements that may beassociated with the mineral deposit in questionare included on the analytical request sheetand that analysis includes possible pathfinderelements. Typical elemental associations arediscussed in detail in Chapter 8.

Analytical techniques

There are a wide variety of analytical tech-niques available to the exploration geologist.The method selected depends upon the ele-ment which is being analysed and upon theamount expected. Amongst the instumentalmethods available are atomic absorption spec-trometry (AAS), X-ray fluorescence (XRF),X-ray diffraction (XRD), neutron activationanalysis (NAA), and inductively coupledplasma emission or mass spectrometry (ICP-ES/MS). These analytical methods and theirapplication to different mineral deposit typesare discussed in Chapter 8. AAS is a relativelyinexpensive method of analysis and someexploration camps now have an instrumentin the field; this ensures rapid analysis of thesamples for immediate follow-up. The othermethods involve purchasing expensive equip-ment and this is usually left to specialistcommercial laboratories.

Detailed identification of individual min-erals is usually undertaken using a scanningelectron microscope (SEM) or an electronmicroprobe, and these and other techniques arediscussed in section 2.2.2.

5.1.6 Prospecting

One of the key exploration activities is thelocation of surface mineralisation and any oldworkings. Although this often results fromthe follow-up of geochemical and geophysicalanomalies or is part of routine geological map-ping, it can also be the province of less formallytrained persons. These prospectors compensatefor their lack of formal training with a detailedknowledge of the countryside and an acute eye– an “eye for ore.” This eye for ore is the result

Page 96: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 79

FIG. 5.4 Gossan (above adit) overlying sulfidesexposed in the pit wall of the San Miguel deposit,Rio Tinto, southern Spain.

or secondary minerals. For example, coppersulfides oxidize to secondary minerals thathave distinctive green or blue colors such asmalachite and azurite. Metals such as lead andzinc normally form white secondary mineralsin carbonate areas that are not easily distin-guished on color grounds from the host carbon-ates. A fuller list is given in Table 5.3. Besidesproviding ions to form secondary minerals, sul-fides often leave recognizable traces of theirpresence in the form of the spaces that theyoccupied. These spaces are relic textures,often known as boxworks from their distinct-ive shapes, and are frequently infilled withlimonite and goethite. Ironstones overlyingsulfides have a varied appearance with much

alternation of color and texture giving rise tothe term “live,” in contrast to ironstones ofnonsulfide origin that show little variation andare known as “dead” ironstones. Ironstonesalso preserve chemical characteristics of theirparent rock, although these can be consider-ably modified due to the leaching of mobileelements.

The recognition of weathered sulfides over-lying base metal or gold deposits and theprediction of subsurface grade is therefore ofextreme importance. This particularly appliesin areas of laterite development. These areas,such as much of Western Australia, have beenstable for long periods of geological time, allrocks are deeply weathered, and the percent-age of ironstones that overlie nonbase metalsulfides (known rather loosely as false gossans)is large. In Western Australia the main tech-niques for investigating these are (Butt &Zeegers 1992):1 visual description of weathered rocks;2 examination for relic textures;3 chemical analysis.

Visual recognition requires experience anda good knowledge of primary rock textures.Visual recognition can also be supplementedby chemical analysis for trace and major ele-ments. Hallberg (1984) showed that Zr-TiO2

plots are extremely useful in confirming rocktypes. Both elements are immobile and easyand cheap to determine at the levels required.They may be supplemented by Cr determina-tions in ultrabasic areas, although Cr is moremobile than the other two elements.

Relic textures can sometimes be observedin hand specimen but this is usually supple-mented by microscopic examination usingeither a binocular microscope or a petrographicmicroscope and an impregnated thin section.In a classic study of relic textures, Blanchard(1968) described the textures resulting from theweathering of sulfides. Major types are shownin Fig. 5.5. Although boxwork textures are dia-gnostic they are, unfortunately, not present inevery gossan overlying base metal sulfides andtextural examination must be supplementedby chemical analysis.

The choice of material to be sampled and theinterpretation need to be carefully scrutinized.Andrew (1978) recommends taking 20 samplesas representative of each gossan and using a

Page 97: Introduction to mineral exploration

80 C.J. MOON & M.K.G. WHATELEY

TABLE 5.3 Colours of minerals in outcrop.

Mineral or metal

Iron sulfidesManganeseAntimonyArsenicBismuthCadmiumCobaltCopperLeadMercuryMolybdenumNickelSilverUraniumVanadiumZinc

Mineral/compound in outcrop

Goethite, hematite, limonite, sulfatesMn oxides, wadAntimony bloomIron arsenateBismuth ochresCadmium sulfideOxides, erythriteCarbonates, silicates, sulfates, oxides, native CuCerussite, anglesite, pyromorphiteCinnabarMolydenum oxides, iron molybdateAnnabergite, garnieriteChlorides, native AgTorbernite, autuniteVanadatesSmithsonite

(a) (b) (c)

(d) (e) (f)

(g) (h) (i) (j)

FIG. 5.5 Typical boxwork structures. Primary oreminerals were: (a–c) galena (a – cleavage, b – mesh,c – radiate); (d,e) sphalerite (d – sponge structure,e – cellular boxwork); (f) chalcopyrite; (g,h) bornite(triangular cellular structure); (i,j) -tetrahedrite(contour boxwork). Approximately 4×magnification. (After Blanchard & Boswell 1934.)

Outcrop color

Yellows, browns, chestnuts, redsBlacksWhiteGreenish, greens, yellowishLight yellowLight yellowBlack, pink, sometimes violetGreens, bluesWhite, yellowRedBright yellowGreenWaxy green, yellowBright green, yellowGreen, yellowWhite

multielement analysis, either XRF or ICP–ES,following a total attack to dissolve silica. Theinterpretation needs to be treated with care. Alarge amount of money was wasted in West-ern Australia at the height of the nickel boomin the early 1970s because ironstones wereevaluated for potential on the basis of theirnickel content. While most ironstones withvery high nickel contents do overlie nickelsulfide deposits, a number have scavengedthe relatively mobile nickel from circulatingground waters or overlie silicate sources ofnickel and a number of deposits have a weaknickel expression at surface. More carefulresearch demonstrated that it is better to con-sider the ratio of nickel to more immobile ele-ments, such as copper, or even better immobileiridium that is present at the sub-ppm level inthe deposits. These elements can be combinedinto a discriminant index (Travis et al. 1976,Moeskops 1977). Similar considerations alsoapply to base metal deposits; barium and leadhave been shown to be useful immobile tracersin these (summarized in Butt & Zeegers 1992).

Field tests

Although a large number of field tests havebeen proposed in the literature only a few of thesimplest are in routine use. At present only twogeophysical instruments can be routinely car-ried by man, a scintillometer to detect gamma

Page 98: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 81

ZONESECONDARY

MINERALSRELATIVE

Cu CONTENTO

XID

IZIN

GR

ED

UC

ING

Leaching

Oxide enrichment

Secondary sulfideenrichment

Primary sulfides

Gossan

Soluble elementsremoved

Disseminatedchalcopyrite–pyrite

MalachiteAzuriteCupriteNative copperChrysocolla

ChalcociteBorniteCovelliteChalcopyritePyrite

Water table

Leaching

Leaching

Iron oxides

0

Dep

th (

m)

100

50

FIG. 5.6 Diagram showing the gossan and copper minerals formed by weathering of a copper sulfide vein.(After Levinson 1980.)

radiation and a magnetic susceptibility meterto determine rock type. The application ofthe former is discussed in section 7.5. Ultra-violet lamps to detect fluorescence of mineralsat night are commonly used, particularly forscheelite detection, although a number of otherminerals glow (Chaussier & Morer 1987). Inmore detailed prospecting darkness can becreated using a tarpaulin.

A number of simple field chemical tests aidin the recognition of metal enrichments in thefield, usually by staining. Particularly usefultests aid the recognition of secondary lead andzinc minerals in carbonate areas. Lead mineralscan be identified in outcrops by reaction withpotassium iodide following acidification withhydrochloric acid. Lead minerals form a brightyellow lead iodide. A bright red precipitateresults when zinc reacts with potassium fer-rocyanide in oxalic acid with diethylaniline.Gray copper minerals can be detected withan acidified mixture of ammonium pyropho-sphate and molybdate and nickel sulfides withdimethyl glyoxine. Full details are given inChaussier and Morer (1987).

Supergene enrichment

As metallic ions are leached at the surfacethey are moved in solution and deposited atchanges in environmental conditions, usuallyat the water table, where active oxidation andgeneration of electrical charge takes place. Thisleaching and deposition has very importanteconomic implications that are particularlysignificant in the case of copper deposits asillustrated in Fig. 5.6. Copper is leached fromthe surface, leaving relic oxidized mineralsconcentrated at the water table in the form ofsecondary sulfides. A simple Eh–pH diagramenables the prediction of the occurrence ofthe different copper minerals. From the sur-face these are malachite, cuprite, covellite,chalcocite, bornite, and chalcopyrite (the pri-mary phase). The changes in mineralogy haveimportant implications for the recovery ofthe minerals and the overall grade of the rock.Oxide minerals cannot be recovered by theflotation process used for sulfides and theirrecovery needs a separate circuit. The overallgrade of the surface rock is much lower than

Page 99: Introduction to mineral exploration

82 C.J. MOON & M.K.G. WHATELEY

the primary ore and sampling of the surfaceoutcrops does not reflect that of the primaryore. Near the water table the copper gradeincreases, as the result of the conversion ofchalcopyrite (35% Cu) to chalcocite (80% Cu).This enrichment, known as supergene enrich-ment, often provides high grade zones in dis-seminated copper deposits, such as those of theporphyry type. These high grade zones pro-vide much extra revenue and may even be thebasis of mines with low primary grades. In con-trast to the behavior of copper, gold is lessmobile and is usually concentrated in oxidezones. If these oxide zones represent a signific-ant amount of leaching over a long period thenthe gold grades may be become economic. Anexample of this is the porphyry deposit of OkTedi discussed in Box 11.4. These near surface,high grade zones are especially important asthey provide high revenue during the earlyyears of a mine and the opportunity to repayloans at an early stage. These gold-rich gossanscan also be mechanically transported for con-siderable distances, e.g. the Rio Tinto Mine insouthern Spain.

Float mapping

The skill of tracing mineralized boulders orrock fragments is extremely valuable in areasof poor exposure or in mountainous areas.In mountainous areas the rock fragmentshave moved downslope under gravity and thelithology hosting the mineralisation can bematched with a probable source in a nearbycliff and a climb attempted. Float mapping andsampling is often combined with stream sedi-ment sampling and a number of successfulsurveys have been reported from Papua NewGuinea (Lindley 1987). In lowland areas miner-alized boulders are often disturbed during cul-tivation and may be moved to nearby walls.In this case it is often difficult to establish asource for the boulders and a soil survey and asubsequent trench may be necessary. Burrow-ing animals may also be of help. Moles or rab-bits, and termites in tropical areas, often bringsmall fragments to the surface.

In glaciated areas boulders may be movedup to tens of kilometers and distinct bouldertrains can be mapped. Such trains have beenfollowed back to deposits, notably in central

Canada, Ireland, and Scandinavia (Fig. 5.7). InFinland the technique proved so successfulthat the government offered monetary rewardsfor finding mineralized boulders. Dogs are alsotrained to sniff out the sulfide boulders as theirsense of smell is more acute than that of theexploration geologist!

5.1.7 Physical exploration: pitting andtrenching

In areas of poor to moderate outcrop a trench(Fig. 5.8) or pit is invaluable in confirming thebedrock source of an anomaly, be it geological,geochemical, or geophysical. The geology ofa trench or pit wall should be described andillustrated in detail (Fig. 5.9). For further detailssee section 9.2. Trenches and pits also providelarge samples for more accurate grade estim-ates as well as for undertaking pilot processingplant test work to determine likely recoveries.

Some operators in remote areas, particu-larly in central Canada, strip relatively largeareas of the overburden to enable systematicmapping of bedrock. However, this wouldnow (probably) be regarded as environmentallyunfriendly.

Lake

N

0 1 km

Mineralized boulder

Subcrop of deposit

FIG. 5.7 Train of mineralized boulders from theLaisvall lead deposit, Sweden.

Page 100: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 83

surveys. Not all the information obtained willbe useful, indeed some may be misleading, anddistinction should be clearly made betweenmeasured and interpreted data. Geophysicaland geochemical surveys can indicate the sur-face or subsurface geology of the potential hostrocks or more directly the presence of mineral-isation as discussed in Chapters 7 and 8. Tradi-tionally the results of these surveys have beencombined by overlaying colored transparentcopies of the data on a topographical or geolog-ical paper base. It is then possible to determinethe interrelation between the various datasets. More recent developments have allowed

750

1410

2510195040

7525

345Marble with local

traces of malachite

Marble withtraces malachiteMarble, weakly

mineralized withmalachite

Barren pinkmarble

Finely banded qtz - micaschist, thin qtz horizons,

possible isoclinalfolding, flat dip

Barren pinkmarble

Banded mica m arblewith 3 cm band moderatelymineralized with malachite

Str. 090 mag.dip 21°N

Str. 040 mag.dip 18°NW

470

750 Sample with ppm Cu

29°N

0

Distance (m)

5 10 15 20 25

S N

FIG. 5.9 Sketch of a trench testing a copper anomaly in central Zambia. (From Reedman 1979.)

An alternative to disturbing the environ-ment by trenching is to use a hand-held drillfor shallow drilling. This type of drill is light-weight and can be transported by two people.It produces a small core, usually around 25–30 mm in diameter. Penetration is usually lim-ited, but varies from around 5 m to as much as45 m depending on the rig, rock type, and skillof the operators!

5.1.8 Merging the data

The key skill in generating a drilling target isintegrating the information from the various

FIG. 5.8 Typical explorationtrench, in this case on the SouthAfrican Highveld.

Page 101: Introduction to mineral exploration

84 C.J. MOON & M.K.G. WHATELEY

the use of computerized methods (GIS) whichallow rapid integration and interrogation of thedatabases and are discussed in detail in section9.2. A straightforward example of the overlay-ing of different data types is shown in Fig. 5.10taken from the work of Ennex plc in NorthernIreland as detailed in Clifford et al. (1990,1992). The area was chosen for investigationbecause of the presence of favorable geology,placer gold in gravels and anomalous arsenicgeochemistry discovered during governmentsurveys. A licence was applied for in 1980.Ennex geologists investigated the anomalousareas and by careful panning and float mappingidentified the source of the gold as thin quartzveins within schist. Although the veins iden-tified assayed about 1 g t−1 Au and were notof economic grade, they were an indicationthat the area was mineralized and prospect-ing was continued over a wider area. In late1983 economically significant gold in bedrockwas found and subsequently three other gold-bearing quartz veins were found where a smallstream has eroded to bedrock. More than 1300mineralized boulders were mapped over astrike of 2700 m. Channel samples of the firstidentified vein gave assays of 14 g t−1 Au over3.8 m with selected samples assaying over150 g t−1 Au.

Detailed follow-up concentrated on buildingup a systematic picture of the area where thereis no outcrop. This focused on boulder mappingand sampling, deep overburden geochemistryand geophysical mapping using very low fre-quency electromagnetic resistivity (VLF–EM/R) surveys. Deep overburden geochemistrywas undertaken using a small core overburdendrill to extract 250 g samples on a 50 × 20 mgrid. This spacing was decided on follow-ing an orientation study over a trenched vein.Geochemical sampling defined 33 anomalies(>100 ppb Au) (Fig. 5.10a). A variety of geo-physical methods were tested including in-duced polarization and VLF–EM/R. No methodlocated the quartz veins but the VLF–EM/Rmethod did define previously unidentified goldbearing shears (Fig. 5.10b) and the subcrop ofgraphitic pelites. The nature of the geophysicaland geochemical anomalies was defined by a2850 m trenching program (Fig. 5.10c). Besidesincreasing the number of gold-bearing struc-tures to 16, this trenching was also instru-

mental in demonstrating the mineralogy ofthe mineralized structures. The quartz veinscontain significant sulfides, mainly pyritebut with lesser chalcopyrite, galena, sphalerite,native copper, and tetrahedrite–tennantite. Bycontrast, the shears are richer in arsenopyritewith chloritization often observed. Thus geo-chemistry defined the approximate locationof the quartz veins and auriferous shears, geo-physics the location of the shears, and tren-ching the subcrop of the quartz veins.

Once the gold-bearing structures had beendefined their depth potential was tested by dia-mond drilling (Fig. 5.10c). The diamond drilling(6490 m in 63 holes) took place during 1985–86and defined a resource of 900,000 t at 9.6 g t−1.Subsequent underground exploration con-firmed the grades and tonnages indicated bydrilling but the deposit is undeveloped becauseof a ban on the regular use of explosives inNorthern Ireland by the security authorities.

5.2 WHEN TO DRILL AND WHEN TO STOP

One of the hardest decisions in exploration isto decide when to start drilling, and an evenharder one is when to stop.

The pressure to drill will be evident whenthe program has identified surface mineralisa-tion. Management will naturally be keen totest this and gain an idea of subsurface miner-alisation as soon as possible. However, thispressure should be resisted until there is areasonable idea of the overall surface geologyand the inferences that can be made from thisknowledge concerning mineralisation in depth.

5.2.1 Setting up a drilling program

The geologist in charge of a drilling program isfaced with a number of problems, both logist-ical and geological. There must be a decision onthe type of drilling required, the drillhole spac-ing (see section 10.4.4), the timing of drilling,and the contractor to be used. The logistics ofdrilling should be considered carefully as thedrill will need drill crews, consumables, andspare parts; this will require helicopter supportin remote areas and vehicle access in morepopulated areas. Many drills require vehicleaccess and access roads must be made and pads

Page 102: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 85

FIG. 5.10 Drilling based on acombination of (a) geochemistry,(b) geophysics, and (c) trenchingand prospecting, SperrinMountains, Northern Ireland.(After Clifford et al. 1992.)

Extent of survey

Road

Psammite

Psammite

Psammite

Road

Road

Owenkillewvein

O'Haganvein

Road veinSheep Dip vein

Mullan vein

No.1 veinBend vein

Crows Foot vein

Kiln shear

Shear

Cur

ragh

ina l

t Bur

n

Owenkillew River

Owenkillew River

Owenkillew River

(a)

>100 ppb Au

Auriferous structure

Mineralized boulder

(b)

>10 VLF e.m. Fraser FilterGraphitic pelite

(c)

Auriferous structureLithological boundaryFault

0 300 m

0 300 m

0 300 m

Trench

Auriferous structure

Diamond drillhole

N

N

N

Page 103: Introduction to mineral exploration

86 C.J. MOON & M.K.G. WHATELEY

tainous areas. In areas without access prob-lems, typical drill hole patterns are square witha regular pattern or with rows of holes offsetfrom adjacent holes. The first hole normallyaimed at the down dip projection of surfaceanomalies or the interpreted centre of subsur-face geophysics. Most programs are plannedon the basis of a few test holes per target with areview of results while drilling. The spacingbetween holes will be based on anticipatedtarget size, previous company experience withdeposits of a similar type, and any informationon previous competitor drilling in the district(Whateley 1992). The subsequent drilling loca-tion and orientation of the second and thirdholes will depend on the success of the firsthole. Success will prompt step outs from thefirst hole whereas a barren and geologicallyuninteresting first hole will suggest that an-other target should be tested.

Once a deposit has been at least partlydefined then the continuity of mineralisationmust be assessed. The spacing between holesdepends on the type of mineralisation and itsanticipated continuity. In an extreme case, e.g.some vein deposits, boreholes are mainly ofuse in indicating structure and not much usein defining grade, which can only be accur-ately determined by underground sampling (seeChapter 10). Typical borehole spacing for avein deposit is 25–50 m and for stratiformdeposits anything from 100 m to several hun-dreds of meters. Examples of drill spacing andorientation for a variety of deposits is shown inFig. 5.12 (pp. 87–89).

5.2.2 Monitoring drilling programs

Monitoring the geology and mineralisationintersected during a drilling program is vitalin controlling costs. In the initial phases ofdrilling this may involve the geologist stayingbeside the rig if it is making rapid progress, e.g.when using percussion drilling, and loggingmaterial as it comes out of the drill hole. In thecase of diamond drilling twice daily visits toexamine core, make initial logs, and decide onthe location of the next drill holes are usuallysufficient, although longer visits will be neededwhen cutting potentially mineralized zones ornearing the scheduled end of the hole. Oftenthe geologist will be required by contractors to

N

S

S N

4 1

6

8

3

2

9

7

5

Drill lineAnomaly

Weathered zone

“Fresh” Bedrock

Overburden

DDH 2

EOH

50°

40°

Anomaly

Minera

lisat

ion

FIG. 5.11 Idealized initial drill grid. EOH, end of hole.(After Annels 1991.)

for drilling constructed so that the drill rigs canbe set on an almost horizontal surface.

The pattern of drilling used is dependenton the assumed attitude and thickness of thedrilling target. This depends on the availableinformation which may, of course, be inaccur-ate. Drilling often causes reconsideration ofgeological ideas and prejudices. Vertical bore-holes are the easiest and cheapest to drill andwidely used for mineralisation with a shallowdip or for disseminated deposits. However,inclined holes are usually preferred for targetswith steep dips. The aim will be to cut the min-eralisation at 90 degrees with the initial hole,cutting immediately below the zone of oxida-tion (weathered zone) (Fig. 5.11).

Drilling is used to define the outlines of anydeposit and also the continuity of mineralisa-tion for purposes of resource estimation. Theinitial pattern of drilling will depend on surfaceaccess, which may be very limited in moun-

Page 104: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 87

FIG. 5.12 (a) Typical combination ofshallow and deeper drilling overthe Doris Creek vein gold depositin Western Australia. The shallow(rotary air blast, RAB) drillingdefines the subcrop of the veinwhereas the deeper (reversecirculation, RC) drilling definesgrades. (After Jeffery et al. 1991.)

5500mN

5000mN

4500mN

4000mN

10,000mE

SECTION 4625mN

Basalt (minor dolerite)

Granodiorite (mylonitic in part)

Qtz – Ser Sch (Sheared porphyry)

Ultramafic intrusive

Black shale, argillite

Dolerite

Au mineralisation

Water tableWT

9900mE 10,000mE 10,100mE 10,200mE

81 m

46 m

DC 33 DC 34

50 m 50 m 50 m 50 m 50 m 50 m 50 m 50 m 50 m 50 m

WT

Vacuum drillsample horizon

Quaternary cover

(a)

NORTHERNGP/BP ZONE(WEST LODE)

EASTERN ZONE(EAST LODE)

To King of Creation

SOUTHERN ZONE(WEST LODE)

Section 4625mN

To Laverton

N

0 300 m

RC Dril lhole

RAB Dril lhole

Surface trace ofmineralisationPotentialresource zone

Doris

Doris CKShear

CK Shear

Page 105: Introduction to mineral exploration

88 C.J. MOON & M.K.G. WHATELEY

00mRL

–300mRL

–600mRL

–900mRL

–1200mRL

–1500mRL500mE 800mE 1100mE 1400mE

0 200 m

W EGloryHole

MA

IN S

HA

FT

Chloritic schist

Cordierite schist

Intermediate schist

Garnet schist

Quartz – feldspar porphyry

Fault

Diamond drillhole

Old stope

Mineralisation

(b)

FIG. 5.12 (b) Typical drilling of a previously mined deep vein gold deposit, Big Bell, Western Australia. Notethe deflections on the deeper holes, the abandoned holes, and combination of drilling both from surface andunderground. (After Handley & Carrey 1990.)

Page 106: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 89

5500N5000N4500N

0

–500

4500EA16 R2 IM6

A11 A14A15 M4

H17H17A H38

635.20 m H38

472.44 mH17A

135.4 mH17

579.12 mA15

355.10 mM4

199.69 mA14252.70 m

A11

553.52 mIM6

46.33 mR2

198.73 mA16

Silurian

Ordovician

Precambrian

Shale and siltstoneSandstone and tuff

Massive quartz

Basic dyke and sillAcid tuff and rhyoliteShale and siltstoneQuartz breccia and stockworkBasicAcid and basic tuffs

Slump brecciaSandstone and tuff

Mona complex thrust breccia

Fault

FIG. 5.13 Geological strip logs and plot of drill holes, Parys Mountain zinc prospect, Anglesey, Wales.(Data courtesy Anglesey Mining, plot by L. Agnew.)

KD21KD69

KD11

KD75

KD2

KD72 KD77 KD19KD

136

KD

55

KD

16

KD

57

KD

20N S

12/10

11/35 7/343/34

2/46

3 /9

8/18

0

50

100

Dep

th (

m)

500 100 150 200 250Distance (m)

Ore zones3 m at 9 kg t–1 U3O8

Diamond drillhole with ore intersection

(c)

FIG. 5.12 (c) Typical drill section of an ill-definedmass, Kintyre uranium deposit, Western Australia.(After Jackson & Andrew 1990.)

sign for progress or the use of casing. Besidesdetailed core logs, down hole geophysical log-ging (see section 7.14) is often used and arrange-ments should be made with contractors fortimely logging as holes can become rapidlyblocked and cleaning of holes is an expensiveundertaking.

Data on mineralisation, the lithologies,and structures hosting it should be recordedand plotted on to a graphic log as soon asthe information becomes available. Initially,strip drill logs (Fig. 5.13) can be used and sec-tions incorporating the known surface geology.Assay information will usually be delayed for afew days as the cores will need to be split or cutand sampled before analysis (see Chapter 10).However estimates of the location and import-ance of mineralisation can be plotted alongsidethe lithological log (Fig. 5.15e). As furtherdrilling proceeds the structural and stratigraph-ical controls on mineralisation should become

Page 107: Introduction to mineral exploration

90 C.J. MOON & M.K.G. WHATELEY

clearer. Ambiguities concerned with the inter-pretation of drilling are common and oftencannot be resolved until there has been under-ground development. A typical example isshown in Fig. 5.14 in which three differentinterpretations are possible from the informa-tion available.

Once a stratigraphy has been established andseveral boreholes are available, more sophistic-ated plotting techniques can be used. Typicalmethods used to plot drillhole informationare structure contour plans, isopach, grade(quality), thickness, and grade multiplied bythickness, known as the grade–thickness prod-uct or accumulation maps. Grade and grade–thickness product maps are extremely usefulin helping to decide on the areal location oforeshoots and of helping direct drilling towardsthese shoots (Fig. 5.15) (pp. 91, 92).

One of the key issues in any drilling programis continuity of mineralisation. This deter-mines the spacing of drill holes and the accur-acy of any resource estimation (as discussed inChapter 10). In most exploration programs thecontinuity can be guessed at by comparisonwith deposits of a similar type in the same dis-trict. However, it is usual to drill holes to testcontinuity once a reasonable sized body hasbeen defined. Typical tests of continuity are todrill holes immediately adjacent to others andto test a small part of the drilled area with acloser spacing of new holes.

5.2.3 Deciding when to stop

Usually the hardest decision when directing adrilling program is to decide when to stop. Themain situations are:1 No mineralisation has been encountered.2 Mineralisation has been intersected, but it isnot of economic grade or width.3 Drill intercepts have some mineralisation ofeconomic grade but there is limited continuityof grade or rough estimates show that the sizeis too small to be of interest.

(a)

(b)

(c)

905 m

865 m

905 m

865 m

Actual ore body sections

Sections obtained by interpolation

Zone of dislocation

Borehole

905 m

865 m

FIG. 5.14 (opposite) (a–c) Three differentinterpretations of drill intersections in an irregulardeposit compared with outlines defined by mining.(After Kreiter 1968.)

Page 108: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 91

11,500E 11,750E 12,000E 12,250E

(a)

(c)

IV59IV58

IV56

IV103

IV55IV54

IV102IV53IV52

IV101IV47

IV46IV45

IV44

IV11718,500N

18,000N

17,500N

IV147

IV37IV38

IV39IV40

IV126

IV116

IV51IV10

IV11IV12

IV60IV35IV34

IV33IV9

IV36IV107

IV108

IV115

IV125

IV61IV8

IV7IV6

IV50IV114

IV124

IV31IV30 IV29

IV5IV32

IV106IV113

IV123

IV127

IV140

IV122

IV112

IV49IV2

IV3IV4

IV13IV27IV14

IV1IV28

IV105IV11

1 IV121

IV148

IV128

IV43IV42

IV41IV104

IV110

IV120

IV129

IV17

IV16IV15

IV119

IV118

IV130

IV131IV139

IV138

IV132

IV133IV134

IV62IV20IV19

IV18IV135

IV149

IV136

IV137

IV146

IV141

IV142

IV143IV144

IV145

Section18000N

18,500N

18,250N

18,000N

17,750N

17,500N

17,250N

17,000N

35

32.5

30

27.5

22.520

25

N

N

FIG. 5.15 (a–d) Typical plots used in interpretation and planning of drilling. The example is a stratiform copperdeposit of the Zambian type: (a) drillhole plan and numbers; (b) structure contour of the top of the mineralizedinterval (meters above sea level); (c) isopach of the mineralized interval (m); (d) grade thickness product(accumulation) of the mineralized interval (m%).

>150 m%

100 – 150 m%

50 – 100 m%

10 – 50 m%

4 – 10 m%

(b)

(d)

18,500N

18,250N

18,000N

17,750N

17,500N

17,250N

17,000N

190019

5020002050

2100

2150

2200

18,500N

18,000N

17,500N

11,500E 12,000E 12,500E

N

N

Page 109: Introduction to mineral exploration

92 C.J. MOON & M.K.G. WHATELEY

processes forming economic grade mineralisa-tion were operative but not at a sufficient scaleto produce a large deposit, at least not in thearea drilled. The encouragement of high grademineralisation will often result in severalphases of drilling to test all hypotheses andsurrounding areas to check whether the sameprocesses may have taken place on a largerscale. This type of prospect is often recycled(i.e. sold) and investigated by several differentcompanies before being abandoned or a deposit

Zone of copper mineralisation

(e)

2420 IV27 IV1 IV2 IV49 IV113 IV114 IV115

200 m

300 m

400 m

600 m 600 m 600 m 600 m

2000

11,500E 12,000E

Soil

Black shale

QuartzitesQuartzitic sandstones

Basement complex

ArgillitesFine-grained limestones

Weathered zone

Conglomerate

MarlsDolomitic limestonesSandy limestones

Sandy shalesArgillaceous sandstones

FIG. 5.15 (e) section through the deposit along line 18,000N shown in Fig. 5.15(a).

4 A body of potentially economic grade andsize has been established.5 The budget is exhausted.

The decision in the first case is easy,although the cause of the surface or subsurfaceanomalies that determined the siting of thedrill hole should be established. The secondcase is more difficult and the possibility ofhigher grade mineralisation should have beeneliminated. The presence of some interestingintersections in a prospect suggests that the

Page 110: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 93

discovered. Often it is only some factor suchas the impending deadline on a mineral rightsoption that will result in a decision to exercisethe option or drop this type of prospect.

The definition of a potentially economicdeposit is an altogether more agreeable prob-lem, and the exact ending of the explorationphase will depend on corporate policy. If asmall company is having to raise money forthe evaluation phase, it will probably leave thedeposit partly open (i.e. certain boundariesundetermined), as a large potential will en-courage investors. Larger companies normallyrequire a better idea of the size, grade, andcontinuity of a deposit before handing over thedeposit to their pre-development group.

It is usual for a successful program to extendover several years, resulting in high costs.Exhausting a budget is common; however over-spending may result in a sharp reprimandfrom accountants and the painful search fornew employment!

5.3 RECYCLING PROSPECTS

The potential of many prospects is not resolvedby the first exploration program and they arerecycled to another company or entrepreneur.This recycling has many advantages for theexplorer. If for some reason the prospect doesnot meet their requirements, they are oftenable to recover some of the investment in theprospect with the sale of the option or the datato a new investor. Often it is not an outrightsale but one in which the seller retains someinterest. This is termed a joint venture or farmout. Typical agreements are ones in which thebuyer agrees to spend a fixed amount of moneyover a fixed period of time in order to earn a pre-agreed equity stake in the prospect. The periodover which the funding takes place is known asthe earn-in phase. Normally the seller retainsan equity stake in the project and the rightto regain complete control if the buyer doesnot meet its contractual obligations. The typeof equity retained by the seller varies; it may in-clude a duty to fund its share of developmentcosts or it may have no further obligations – afree or carried interest. The seller may as analternative choose merely to have a royalty ifthe project is eventually brought to production.

Joint venturing is normally carried out by cir-culating prospective partners with brief detailsof the project, but this may exclude locationand other sensitive details. Seriously interestedparties will then be provided with full detailsand the possibility of visiting the prospect togain a more accurate impression before com-mitting any money.

Buying or re-examining a prospect that aprevious party has discarded is tempting, butthe buyer must be satisfied that they can bemore successful than the other explorer. Thissuccess may result from different explorationmodels or improved techniques.

Mulberry Prospect

A good example of this is the exploration ofthe area on the north side of the St AustellGranite in southwestern England, here termedMulberry after the most productive old mine(Dines 1956). It was examined by three differ-ent companies from 1963 to 1982, as reportedin British Geological Survey (BGS) open file re-ports. The Mulberry area is of obvious interestto companies as it is one of the few in Cornwallthat supported open pit mining of dissemin-ated tin deposits in the nineteenth centuryand might therefore be more efficiently minedusing advances in mining technology. Thedeposits worked (Fig. 5.16) were the Mulberryopen pit with a N–S strike and a series of openpits on the Wheal Prosper structure which runsE–W. All the open pits were based on a sheetedvein complex or stockwork of thin veins con-taining cassiterite, tourmaline, and quartz. Inaddition there were a number of small under-ground workings based on E–W copper- and N–S iron-bearing veins. All the deposits are withinthe metamorphic aureole of the St AustellGranite, hosted in slate and another metamor-phic rock known locally as calc-flinta. Thisconsists of garnet, diopside, actinolite, quartz,and calcite and was probably originally an im-pure tuff or limestone.

The initial examination in 1964 resultedfrom a regional appraisal in which Consoli-dated Gold Fields (CGF) sought bulk miningtargets that were amenable to improved min-ing and processing techniques. CGF immedi-ately identified Mulberry as being of interestand confirmed this by sampling the old open

Page 111: Introduction to mineral exploration

94 C.J. MOON & M.K.G. WHATELEY

Limit of GeochemicalSoil Survey

●●

●●

92m

A6

A4

A1&2

A3

A5

A15A14

A7

A8

A9

A10

A11

153

m

122 m

Adit

61 m

92 m

Mulberry mine

Prosperquarry

302010

40

50

60

0 300 mElvan

Meadfoot bedsCalc flinta

N

FIG. 5.16 Consolidated Gold Fields Program 1964–66: drilling along strike from the Mulberry open pit.

Page 112: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 95

pit and the underground accesses. In additionthe company investigated the potential of thearea around the open pit and particularly itsstrike extensions by taking soil samples forgeochemical analysis along E–W lines (Hosking1971). These suggested that the deposit contin-ued to the south into the calc flintas. CGF thenundertook a drilling program to test the poten-tial of the deposit at depth (Fig. 5.16). Drillingover the period 1964–66 confirmed that themineralisation continued at roughly the widthof the open pit but that the cassiterite was con-tained in thin erratic stringers. Drilling to thenorth of the open pit showed a sharp cut-offwhereas drilling to the south showed that tinvalues continued from the slate host into a calcflinta host where mineralisation became muchmore diffuse. Although the grades intersectedwere similar to those in the open pit, the sizeand grade were not considered economic, espe-cially when compared with Wheal Jane (even-tually a producing mine) which CGF was alsoinvestigating, and the lease was terminated.

The next phase of exploration was from1971 to 1972 when the area was investigatedby Noranda–Kerr Ltd, a subsidiary of the Cana-dian mining group Noranda Inc. This explora-tion built on the work of CGF and investigatedthe possibility that the diffuse mineralisationwithin the calc silicates might be strataboundand more extensive than previously suspected.Noranda conducted another soil survey (Fig.5.17) that showed strong Sn and Cu anomaliesover the calc silicates. Four drillholes werethen sited to test these anomalies after pitswere dug to confirm mineralisation in bedrock.The drillholes intersected the calc silicates butthe structure was unclear and not as predictedfrom the surface. In addition although exten-sive (>20 m thick) mineralisation was inter-sected it was weak (mainly 0.2– 0.5% Sn) withfew high grade (>1% Sn) intersections. Theprogram therefore confirmed the explorationmodel but was terminated.

As part of an effort to encourage mineralexploration in the UK, the British GeologicalSurvey investigated the southern part of thearea (Bennett et al. 1981). Their surveys alsodefined the southern geochemical anomaly tothe SW of Wheal Prosper, as well as an area ofalluvial tin in the west of the area. The exten-sion of the Mulberry trend was not as clearly

detected, probably because of the use of a widersampling spacing. BGS also drilled two holes(Fig. 5.19) to check the subsurface extension ofthe western end of the Wheal Prosper system aswell as a copper anomaly adjacent to the oldpit. Limited assays gave similar results (up to0.3% Sn) to surface samples.

The third commercial program was a moreexhaustive attempt from 1979 to 1982 by Cen-tral Mining and Finance Ltd (CMF), a subsi-diary of Charter Consolidated Ltd, themselvesan associate company of the Anglo Americangroup. Charter Consolidated had substantialinterests in tin and wolfram mining, includingan interest in the South Crofty Mine and aninterest in Tehidy Minerals, the main mineralrights owners in the area. CMF’s approachincluded a re-evaluation of the Mulberry pitand a search for extensions, as well as a searchfor extensions of the Wheal Prosper systemand an attempt to identify any buried granitecupolas with their associated mineralisation.Their program contained four phases. In thefirst phase the Mulberry pit was resampled andregional soil geochemistry and gravity surveyswere undertaken. The gravity survey identifieda perturbation in the regional gravity fieldalong the Mulberry trend (Fig. 5.18). Regionalshallow soil geochemistry samples were takenon 100 m centers and analysed for Sn, Cu,As, Mo, and W. They showed an ill-definedsouthern continuation of the Mulberry systemin Cu data and a large Cu and As anomaly tothe southwest of the Wheal Prosper system.Resampling of the Mulberry pit led to a re-estimation of the resource to 1.1 Mt at 0.44%Sn using a 0.2% Sn cut off. This would formpart of the resource for any mine.

The Phase 2 program in 1980 consisted of adetailed geochemical soil survey follow-up ofanomalous areas identified in phase 1, as wellas the planning of two holes to test the geo-physical anomaly. The first hole (CMF1, Fig.5.19) was a 540 m deep diamond-cored bore-hole that failed to find any indication of graniteor associated contact metamorphism, althoughit did cut vein mineralisation of 0.51% Sn over3.1 m at 206.5 m depth. Because of the lack ofgranite the planned second hole was cancelled.

The results of the soil geochemistry surveyencouraged CMF to progress to phase 3 witha deep overburden sampling program to the

Page 113: Introduction to mineral exploration

96 C.J. MOON & M.K.G. WHATELEY

Limit ofSurvey

92m

122 m

153

m

Adit

61 m

92 m

Mulberry mine

Prosperquarry

302010

40

50

60

0 300 m>200 ppm Sn

Noranda Drill Hole

Elvan

Meadfoot bedsCalc flinta

N

4

1

23

200 ppm Sn

FIG. 5.17 Noranda Kerr Program 1971–72: soil surveys and drilling of calc silicates.

Page 114: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 97

Adit

–14.4–14.6

–14.

8

–15.0

–15.4

–15.6

–15.8

–16.2–16.4

–16.6

–16.8

–17.0

–17.2

–17.4

–17.6

–17.8

–18.0

CMF2(Proposed)

CMF1

61 m

92 m

Mulberry mine

92m

122 m

153

m

Prosperquarry

302010

40

50

60

0 300 m

N

FIG. 5.18 Central Mining and Finance Program: regional gravity survey and deep drillholes.

Page 115: Introduction to mineral exploration

98 C.J. MOON & M.K.G. WHATELEY

CMF7

CMF2CMF4

CMF6

CMF8 CMF1

IGS1

IGS2

IGS3

CMF3

CMF10 CMF9

CMF5

153

m

122 m

92m

Adit

61 m

92 m

Mulberry mine

Prosperquarry

302010

40

50

60

0 300 m

Elvan

Meadfoot BedsCalc flinta

>200 ppm Cu

Deep overburdenanomaly

N

FIG. 5.19 Central Mining and Finance Program 1980–82: drilling to test deep overburden anomalies. TheBritish Geological Survey drillholes to test the Wheal Prosper system (IGS 1–3) are also shown.

Page 116: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 99

south of the Mulberry trend. This defined twoparallel mineralized zones with encouragingassays in bedrock. Their trends were testedwith five holes (CMF 2,4,6–8, Fig. 5.19), whichdemonstrated that the mineralized zones aresteeply dipping veins. Although some highergrade intersections were made the overall tenordid not appear to be viable, especially whenthe mineralogy was carefully examined. Labor-atory mineral processing trials of the highergrade intersections failed to recover sufficientcassiterite to account for the total tin contentdetermined. It therefore seemed likely thatsome tin was contained in silicates, probablypartly in the tin garnet, malayaite. This tin isnot recoverable during conventional mineralprocessing and considerably devalued these in-tersections. Two holes were also drilled to testthe subsurface extension of the Wheal Prosperpit (CMF 5) and the major copper anomaly(CMF 3). Hole CMF 5 confirmed that surfacegrades extended to depth but the grade was toolow to be of further interest. By contrast, CMF3 was more encouraging with several intersec-tions, including 1.5 m at 0.5% Sn at 95 m.

The encouragement of Phase 3 in the south ofthe area led to a Phase 4 (1982) program whichfurther examined the large copper anomalyat the intersection of the Wheal Prosper andMulberry trends. Two holes were drilled(CMF9 and CMF10) and although one of thesecut high grade mineralisation (0.85 m of 7.0%Sn at 15.9 m) the program was terminated.

5.3.1 Valuation of prospects

The valuation of mineral properties at theexploration stage is not straightforward. If a re-source is defined then the methods discussed inChapter 11 can be used. If however the propertyis an early stage of exploration then valuationis more subjective.

Very little has been published formally onvaluation of early stage exploration properties,although a useful starting point is the shortcourse volume published by the Prospectorsand Developers Association of Canada (PDAC1998). In a review of Australian practice, Law-rence (1998) cites a number of methods:1 Comparable transaction method. In thismethod the value of similar transactions iscompared.

2 Multiple of exploration spending. The valueof the prospect is some multiple (usually be-tween 0.5 and 3) of the amount that has beenspent on the property.3 Joint venture comparison. A value isassigned based on the value of a joint venture ofthe property or a similar one when the jointventure does not involve a company associatedwith the holding company.4 Replacement value. In this method the valueis a multiple of the cost of acquiring and main-taining the property, e.g. taxes and peggingcosts, corrected for inflation.5 Geoscience rating method. The value isbased on a series of geological parameters suchas alteration, width, and grade of any mineral-ized intersections and proximity to depositsusing weightings. The basis for one method ofweightings is discussed in Kilburn (1990).

In a practical example Ward and Lawrence(1998) compared these approaches and pre-ferred a comparable transaction approach. Theyattempted to value a block of ground propect-ive for gold in northern Ireland based on aproprietary database of European and globaltransactions. This gave valuations between$US2.1M and $US15.2M. However the mostsatisfactory estimate was based on the valua-tion of a similar property in Scotland at$US222 ha−1. When this was applied to north-ern Ireland it gave a value of $US9.2M. Theyalso cited examples from Alaska and New-foundland, which showed an inverse relationbetween property size and value (Fig. 5.20) but

00

Size of properties (ha)

10,0

00

20,0

00

30,0

00

40,0

00

50,0

00

60,0

00

70,0

00

100

200

300

400

500

600

Valu

e pe

r ha

($)

FIG. 5.20 Relation between property size and valuefor 26 grassroots properties, Labrador in 1995–97.(After Ward & Lawrence 1998.)

Page 117: Introduction to mineral exploration

100 C.J. MOON & M.K.G. WHATELEY

a strong positive correlation of value withdecreasing distance from major deposits, in thelatter case, the Voisey’s Bay nickel deposit.

The reader should note the impact of explora-tion fads. When the Bre-X frenzy (discussed indetail below) was at its height the number oftransactions of gold properties in Indonesiaincreased in number and value by a factor of 20(Ward & Lawrence 1998). Even large companiescan be sucked into paying a premium for prop-erties, particularly when the property has alarge resource. A good example is the Voisey’sBay nickel deposit which was bought in 1996by Inco for $C4.3 billion from junior companyDiamond Fields Resources after a bidding warwith Inco’s great rival Falconbridge (McNish1999). In 2002 Inco was forced to take a $C1.55billion write off of the value of the projectwhich should eventually come into productionin 2006.

5.4 THE BRE-X MINERALS SAGA

The events surrounding the rise and fall ofBre-X Minerals Ltd provide a good example ofsome the pitfalls surrounding mineral explora-tion and the need to audit data and informationat every stage. Good accounts can be foundin popular books by Goold and Willis (1997)and Wells (1999), as well as a more technicalsummary in Hefferman (1998).

The company was founded in 1989 by DavidWalsh, a stock trader in Calgary, Alberta, toexplore for diamonds in the wake of theDiamet discovery (Chapter 17.2). In 1993, afterbeing personally bankrupted, he spent someof his last $C10,000 on a visit to Jakarta,Indonesia, to meet an exploration geologist,John Felderhof, whom he knew to be active ingold exploration. Felderhof, who was a mem-ber of the team that had discovered the OkTedi deposit for Kennecott Copper in 1968, pro-moted a theory that gold in the Indonesianprovince of Kalimantan was related to volcanicdiatremes. This was plausible as the areaincluded significant gold deposits at MountMouro and Kelian. Felderhof and a Filipinogeologist, Michael de Guzman, were hired bya Scottish-based company, Waverley MiningFinance, to investigate some properties thatthey had bought from Australian companies.

Amongst these was a prospect called Busangthat had been investigated by drilling in 1988and 1989, following its discovery by regionalreconnaissance. The prospect was consideredto report erratic and spotty gold. FollowingFelderhof’s recommendation, Bre-X managedto raise $C80,100 to buy an option on the cen-tral 15,060 ha claim at Busang. The first twoholes in the central block were barren but thethird hole was reported to contain assays of upto 6.6 g t−1 Au.

In order to fund further exploration, Bre-Xsold shares through a Toronto-based broker andby May 1994 had raised $C5.4M. Spectaculardrill intersections were then announced and aresource of 10.3 Mt at 2.9 g t−1 Au for the Cen-tral Zone in the beginning of 1995. However,these announcements were eclipsed by thediscovery of the Southeast zone and two sensa-tional intersections in October 1995 of 301 mgrading 4.4 g t−1 Au and 137 m grading 5.7 g t−1.These results and the announcement by Walshand Felderhof that the resources were of theorder of 30–40 million ounces of gold drove theshare price to $C286 per share and a companyvalue of over $C6 billion (Fig. 5.21). Bre-X thenhired a respected Canadian-based firm of con-sultants to calculate the resource and they con-firmed Bre-X’s estimate based on the core logsand assays provided by Bre-X of 47M oz of goldand an in situ value of $C16 billion.

It was at this stage that Indonesian polit-icians became involved, both in the confirma-tion of Bre-X’s exploration licence for theSoutheast zone and the business of finding amajor mining company partner to develop thedeposit. After a number of offers from twomajor Canadian companies, Bre-X decidedafter pressure from the Indonesian governmentover the licence renewal to enter a joint ven-ture. The partners were to be the Indonesiangovernment (10%), companies of MohammedHasan (an Indonesian friend of PresidentSuharto) with 20%, and Freeport McMoRan, anAmerican company that runs the Grasbergmine in Indonesia with 15%. As part of thedeal Freeport was to invest $US400 million andimmediately began a program to confirm thevalidity of the resource data, a process knownas due diligence. Freeport drilled seven holes atBusang in an attempt to confirm Bre-X’s drill-ing results, both by drilling four holes, collared

Page 118: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 101

19890

Sha

re p

rice

(US

$)

5

10

15

20

25

30

1990 1991 1992 1993 1994 1995 1996 19970

Share price (U

S$)

5

10

15

20

25

30

Bre-X began trading in July 1989 on the Alberta Stock Exchange

On Wednesday, March 26, 1997, North American stock exchanges haltedtrading in Bre-X shares at $15.50 after the company admitted a “strongpossibility” that estimates of how much gold it found in Borneo had beenoverstated. On March 27, the shares reopened at $3.50

The chart reflects a 10-for-1 split on May 22, 1996. The prices are thoseon the Alberta Stock Exchange until April 12, 1996, and on the Toronto orMontreal exchanges thereafter

FIG. 5.21 Share price of Bre-X Minerals. (From Goold & Willis 1998. Source: stock prices from Datastream.)

Mineral Services, to undertake a technicalaudit of Busang. Strathcona mobilized a teamled one of its partners, Graham Farquharson,to go to Jakarta and Busang. After arrival inJakarta the Strathcona team reviewed theFreeport data and advised Bre-X to make anannouncement of the problem on the Torontostock exchange, on which Bre-X was nowlisted. The announcement said “there appearsto be a strong possibility that the potentialgold resources on the Busang project in EastKalimantan, Indonesia have been overstatedbecause of invalid samples and assaying ofthose samples,” This announcement wascoupled with a 1-day suspension of trading ofBre-X at the end of which the share pricedropped from $C15.80 (there had been a 1:10split) to $C2.50, a loss of value in the companyof approximately $C3 billion (Fig. 5.21). TheStrathcona investigation which included drill-ing and assaying in Perth, Australia of 1470 mof core took a further month. At the end of this,Freeeport’s results were confirmed. As thelaboratory used by Bre-X was cleared, the onlyexplanation was that gold was added to thesamples after crushing and before analysis, aprocess generally known as salting. Furtherinvestigations suggested that this was probablycarried out by at least one of the site geologists,

2 m from high-grade Bre-X holes, two holesperpendicular to Bre-X holes, and one to infillbetween Bre-X holes. Samples were sent to fourlaboratories, the laboratory used by Bre-X, twoin the Indonesian capital Jakarta, and one inNew Orleans. In addition, the gold from thecores drilled by Freeport and from samples pro-vided by Bre-X were also selected for SEMexamination. Unusually Bre-X had assayed thewhole core and not kept a portion as a record.The assay results from all laboratories showedlow and erratic gold concentrations in theFreeport cores causing consternation to theFreeport geologists and management. The goldfound in the Freeport cores was very fine andunlike the coarse rounded gold grains found inthe Bre-X samples. As the due diligence pro-gram was bound by conditions of confidential-ity, Freeport reported the results to Bre-X andwaited for an explanation. Before Bre-X wasable to investigate this claim, its site geologist,Michael de Guzman, fell or jumped from ahelicopter on his way back to the Busang sitefrom the Prospectors and Developers Associa-tion Conference in Toronto where Felderhofhad been named Prospector of the Year.

On hearing the news from Freeport,Bre-X executives in Toronto immediately hireda major Canadian consultancy, Strathcona

Page 119: Introduction to mineral exploration

102 C.J. MOON & M.K.G. WHATELEY

including Michael de Guzman. The gold usedappeared to have been alluvial and had beenbought from local small scale miners. In thiscase, the salting had been difficult to detect asthe gold had been added to sample interval thatcorrelated with alteration and it appeared thefraudsters had used the core logs to determinewhich samples to salt. The only person to facejustice was Felderhof who allegedly sold $C83million in shares before the collapse and wentto live in the Cayman Islands, a country withno extradition treaty with Canada. He was(2004) on trial for insider trading, alleging thathe sold substantial shares knowing that theBre-X ownership over Busang was in doubt.The major questions to be addressed in thewake of the investigation were how such afraud could be detected and how further fraudsof this nature could be prevented. There were anumber of irregular procedures at Busang thatshould have been detected:1 The whole core was assayed and no part waskept for a record.2 The gold grains reported were more consist-ent with an alluvial derivation than from abedrock source.3 Pre-feasibility testing indicated a very highrecovery by gravity methods – very unusual forepithermal gold deposits.4 The reported occurrence of near surfacesulfide in a deeply weathered environment.5 The lack of alluvial gold in the stream drain-ing the Southeast zone.6 The wide spaced drilling on the Southeastzone (250 by 50 m) that was much larger thannormal.

The result of the Bre-X fraud has been muchstricter enforcement of disclosure of ore re-serve calculations methods and the develop-ment of formal codes such as the JORC code(see section 10.4). The Bre-X affair also markedthe end of the exploration boom of the late1990s and many investors refused to investin junior mining stocks, although some madesimilar mistakes in the subsequent dotcomboom.

5.5 SUMMARY

Following the identification of areas withpossible mineralisation and the acquisition of

land, the next step is to generate targets fordrilling or other physical examination suchas pitting and trenching or, in mountainousplaces, by driving adits into the mountainside.Budgeting and the organization of suitable staffare again involved.

If the area is remote then suitable topo-graphical maps may not be available and theground must be surveyed to provide a base mapfor geological, geochemical, and geophysicalwork. Some of the geological investigationsmay include the mapping and sampling of oldmines and the dispatch of samples for assaying.Any of this work may locate areas of minera-lisation which may be recognized by topogra-phical effects, coloring, or wall rock alteration.Different ore minerals may give rise to char-acteristic color stains and relic textures inweathered rocks and gossans and below thesesupergene enrichment may have occurred. Allthese investigations together with pitting andtrenching may indicate some promising drill-ing targets.

Geologists should never rush into drillingwhich is a most expensive undertaking! Verycareful planning is necessary (section 5.2.1) anddrilling programs require meticulous monitor-ing. Once the drilling is in full swing decidingwhen to stop may be very difficult and variouspossibilities are reviewed in section 5.2.3.

If an exploration program does not resolvethe potential of some mineralized ground thenthe operating company may sell the option,with or without the data obtained on it, to anew investor. This is often nowadays termedrecycling and it enables the original operator torecover some of its investment. Such recyclingis exemplified by the Mulberry Prospect inCornwall (section 5.3). The buying and sellingof prospects can be fraught with problems, par-ticularly of valuation (section 5.3.1) and alsothe possibility of fraud (section 5.4).

5.6 FURTHER READING

Basic field techniques are covered in Reedman(1979), Peters (1987), Chaussier and Morer(1987), and Majoribanks (1997). Reviews ofgeochemical, geophysical and remote sensingtechniques, covered in Chapters 6–8, as wellas some case histories of their integration,

Page 120: Introduction to mineral exploration

5: FROM PROSPECT TO PREFEASIBILITY 103

(1995, 2000) that of the Pacific Rim area. Thevolume of Goldfarb and Neilsen (2002) pro-vides an update of exploration in the 1990sand early 2000s. Evaluation techniques fromexploration to production are well covered byAnnels (1991).

can be found in the volume edited by Gubins(1997).

Case histories of mineral exploration arescattered through the literature but the volumeof Glasson and Rattigan (1990) compiled muchAustralian experience and those of Sillitoe

Page 121: Introduction to mineral exploration

104 M.K.G. WHATELEY

6Remote Sensing

Michael K.G. Whateley

6.1 INTRODUCTION

Geologists have been using aerial photo-graphy to help their exploration efforts fordecades. Since the advent of satellite imagerywith the launch of the first earth resourcessatellite (Landsat 1) in 1972, exploration geolo-gists are increasingly involved in interpretingdigital images (computerized data) of the ter-rain. Recent technological advances now pro-vide high resolution multispectral satellite andairborne digital data. More recently, geologistsinvolved in research and commercial explora-tion have been seeking out the more elusivepotential mineral deposits, e.g. those hiddenby vegetation or by Quaternary cover. Usu-ally geochemical, geophysical and other mapdata are available. It is now possible to expressthese map data as digital images, allowing thegeologist to manipulate and combine themusing digital image processing software, suchas Erdas Imagine, ERMapper TNT MIPS, andgeographical information systems (GIS). Theselatter techniques are discussed in detail insection 9.2.

As image interpretation and photogeologyare commonly used in exploration programstoday, it is the intention in this chapter todescribe a typical satellite system and explainhow the digital images can be processed, inter-preted, and used in an exploration program toselect targets. More detailed photogeologicalstudies using aerial photographs or high resolu-tion satellite images are then carried out on thetarget areas.

6.1.1 Data collection

Remote sensing is the collection of informa-tion about an object or area without being inphysical contact with it. Data gathering sys-tems used in remote sensing are:1 photographs obtained from manned spaceflights or airborne cameras, and2 electronic scanners or sensors such as multi-spectral scanners in satellites or aeroplanes andTV cameras, all of which record data digitally.

Most people are familiar with the weathersatellite data shown on national and regionaltelevision. For most geologists and other earthscientists, multispectral imagery is synonym-ous with NASA’s Landsat series. It is imagesfrom these satellites that are most readilyavailable to exploration geologists and theyare discussed below. The use of imagery anddigital data from multispectral sources suchas NASA’s flagship satellite, Terra, or commer-cial satellites such as QuickBird and Ikonos,are becoming more widespread. Remote sens-ing data gathering systems are divided intotwo fundamental types, i.e. those with passiveor active sensors.

Passive sensors

These sensors gather data using availablereflected or transmitted parts of the electro-magnetic (EM) spectrum, i.e. they rely on solarillumination of the ground or natural thermalradiation for their source of energy respect-ively. Some examples are:

Page 122: Introduction to mineral exploration

6: REMOTE SENSING 105

100

10

1

10 m 100 m 1 km 10 km

SPOT

LANDSAT

TM MSS

TIROS-N

METEOSAT

Sun synchronous orbits

Geostationary orbits

Freq

uenc

y of

repe

titiv

e co

ver (

days

)

Pixel size

Aster

Terra

FIG. 6.1 Different resolution andfrequencies of some satelliteimagery.

Active sensors

These sensors use their own source of energy.They emit energy and measure the intensity ofenergy reflected by a target. Some examples areRadar (microwave) and Lasers (European SpaceAgency 2004).

6.2 THE LANDSAT SYSTEM

The first Landsat satellite, originally calledEarth Resources Technology Satellite (ERTS),was launched in July 1972. Seven satellites inthe series have been launched (Table 6.1).Landsat has been superseded by the Terrasatellite, hosting the ASTER sensors (Abrams& Hook 2002).

Each satellite is solar powered and has a datacollecting system which transmits data to thehome station (Fig. 6.2). On Landsat 1 to 3 datawere recorded on tape when out of range of thehome station for later transmission. Trackingand data relay system satellites (TDRSS) wereoperational from March 1984 (Fig. 6.2). Landsat7 has only a direct downlink capability withrecorders.

1 Landsat Multispectral Scanner (MSS);2 Landsat Thematic Mapper (TM) utilizesadditional wavelengths, and has superior spec-tral and spatial resolution compared with MSSimages;3 Advanced Spaceborne Thermal Emission andReflection Radiometer (ASTER) on NASA’ssatellite Terra;4 SPOT, a French commercial satellite withstereoscopic capabilities;5 high resolution commercial satellites, suchas Ikonos (Infoterra 2004), the first of the nextgeneration of high spatial resolution satellites,and QuickBird (Ball Aerospace & TechnologiesCorp 2002);6 Space Shuttle;7 airborne scanning systems that have evengreater resolution and can look at more andnarrower wavebands.

Different satellites collect different data withvarious degrees of resolution (Fig. 6.1) depend-ing on their application. For geological ap-plications resolution of the order of metersof Landsat is required in contrast to the tensof kilometers of NOAA (National Oceanic andAtmospheric Administration) weather satellites.

Page 123: Introduction to mineral exploration

106 M.K.G. WHATELEY

TA

BLE

6.1

Ch

arac

teri

stic

s of

sev

eral

rem

ote

sen

sin

g sy

stem

s.

Lau

nch

Alt

itu

de (k

m)

Rep

eat

cove

rage

(day

s)

Spec

tral

ban

ds (µ

m)

IFO

V (m

)

Scen

e di

men

sion

s (k

m)

Pix

els

per

scen

e (×

106 )

* T

he

MSS

ban

ds o

n L

ands

ats

1–3

wer

e n

um

bere

d 4,

5, 6

, an

d 7

beca

use

of a

th

ree-

ban

d R

BV

sen

sor

on L

ands

ats

1 an

d 2.

P, p

anch

rom

atic

mod

e.

Land

sat M

SS

1 Ju

ly 1

972

–Jan

197

82

Jan

197

5–Ju

ly 1

983

3 M

ar 1

978–

Sept

198

34

July

198

2–J

uly

198

75

Mar

198

4–p

rese

nt

6 O

ct 1

993–

Oct

199

3

918

(1–3

)

18 4*0.

50–0

.60

50.

60–0

.70

60.

70–0

.80

70.

80–1

.10

79×

79 (1

–3)

82×

82 (4

–5)

185

×18

5

28

Enha

nced

TM

6 O

ct 1

993–

Oct

199

37

Apr

il 1

999–

pres

ent

705

16 10.

45–0

.551

20.

525–

0.60

53

0.63

0–0

.690

40.

750

–0.9

005

1.55

0–1

.750

610

.40

–12.

57

2.09

–2.3

5P

0.52

–0.9

0

15×

15 (P

)30

×30

(ban

ds 1

–5, 7

)60

×60

(ban

d 6)

183

×17

5

231

AST

ER

1 D

ec 1

999–

pres

ent

705

10.

52–0

.69

20.

63–0

.69

30.

78–0

.86

51.

60–1

.70

62.

145–

2.18

57

2.18

5–2.

225

82.

295–

2.36

59

2.36

0–2

.430

108.

125

–8.4

7511

8.47

5–8

.825

128.

925

–9.2

7513

10.2

5–1

0.95

1410

.95

–11.

65

15×

15 (b

ands

1–3

)30

×30

(ban

ds 4

–9)

90×

90 (b

ands

10

–14)

60×

60

21

The

mat

ic m

appe

r(T

M)

4 Ju

ly 1

982

–Ju

ne

2001

5 M

ar 1

984

–pre

sen

t

705

16 10.

45–0

.53

20.

52–0

.60

30.

63–0

.69

40.

76–0

.90

51.

55–1

.75

610

.40

–12.

57

2.08

–2.3

5

30×

30 (b

ands

1–5

, 7)

120

×12

0 (b

and

6)

185

×18

5

231

SPO

T/H

RV

1 Ja

n 1

986

–Dec

199

32

Jan

199

0–D

ec 1

997

3 Se

pt 1

993–

1998

4 19

97–2

001

5 19

98–P

rese

nt

822

26 10.

50–0

.59

20.

61–0

.68

30.

79–0

.89

41.

58–1

.75

(SP

OT

4)

P0.

51–0

.73

20×

20 (b

ands

1–3

)10

×10

(P)

60×

60

27 (b

ands

1–3

)

Page 124: Introduction to mineral exploration

6: REMOTE SENSING 107

SCANNER OPTICS

6 Detectors per band

Groundreceivingstation

Data storage

Oscillatingscan mirror

4 GR5 RED6 IR7 IR

918

km A

ltitu

de

185 km

Activescan

Field of view = 11.56°

185 km

6 Lines/scan/band

Path ofspacecraft travel

TDRS

North

South

EastWest

FIG. 6.2 Schema of a satellite system. MSS data are collected on Landsats 1–3 in four bands and transmittedvia the TDRSS (relay satellite) to the ground receiving station. (Modified after Lo 1986, Sabins 1997.)

violet, and blue light are particularly affectedby scattering within the atmosphere and areof little use in geology. Blue light is onlyusefully captured using airborne systems.The blanks in the spectrum are due mainlyto absorption by water vapor, CO2 and O3

(Fig. 6.3). Only wavelengths greater than 3 µm,e.g. microwave wavelengths (or radar), canpenetrate cloud (Fig. 6.3), so radar imageryis particularly useful in areas of considerablecloud cover (e.g. equatorial regions).

6.2.1 Characteristics of digital images

MSS data are recorded by a set of six detectorsfor each spectral band (Table 6.1). Six scan linesare simultaneously generated for each of thefour spectral bands (Fig. 6.3). Data are recordedin a series of sweeps and recorded only onthe eastbound sweep (Fig. 6.2). TM data arerecorded on both east- and westbound sweepsand each TM band uses an array of 16 detectors(band 6 uses only four detectors). There iscontinuous data collection and the data are

The Landsat series was designed initially toprovide multispectral imagery for the studyof renewable and nonrenewable resources, par-ticularly land use resources using MSS andReturn Beam Vidicon (RBV) cameras. RBV im-ages are not commonly used today. Geologistsimmediately recognized the geological poten-tial of the Landsat images and in July 1982when Landsat 4 was launched it housed, inaddition to MSS, a Thematic Mapper whichscans in seven bands, two of which (5 and 7)were chosen specifically for their geologicalapplicability (Table 6.1 & Fig. 6.6). Landsats 5and 7 also have TM. Landsat 6 was lost atlaunch. The additional features of the ASTERsensors are their high spatial and radiometricresolution over similar parts of the spectrum asTM, broad spectral coverage (visible- throughthermal-infrared), and stereo capability on asingle path.

Not all wavelengths are available for remotesensing (Fig. 6.3). Even on an apparently clearday ultraviolet (UV) light is substantially ab-sorbed by ozone in the upper atmosphere. UV,

Page 125: Introduction to mineral exploration

108 M.K.G. WHATELEY

H2OCO2H2O H2O

Wavelength 1 nm 1 mm 1 mm 1 cm 1 m

GAMMARAY

X - RAY ULTRA-VIOLET

VISIBLE

MICROWAVEINFRARED RADIO

Ref

lect

ed I

R

The

rmal

IR

(a)

Aircraft(typical)

Seasat

SIR A & B

Normal color photos

IR color photos

AircraftIR scanners

Landsat thematic mapperIMAGINGSYSTEMS 1 2 3 4 5 67

Landsat multispectral scanner SatelliteIR scanners4 5 6 7

(b)

30

Wavelength (mm)

0.5 1.0 1.5 2.0 3.0 4.0 5.0 10 15 200

100

Atm

osph

eric

tran

smis

sion

%

Pho

to.

UV

Blu

e

Gre

en

Red

Pho

to.

IR

1.6

µm

2.2

µm

3–5

µm

8–14

µm

H2O

O3O3

Reflected IR Thermal IR

CO2

INFRAREDVISIBLEULTRA -VIOLET

80

60

40

20

CO2

10–12 10–11 10–10 10–9 10–8 10–7 10–6 10–5 10–4 10–3 10–2 10–1 100 101

Aster

Expandedfigure (b)

10–12 13–141 2 3 4 5–9

FIG. 6.3 (a) The electromagnetic spectrum. (b) Expanded portion of the spectrum showing the majorityof the data-gathering wavelengths. Additional data are acquired in the microwave part of the spectrumfor use by aircraft and to measure the surface roughness of either wave action or ice accumulation. Certainparts of the spectrum are wholly or partly absorbed by atmospheric gases. Atmospheric windows wheretransmission occurs are shown and the imaging systems that use these wavelengths are indicated.(Modified after Sabins 1997.)

Page 126: Introduction to mineral exploration

6: REMOTE SENSING 109

transmitted to earth for storage on ComputerCompatible Tapes (CCTs). These tapes are thenprocessed on a computer to produce images.The width of scene dimensions on the surfaceunder the satellite (its swath) is 185 km. Thedata are (for convenience) divided into sets thatequal 185 km along its path.

The satellites have a polar, sun synchronousorbit, and the scanners only record on thesouthbound path (Fig. 6.2) because it is night-time on the northbound path. The paths ofLandsats 1 to 3 shifted west by 160 km at theequator each day so that every 18 days thepaths repeat resulting in repetitive, worldwide,MSS coverage. Landsats 4, 5 and 7 and the Terrasatellite have a slightly different orbit result-ing in a revisit frequency of 16 days. Imagesare collected at the same local time on eachpass – generally between 9.30 and 10.30 a.m.to ensure similar illumination conditions onadjacent tracks. Successive paths overlap by34% at 40°N and 14% at the equator. Landsatdoes not give stereoscopic coverage, althoughit can be added digitally. ASTER does givestereoscopic coverage.

6.2.2 Pixel parameters

Digital images consist of discrete picture ele-ments, or pixels. Associated with each pixel isa number that is the average radiance, or bright-

ness, of that very small area within the scene.Satellite scanners look through the atmosphereat the earth’s surface, and the sensor thereforemeasures the reflected radiation from the sur-face and radiation scattered by the atmosphere.The pixel value represents both. Fortunately,the level of scattered radiation is nearly con-stant, so changes in pixel value are essenti-ally caused by changes in the radiance of thesurface.

The Instantaneous Field of View (IFOV) isthe distance between consecutive measure-ments of pixel radiance (79 m × 79 m in MSS),which is commonly the same as the pixel size(Table 6.1). However, MSS scan lines overlap,hence the pixel interval is less than the IFOV(56 m × 79 m). The TM and ASTER detectorshave a greatly improved spatial resolution,giving an IFOV of 30 m × 30 m.

The radiance for each pixel is quantified intodiscrete gray levels and a finite number of bitsare used to represent these data (Table 6.2).Table 6.2 shows how decimal numbers in therange 0–255 can be coded using eight indi-vidual bits (0 or 1) of data. Each is grouped as an8 bit “byte” of information, with each bit usedto indicate ascending powers of 2 from 20 (=1) to27 (=128). This scheme enables us to code just256 (0–255) values. On the display terminal,256 different brightness levels are used, cor-responding to different shades of gray ranging

TABLE 6.2 The principle of binary coding using 8 bit binary digits (bits).

Place values

Binary 27 26 25 24 23 22 21 20

Decimal 128 64 32 16 8 4 2 1

0 0 0 0 0 0 0 0 = 00 0 0 0 0 0 0 1 = 10 0 0 0 0 0 1 0 = 20 0 0 0 0 1 0 0 = 40 0 0 0 1 0 0 0 = 80 0 0 1 0 0 0 0 = 160 0 1 0 0 0 0 0 = 320 1 0 0 0 0 0 0 = 641 0 0 0 0 0 0 0 = 1281 0 0 0 0 1 0 1 = 128 + 4 + 1

= 1331 1 1 1 1 1 1 1 = 128 + 64 + 32 + 16 + 8 + 4 + 2 + 1

= 255

Page 127: Introduction to mineral exploration

110 M.K.G. WHATELEY

TABLE 6.3 False colors assigned to Landsat MSS bands.

MSS band Wavelength (µµµµµm) Natural color recorded False color assigned

4 0.5–0.6 Green Blue5 0.6–0.7 Red Green6 0.7–0.8 IR Red7 0.8–1.1 IR Red

from black (0) to white (255), giving us a grayscale. Commonly 6 bits per pixel (64 graylevels) are used for MSS and 8 bits per pixel (256gray levels) for TM and ASTER. A pixel cantherefore be located in the image by an x andy co-ordinate, while the z value defines thegray-scale value between 0 and 255. Pixel sizedetermines spatial resolution whilst the radi-ance quantization affects radiometric resolu-tion (Schowengerdt 1983).

6.2.3 Image parameters

An image is built up of a series of rows andcolumns of pixels. Rows of pixels multiplied bythe number of columns in a typical MSS image(3240 × 2340 respectively) gives approximately7M pixels per image. The improved resolu-tion and larger number of channels scannedby the TM results in nine times as many pixelsper scene, hence the need for computers toanalyze the data. Remote sensing images arecommonly multispectral, e.g. the same sceneis imaged simultaneously in several spectralbands. Landsat MSS records four spectral bands(Table 6.3), TM records seven, and ASTERrecords 14 spectral bands. Visible blue light isnot recorded by Landsat MSS or ASTER, but itis recorded in TM band 1. By combining threespectral bands and assigning a color to eachband (so that the gray scale becomes a red,green, or blue scale instead), a multispectralfalse color composite image is produced.

The image intensity level histogram is a use-ful indicator of image quality (Fig. 6.4). Thehistogram describes the statistical distributionof intensity levels or gray levels in an image interms of the number of pixels (or percentage ofthe total number of pixels) having each graylevel. Figure 6.4 shows the general character-istics of histograms for a variety of images.

A photographic print of a geometricallycorrected image (Fig. 6.5) is a parallelogram.The earth rotates during the time it takesthe satellite to scan a 185 km length of itsswath, resulting in the skewed image (Fig.6.11f).

6.2.4 Availability of Landsat data

1 Landsat CCTs or CDs of MSS, TM, orASTER imagery are available for computerprocessing (Box 6.1).2 Black and white, single band prints in astandard 23 cm × 23 cm format at a scale of1:1000,000 are available from the abovesources.3 Color composite prints in a similar formatare also available. It is possible to enlargeLandsat MSS images up to 1:100,000, before thepicture quality degrades severely.

6.2.5 Mosaics

Each satellite image has a uniform scale andrelative lack of distortion at the edges. Blackand white or color prints of each scene can becut and spliced into mosaics. With the repetit-ive cover most of the world has cloud-freeLandsat imagery available. The biggest diffi-culty is in matching the gray tone during print-ing, and the images from different seasons.Scenes can be merged digitally to alleviate graytone differences.

6.2.6 Digital image processing

Data from the satellites are collected at theground station on magnetic tape, i.e. in a digitalform. Data are loaded into image processingcomputer systems from CCTs or CDs. Themain functions of these systems in a geological

Page 128: Introduction to mineral exploration

6: REMOTE SENSING 111

FIG. 6.4 Histograms of the gray levels (intensity levels) of three different types of image. The image is partof Landsat 5, TM band 3, taken on 26 April 1984 of the High Peak District of northern England. The areais approximately 15 × 15 km. The dark areas are water bodies; those in the the upper left corner are theLongdale reservoirs while those in the right center and lower right are the Derwent Valley reservoirs.The two E–W trending bright areas in the lower right of the images are the lowland valleys of Edale andCastleton. (From Mather 1987.) (a) A bright, low intensity image histogram with low spectral resolution.(b) A dark, low intensity, low spectral resolution, image histogram. (c) A histogram with high spectralresolution but low reflectance values. Increased spectral resolution has been achieved with a simplehistogram transformation (also known as a histogram stretch). (Modified after Schowengerdt 1983.)

Input gray level 0 30 170 255

(b)

Transformed gray level

Ref

lect

ance

%

Tran

sfor

med

gra

y le

vel

(c)

30 170 0 255Transformed gray level

Ref

lect

ance

%

Input gray level

Tran

sfor

med

gra

y le

vel

Transformed gray level

Ref

lect

ance

%

0 255

(a)

Page 129: Introduction to mineral exploration

112 M.K.G. WHATELEY

FIG. 6.5 A black and white Landsat 1 (ERTS), band 7 photograph of part of the Bushveld Igneous Complex inSouth Africa. This summer image (wet season) illustrates the strong geological control over vegetation andvariations in soil moisture. The circular Pilanesberg Alkali Complex is seen in the north with the layeredbasic and ultra-basic intrusion to the southeast of it (dark tone). The banding in the Transvaal Supergroupdolomites, in the south center (medium gray tone) was unsuspected until viewed on Landsat photographs.

Page 130: Introduction to mineral exploration

6: REMOTE SENSING 113

BOX 6.1 Some sources of remotely sensed data.

• US Geological Survey 1981. Worldwide directory of national earth science agencies and relatedinternational organizations. Geological Survey Circular 834, USGS, 507 National Center, Reston,Virginia, 22092, USA. http://www.usgs.gov and http://glovis.usgs.gov/• SPOT-Image, 18 avenue Eduoard Belin, F31055, Toulouse Cedex, France. http://www.spot.com• Meteorological satellite imagery.US Department of Commerce, NOAA/NESDIS/NSDC, Satellite Data Services Division (E/CCGI),World Weather Building, Room 10, Washington DC, 20233, USA http://db.aoml.noaa.gov/dbweb/• Landsat (pre-September 1985).US Geological Survey, EROS Data Center (EDC), Mundt Federal Building, Sioux Falls, South Dakota,57198, USA• Landsat (post September 1985).Earth Observation Satellite Company (EOSAT), 1901 North Moore Street, Arlington, Virginia, 22209,USA. http://edcimswww.cr.usgs.gov/pub/imswelcome/• Landsat imagery is also available from local and national agencies and free from the GLCF.http://glcf.umiacs.umd.edu/intro/landsat7satellite.shtml• Infoterra, Europa House, The Crescent, Southwood, Farnborough, Hampshire, GU14 0NL, UK.http://www.infoterra-global.com

exploration program are: image restoration,image enhancement, and data extraction(Sabins 1997). Drury (2001) explains themethods of digital image processing in moredetail and shows how they can be applied togeological remote sensing.

Image restoration

Image restoration is the process of correctinginherent defects in the image caused duringdata collection. Some of the routines used tocorrect these defects are:1 replacing lost data, i.e. dropped scan lines orbad pixels;2 filtering out atmospheric noise;3 geometrical corrections.

The last named correct the data for carto-graphic projection, which is particularly im-portant if the imagery is to be integrated withgeophysical, topographical, or other map-baseddata.

Image enhancement

Image enhancement transforms the originaldata to improve the information content. Someof the routines used to enhance the images areas follows:

1 Contrast enhancement. An image histogramhas already been described in section 6.2.3(Fig. 6.4). A histogram of a typical untrans-formed image has low contrast (Fig. 6.4a) andin this case the input gray level is equivalentto the transformed gray level (Schowengerdt1983, Drury 2001). A simple linear transforma-tion, commonly called a contrast stretch, isroutinely used to increase the contrast of a dis-played image by expanding the original graylevel range to fill the dynamic range of thedisplay device (Fig. 6.4b).2 Spatial filtering is a technique used toenhance naturally occurring straight featuresuch as fractures, faults, joints, etc. Drury(2001) and Sabins (1997) described this in moredetail.3 Density slicing converts the continuousgray tone range into a series of density intervals(slices), each corresponding to a specific digitalrange. Each slice may be given a separate coloror line printer symbol. Density slicing hasbeen successfully used in mapping bathymetricvariations in shallow water and in mappingtemperature variations in the cooling water ofthermal power stations (Sabins 1997).4 False color composite images of three bands,e.g. MSS bands 4, 5, and 7, increase the amountof information available for interpretation.

Page 131: Introduction to mineral exploration

114 M.K.G. WHATELEY

Information extraction

Routines employed for information extractionuse the speed and decision-making capabilityof the computers to classify pixels on somepredetermined gray level. This is carried outinteractively on computers by band ratioing,multispectral classification, and principal com-ponent analysis. These can all be used toenhance specific geological features.1 Ratios are prepared by dividing the gray levelof a pixel in one band by that in another band.Ratios are important in helping to recognizeferruginous and limonitic cappings (gossans)(Fig. 6.6). Rocks and soils rich in iron oxidesand hydroxides absorb wavelengths less than0.55 µm and this is responsible for their strongred coloration (Drury 2001). These iron oxidesare often mixed with other minerals whichmask this coloration. A ratio of MSS band 4over band 5 will enhance the small contribu-

tion of iron minerals (Fig. 6.6). Similarly a ratioof MSS band 6 over band 7 will discrimin-ate areas of limonite alteration. Unfortunatelylimonite also occurs in unaltered sediment-ary and volcanic rocks. However, by ratioingTM bands, mineralogical spectral character-istics related to alteration can be detected.The Landsat TM scanner has two critical,additional bands (TM bands 5 and 7), whichare important in being used to identify hydro-thermally altered rocks. Absorption in the“clay band” causes low reflectance at 2.2 µm(TM band 7), but altered rocks have a highreflectance at 1.6 µm (TM band 5). A ratio ofbands 5/7 will result in enhancement of alteredrocks (Fig. 6.6). ASTER has even finer spectralbands in the “clay band” which gives enhancedalteration discrimination (Table 6.1).2 Multispectral classification. Using thisroutine a symbol or color is given to a pixelor small group of pixels which have a high

Water absorption bandIron and hydroxyl bearing soil or rock

Green vegetation

3

54 6 7

60

50

40

30

20

10

0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4

1 2 4 5 7

1 2 3 4 5–9

Multispectral scanner(MSS) Thematic mapper

(TM)

“CLAY BAND”“IRON BAND”

Ref

lect

ance

%

Wavelength (mm)

ASTER

FIG. 6.6 A reflectance profile of the visible and IR parts of the EM spectrum, showing the changing reflectanceprofile of soil associated with “typical” hydrothermal alteration compared with that of green vegetation.(Modified after Settle et al. 1984.)

Page 132: Introduction to mineral exploration

6: REMOTE SENSING 115

probability of representing the same kind ofsurface material, e.g. alteration. This deci-sion is best made by computers. Multispectralclassification is particularly useful for a largearea or a region covered by dense vegetation,as differences in the vegetation reflect theunderlying geology.3 Principal component analysis is used toenhance or distinguish lithological differences,because spectral differences between rock typesmay be more apparent in principal componentimages than in single bands. The reflectancesin different bands in MSS, TM, or ASTERimages have a high degree of correlation. Prin-cipal component analysis is a commonly usedmethod to improve the spread of data by redis-tributing them about another set of axes. Thiseffectively exaggerates differences in the data.

6.2.7 Interpretation

Two approaches are used to extract geologicalinformation from satellite imagery.1 Spectral approach. Spectral properties areused to separate units in image data based onspectral reflectance. This is done interactivelyon computers using multispectral data in areaswith or without dense vegetation.2 Photogeological approach. Known weather-ing and erosional characteristics (topographicalexpression) are used, on black and white orcolor prints, to imply the presence of geologicalstructure or lithology. Photogeological ele-ments include topography, erosion, tone, tex-ture, drainage pattern, vegetation, and land use.These elements are discussed in section 6.4.

6.2.8 Applications of Landsat satelliteimagery

The use of satellite imagery is now a standardtechnique in mineral exploration (Nash et al.1980, Goetz & Rowan 1981, Peters 1983, Drury2001). It has also been used in structural inves-tigations (Drury 1986) and in hydrogeology(Deutsch et al. 1981). In mineral explorationLandsat imagery has been used to provide basicgeological maps, to detect hydrothermal altera-tion associated with mineral deposits, and toproduce maps of regional and local fracturepatterns, which may have controlled miner-alisation or hydrocarbon accumulations.

Structural maps

The synoptic view afforded by MSS and TMimagery is ideal for regional structural ana-lyses, especially if the scene chosen is illu-minated with a low sun angle (e.g. autumn orspring) which emphasizes topographical fea-tures. Manual lineament analysis on overlaysof photographic prints is then carried out (usu-ally after a spatial filtering technique hasbeen employed) and lineaments are digitized.The original apparently random pattern canbe quantified by computer processing, thusproviding an objective method for evaluatinglineaments. A rose diagram is prepared todepict the strike–frequency distribution. At theHelvetia porphyry copper test site in Arizona,the three trends identified by these methodscorrespond to deformation events in the Pre-cambrian, the Palaeozoic, and the LaramideOrogeny (Abrams et al. 1984).

The most common structures seen onLandsat images are faults, fractures, linea-ments of uncertain nature (see Fig. 9.10), andcircular features. In conjunction with the stat-istical approach above, deductions regardingstress patterns in an area can be made. Thesestructural studies may yield clues to the loca-tion of concealed mineral deposits. Linearzones, such as the 38th Parallel Lineament ofthe USA (Heyl 1972), consist of weak crustalareas that have provided favorable sites forupwelling intrusives and mineralizing fluids.These areas contain faults or fractures asso-ciated with deep-seated basement structuresand the intersections of these features withother faults can provide favorable loci formineralisation.

Lithology and alteration maps

Landsat imagery has made significant con-tributions to the advancement of geologicalmapping both in known and unmapped areasof the world. Many third world countries nowhave geological maps which were previouslytoo expensive to produce by conventional fieldmapping. Today regional geological mappingoften starts with Landsat imagery, followedby rapid field reconnaissance for verifica-tion where possible. The maps thus producedare used to select exploration target zones on

Page 133: Introduction to mineral exploration

116 M.K.G. WHATELEY

the basis of favorable geology and structure.Detailed follow-up photogeology and fieldwork then takes place.

Ferruginous residual deposits (gossans)which overlie mineralized ground can often beidentified from color anomalies on enhancedfalse color ratio composites, whereas thesemay not be detected on standard imagery, e.g.at the Silver Bell porphyry copper site (Abramset al. 1984) and the Cuprite Mining District inNevada (Abrams & Hook 2002).

6.2.9 Advantages of Landsat imagery

1 This imagery provides a synoptic view oflarge areas of the Earth’s crust (185 km ×185 km), revealing structures previously un-recognized because of their great extent.2 The evolution of rapidly developing dyn-amic geological phenomena can be examinedthrough the use of successive images of thesame area produced at 16- to 26-day intervals,e.g. delta growth or glaciation.3 Some geological features are only inter-mittently visible, e.g. under certain conditionsof climate or vegetation cover. Landsat offers“revisit capability” and multitemporal cover-age (Viljoen et al. 1975).4 Digital images can be displayed in color.This is useful in mapping rock types and altera-tion products, either direct from rocks or fromvegetation changes (Abrams & Hook 2002).5 Landsat is valuable in providing a tool forrapid mapping of regional and local fracturesystems. These systems may have controlledore deposit location.6 It is very cost effective, with an outlay ofonly a few pence km−2 for map production. TheGlobal Land Cover Facility of University ofMaryland (GLCF 2004) offers the largest freesource of Landsat data.7 Computer processing enables discriminationand detection of specified rocks or areas.

6.3 OTHER IMAGING SYSTEMS

6.3.1 SPOT

The first French satellite, SPOT, was launchedin February 1986, followed by others withimproved technologies (Table 6.1). They are

more sophisticated than Landsat, having a highresolution visible (HRV) imaging system, witha 20 m (Spot 5, 10 m) ground resolution in MSmode and a 10 m (Spot 5, 5 m) resolution inpanchromatic mode (Table 6.1). SPOT also hasan off-nadir viewing capability which meansthat an area at 45°N can be imaged 11 times inthe 26-day orbital cycle. Repeated off-nadirviewing introduces parallax from which stereo-scopic image pairs can be produced. The com-mercial nature of SPOT means that imagesare slightly more expensive, but the increasedresolution and stereoscopic capability meanthat the additional cost may be warranted inan exploration program, although SPOT doesnot have as good a spectral range as TM.

6.3.2 Aster

NASA’s satellite, Terra, was launched inDecember 1999. Terra is the first of a seriesof multi-instrument spacecraft that are part ofNASA’s Earth Observing System (EOS). Theprogram comprises a science component and adata system supporting a coordinated seriesof polar-orbiting and low inclination satellitesfor long-term global observations of the landsurface, biosphere, solid Earth, atmosphere,and oceans. It houses the Advanced SpaceborneThermal Emission and Reflection Radiometer(ASTER) sensors. ASTER covers a wide spectralregion with 14 bands from the visible to thethermal infrared parts of the spectrum withhigh spatial, spectral, and radiometric resolu-tion (Abrams & Hook 2002). An additionalbackward-looking near-infrared band providesstereo coverage. The spatial resolution varieswith wavelength (Table 6.1).

6.3.3 High resolution satellite systems

The Ikonos-2 satellite was launched in Septem-ber 1999 and has been delivering commercialdata since early 2000 (Infoterra 2004). Ikonosis the first of the next generation of high spatialresolution satellites. Ikonos data records fourchannels of multispectral data (0.45–0.53,0.52–0.61, 0.64–0.72, and 0.77–0.88 µm) witha resolution of 4 m and one panchromaticchannel with 1 m resolution (0.45–0.90 µm).Ikonos radiometric resolution is far greaterthan the Landsat scenes, with data collected as

Page 134: Introduction to mineral exploration

6: REMOTE SENSING 117

11 bits per pixel (2048 gray tones). Specialistimage processing software is required to pro-cess these images. Ikonos instruments haveboth cross- and along-track viewing capabilit-ies, which allows frequent revisit capability,e.g. 3 days at 1 m resolution.

The high resolution imagery satelliteQuickBird was launched in October 2001(DigitalGlobe 2004). It provides a panchro-matic channel with 0.6 m resolution (0.4–0.9 µm) and four channels of multispectral,stereoscopic data (0.45–0.52, 0.52–0.60, 0.63–0.69, and 0.76–0.90 µm) with a resolution of2.44 m. QuickBird also collects 11 bits perpixel. The satellite operates in a 450 km 98degrees sun-synchronous orbit, with eachorbit taking 93.4 minutes. Each scene covers16.5 × 16.5 km.

6.3.4 Hyperspectral airborne systems

Hyperspectral airborne systems acquire spec-tral coverage in the 0.4–2.4 µm range. Thesesensors are capable of acquiring any bandcombination ranging from an optimal numberbetween 10 and 70 multispectral bands to a fullhyperspectral data set of 286 bands (SpectrumMapping 2003). Using a band configuration toolthe user can define a band file for the desirednumber of bands and individual bandwidthfor each band. This band file is then selectedin-flight using flight operations software.

6.4 PHOTOGEOLOGY

Photogeology is the name given to the use ofaerial photographs for geological studies. Toget the best out of photographs geologists mustplan the photogeological work in the office andin the field. A typical scheme is:1 annotation of aerial photographs;2 compilation of photogeology on to topo-graphical base maps;3 field checking;4 re-annotation;5 re-compilation for production of a finalphotogeological map.

The common types of aerial photos are:panchromatic black and white photographs(B&W), B&W taken on infrared (IR) sensitivefilm, color photographs, and color IR.

These films make use of different parts ofthe spectrum, e.g. visible light (0.4–0.7 µm)and photographic near infrared (0.7–0.9 µm).Panchromatic film produces a print of graytones between black and white in the visiblepart of the spectrum. These are by far the mostcommon and cheapest form of aerial photogra-phy. Color film produces a print of the visiblepart of the spectrum but in full natural color. Itis expensive but very useful in certain terrains.By contrast, color infrared film records thegreen, red, and near infrared (to about 0.9 µm)parts of the spectrum. The dyes developed ineach of these layers are yellow, magenta, andcyan. The result is a “false color” film in whichblue areas in the image result from objectsreflecting primarily green energy, green areasin the image result from objects reflectingprimarily red energy, and red areas in the imageresult from objects reflecting primarily in thephotographic near infrared portion of thespectrum. Vegetation reflects IR particularlywell (Fig. 6.6) and so IR is used extensivelywhere differences in vegetation may help inexploration. Aerial photographs are generallyclassified as either oblique or vertical.

Oblique photographs

Oblique photographs can be either high angleoblique photographs which include a horizonor low angle oblique photographs which donot include a horizon. Oblique photographsare useful for obtaining a permanent record ofcliffs and similar features which are difficult toaccess, and these photographs can be studied atleisure in the office. Similarly, studies can bemade of quarry faces to detect structural prob-lems, to plan potential dam sites, etc.

Vertical photographs

Vertical photographs are those taken by a cam-era pointing vertically downward. A typicalaerial photograph is shown in Fig. 6.7. Theprincipal point is the point on the photographthat lies on the optical axis of the camera. It isfound on a photograph by joining the fiducialmarks. Normally a title strip includes bubblebalance, flight number, photograph number,date and time of the exposure, sun elevation,flight height, and camera focal length.

Page 135: Introduction to mineral exploration

118 M.K.G. WHATELEY

FIG. 6.7 A typical black and white, vertical, aerial photograph stereo pair from Guinea. If you view the pair instereo you should see the stack in the center. The black marginal rectangles locate the fiducial marks.

best fit approach is used and photographs arecut and matched but of course they are thenunusable for other purposes. The advantageis that apparent topographical mismatches andtonal changes are reduced to a minimum andone achieves a continuous photographic cover-age. Unfortunately the scale varies across themosaic and there are occasional areas of loss,duplication, or breaks in topographical detail.

These imperfections can be corrected bymaking a controlled mosaic using scale-corrected, rectified, matched (tonal) prints.However this can be very expensive. Whenphotographs are scale corrected, matched, andproduced as a series with contours on them,they are known as orthophotographs.

6.4.1 Scale of aerial photographs

The amount of detail in an aerial photograph(resolution) is dependent upon the scale of thephotograph. The simplest way to determine aphotograph’s scale is to compare the distancebetween any two points on the ground and onthe photographs.

Photographs are taken along flight lines(Fig. 6.8). Successive exposures are taken sothat adjacent photographs overlap by about60%. This is essential for stereoscopic cover-age. Adjacent pairs of overlapping photographsare called stereopairs. To photograph a largeblock of ground a number of parallel flightlines are flown. These must overlap laterally toensure that no area is left unphotographed.This area of sidelap is usually about 30%.Less sidelap is needed in flat areas than inmountainous terrain. Flight lines are recordedon a topographical map and drawn as a flightplan. Each principal point is shown and labeledon the flight plan. These are used for orderingphotographs, from local and national Govern-ment agencies. At the start of a photogeologicalexercise, a photographic overview of the area isachieved by taking every alternate photographfrom successive flight lines and roughly joiningthem to make a print laydown. The photo-graphs are not cut, but simply trimmed to over-lap and then laid down.

A slightly more sophisticated approach isto make an uncontrolled mosaic. For this, the

Page 136: Introduction to mineral exploration

6: REMOTE SENSING 119

FIG. 6.8 Photographic coverage. (a) Showing the overlap of photographs during exposure. (b) A print laydownof the photographs from a single strip. (c) A print laydown of photographs from adjacent flight linesillustrating sidelap and drift. (Modified after Lillesand et al. 2004.)

Stereoscopicoverlap area

60%overlap

Edge variationscaused by aircraft drift

Nadir line(ground trace of aircraft)

1 23 4

5

1 2 3 4 5 6

6

Coverage ofsingle photo

(b)

(a)

Flight line

Sidelap

Aircraft drift

30% Sidelap

(c)

Page 137: Introduction to mineral exploration

120 M.K.G. WHATELEY

f

h

a b

A B

Lens

Negative

Ground level

Sea level

(H-h)=H′H

L

FIG. 6.9 The factors used to calculate the scale ofaerial photographs. (Modified after Allum 1966.)

Top

Base

PX1

P1 P2

Photograph 1

PX4

Top

Base

PX2

P1 P2 P3

PX3

Photograph 2

FIG. 6.10 Distortion on aerial photographs,where objects appear to lean away radially fromthe principal point (P1 on photograph 1 and P2

on photograph 2) of the photograph, is knownas relief displacement. The principal pointsare joined to construct the baseline of thephotograph and a perpendicular line is droppedfrom the base of the object to the baseline oneach photograph. The distance from the principalpoint on photograph 1 to this intersection pointis measured (PX1). The same exercise is carriedout on photograph 2 (PX2). The parallax of thebase of the object is given by the sum of PX3 and PX4.It is obvious that the parallax of the base is lessthan at the top. The top has greater parallax.(Modified after Allum 1966.)

In vertical photographs taken over flat ter-rain, scale (S) is a function of the focal length (f)of the camera and the flying height above theground (H′) of the aircraft (Fig. 6.9).

Scale( ) S

fH

=′

H′ is obtained by subtracting the terrain eleva-tion (h) from the height of the aircraft above adatum (H), usually sea level which is the valuegiven by the aircraft’s altimeter. The import-ant principle to understand is that photographscale is a function of terrain elevation. A planeflies at a constant (or nearly constant) height.When a plane flies over varying terrain eleva-tion, such as in mountainous areas, then thescale will vary rapidly across the photographs.

6.4.2 Parallax

All points on a topographical map are shownin their true relative horizontal positions, butpoints on a photograph taken over terrain ofvarying height are displaced from their actualrelative position. This apparent displacementis known as parallax. Objects at a higher eleva-tion lie closer to the camera and appear largerthan similar objects at a lower elevation. Thetops of objects are always displaced relativeto their bases (Fig. 6.10). This distortion onaerial photographs is known as relief displace-ment and results in objects appearing to leanaway radially from the principal point of aphotograph (Fig. 6.7).

The effect of relief displacement is illus-trated in Fig. 6.10 where the radial displace-

Page 138: Introduction to mineral exploration

6: REMOTE SENSING 121

ment differs on two adjacent photographs. Itcan be seen that the parallax of the base atthe object is less than the parallax of the topof the object. Thus, a difference in elevation,in addition to producing radial displacement,also produces a difference in parallax. It isthis difference in parallax that gives a three-dimensional effect when stereopairs are viewedstereoscopically. Allum (1966) gives a gooddescription of relief and parallax.

6.4.3 Photographic resolution

The resolution of a photographic film is influ-enced by the following factors:1 Scale, which has already been discussed.2 Resolving power of the film. A 25 ISO film(a slow film) has a large number of silver halidegrains per unit area and needs a long exposuretime to obtain an image. A 400 ISO film (a fastfilm) has relatively fewer grains per unit areabut requires less exposure time. The 25 ISOfilm has better resolution, while the 400 ISOfilm tends to be grainy.3 Resolving power of the camera lens. Com-pare the results from a high quality Nikon lenswith those of a cheap camera lens.4 Uncompensated camera motion duringexposure.5 Atmospheric conditions.6 Conditions of film processing.

The effect of scale and resolution is ex-pressed as ground resolution distance (GRD).Typically GRDs for panchromatic film varyfrom centimeters for low flown photography toabout a meter for high flown photography.

6.4.4 Problems with aerial photographs andflying

1 Drift – edges of the photograph are parallel tothe flight line but the plane drifts off course(Fig. 6.8c).2 Scale – not uniform because of parallax,terrain elevation, aircraft elevation (Fig. 6.11a),and the factors described below.3 Tilt – front to back or side to side, causingdistortions of the photograph (Fig. 6.11b,c).4 Yaw (or crab) – the plane turns into wind tokeep to the flight line, resulting in photographswhose edges are not parallel to the flight line(Fig. 6.11d).

5 Platform velocity – changing speed of theaircraft during film exposure (Fig. 6.11e). Thisapplies particularly to satellite imagery.6 Earth rotation – see section 6.2.3 (Fig. 6.11f).

6.4.5 Photointerpretation equipment

The two main pieces of photointerpretationequipment that exploration geologists use are:(i) field stereoscopes capable of being takeninto the field for quick field checking, and (ii)mirror stereoscopes which are used mainlyin the office and can view full 23 cm × 23 cmphotographs without overlap. These pieces ofequipment and their use are fully described byMoseley (1981) and Drury (2001).

Extra equipment which can be utilized in anexploration program includes a color additiveviewer and an electronic image analyzer:1 A color additive viewer color codes andsuperimposes three multispectral photographsto generate a false color composite. Multispec-tral photographs need three or four camerastaking simultaneous images in narrow spectralbands, e.g. 0.4–0.5, 0.5–0.6, 0.6–0.7 µm. Coloris far easier to interpret because the human eyecan differentiate more colors than gray tones.2 The electronic image analyzer consists of aclosed circuit TV scanner which scans a blackand white image and produces a video digitalimage. The tape is then processed by computerand the results are shown on a TV monitor.This brings us back to the digital image pro-cessing described in section 6.2.6.

High resolution scanners mean that aerialphotographs can now be converted into digitalimages. To some extent these can be treated ina similar way to the satellite digital imagery.Low altitude aerial photography can also beacquired in a high resolution digital form,with numerous spectral bands. These highresolution images can be combined with lowerresolution satellite digital imagery to preservethe best features in both. For example, a highspatial resolution aerial photography can becombined with high spectral resolution satel-lite imagery (Department of Land Information2003). This produces a product that can dis-tinguish roads, cultural and urban featuresderived from the aerial photograph, and alsoprovide information on vegetation cover andgeology derived from the satellite imagery.

Page 139: Introduction to mineral exploration

122 M.K.G. WHATELEY

(a) (b)

Altitude variation Pitch variation

(c) (d)

Yaw variationRoll variation

(e) (f)

Platform velocity Earth rotation

FIG. 6.11 Some of the problems thatcan arise as a result of variations inthe movement of the platform(aeroplane) on which the imagingdevice is mounted. (Modified afterDrury 2001.)

6.4.6 Elements of aerial photographinterpretation

Aerial photograph interpretation is based ona systematic observation and evaluation ofkey elements, and involves the identificationfrom pairs of stereophotographs of relief, tonaland textural variations, drainage patterns andtexture, erosion forms, vegetation, and land use(Lillesand et al. 2004). Drury (2001) includesadditional characteristics of the surface such asshape (of familiar geological features), context,and scale of a feature.

Topography (relief)

Each rock type generally has its own charac-teristic topographical form. There is often adistinct topographical change between tworock types, e.g. sandstone and shale. How-ever, it is not an absolute quantity becausea dolerite dyke may form a positive featurein certain climatic conditions but a negativefeature elsewhere.

Vertical photographs with their 60% overlapresult in an exaggeration of relief by three orfour times. Consequently slopes and bedding

Page 140: Introduction to mineral exploration

6: REMOTE SENSING 123

dips appear steeper than they actually are. Itis necessary to judge dips and group theminto estimated categories, e.g. 0–5°, 6–10°,11–25°, 26–45°, 46–85°, vertical. The verticalexaggeration works to the advantage of thephotogeologist, as it may lead to the identifica-tion of subtle changes in slope in otherwiserounded and subdued topographical features.

Tone

Tone refers to the brightness at any point on apanchromatic photograph. Tone is affected bymany factors such as: nature of the rock (sand-stone is light, but shale is dark), light condi-tions at the time of photographing (cloud, haze,sun angle), film, filters, and film processing.

These effects mean that we are interpret-ing relative tone values. In general terms basicextrusive and intrusive igneous rocks (lava,dolerite) have a darker tone, while bedded sand-stone, limestone, quartz schists, quartzite,and acid igneous rocks are generally lighter.Mudstone, shale, and slate have intermediatetones.

Subtle differences in rock colors are morereadily detected using color photographs, butthese are more expensive. Subtle differences insoil moisture and vegetation vigor can be morereadily detected using color IR, but even thesechange with the time of the year.

Texture

There is a large variation in apparent textureof the ground surface as seen on aerial photo-graphs. Texture is often relative and subjective,but some examples are limestone areas whichmay be mottled or speckled, whilst shale isgenerally smooth, sandstone is blocky, andgranite is rounded.

Drainage pattern

This indicates the bedrock type which in turninfluences soil characteristics and site drainageconditions. The six most common drainage pat-terns are: dendritic, rectangular, trellis, radial,centripetal, deranged (Fig. 6.12). Dendriticdrainage (Fig. 6.12a) occurs on relativelyhomogenous material such as flat-lying sedi-mentary rock and granite. Rectangular drainage

(Fig. 6.12b) is a dendritic pattern modified bystructure, such as a well-jointed, flat-lying,massive sandstone. Trellis drainage (Fig. 6.12c)has one dominant direction with subsidiarydrainage at right angles. It occurs in areasof folded sedimentary rocks. Radial drainage(Fig. 6.12d) radiates outwards from a centralarea, typical of domes and volcanoes. Centrip-etal drainage (Fig. 6.12e) is the reverse of radialdrainage where drainage is directed inwards. Itoccurs in limestone sinkholes, glacial kettleholes, volcanic craters, and interior basins (e.g.Lake Eyre, Australia and Lake Chad). Derangeddrainage (Fig. 6.12f) consists of disordered,short, aimlessly wandering streams typical ofablation till areas.

These are all destructional drainage patterns.There are numerous constructional landformssuch as alluvial fans, deltas, glacial outwashplains, and other superficial deposits. These areonly indirectly of value in exploration studies(except for placer and sand and gravel explora-tion) and are described in detail by Siegal andGillespie (1980) and Drury (2001).

Drainage texture

Drainage texture is described as either coarseor fine (Fig. 6.12h). The coarse texture developson well-drained soil and rock with little sur-face run off, e.g. limestone, chalk. Fine texturedevelops where soils and rocks have poorinternal drainage and high run off, e.g. lava andshale.

Erosion

Erosion is a direct extension of the descriptionof drainage above, but gullies, etc., often followlines of weakness and thus exaggerate featuressuch as joints, fractures, and faults.

Vegetation and land use

The distribution of natural and cultivatedvegetation often indicates differences in rocktype, e.g. sandstone and shale may be cultiv-ated, while dolerite is left as rough pasture. Onthe other hand forests may well obscure dif-ferences so great care must be taken to drawmeaningful conclusions when annotating areaswhich have a dense vegetation cover. If an area

Page 141: Introduction to mineral exploration

124 M.K.G. WHATELEY

(a) (b)

(c) (d)

(e) (f)

(g) (h)

FIG. 6.12 Typical drainage patternswhich can be used to interpret theunderlying geology (a–f), see textfor details. Drainage texture can becoarse (g) or fine (h). (Modified afterLillesand et al. 2004.)

is heavily forested, such as tropical jungles,then the use of conventional photographybecomes limited and we have to look at alter-native imagery such as reflected infrared (IR)and side-looking airborne radar (SLAR) (Sabins1997, Drury 2001).

Scanners with narrow apertures viewing thereflectance from vegetation (thematic mappers)

are now being used (Goetz et al. 1983). Oneway in which all imagery can be used is by tak-ing pictures of the same area at different timesof the year (multitemporal photography). Thismay emphasize certain features at a given timeof the year. Lines of vegetation, e.g. bushes,trees, etc., are a good indicator of fractures,faults, veins, and joints. The joints are more

Page 142: Introduction to mineral exploration

6: REMOTE SENSING 125

porous and have more available ground water,producing lines of more vigorous vegetation.

Lineament

This is a word used in photogeology to de-scribe any line on an aerial photograph thatis structurally controlled by joints, fractures,faults, mineral veins, lithological horizons,rock boundaries, etc. These have already beendiscussed above, e.g. streams, gullies, lines ofvegetation, etc. It is important to differentiatethese from fence lines, roads, rails, etc.

6.4.7 Interpretation

By using the elements of aerial photographinterpretation described above we identifylandforms, man-made lines versus photogeo-logical lineaments, vegetation, rock outcropboundaries, rock types, and rock structures(folds, faults, fractures). Once identified, thegeological units and structure are transferredon to a topographical base map.

Most near-surface mineral deposits in acces-sible regions have been discovered. Emphasisis now on remote regions or deep-seateddeposits. The advantages of using imageryin remote, inaccessible regions are obvious.Much information about potential areas withdeep-seated mineralisation can be provided byinterpretation of surface features, e.g. possibledeep-seated faults and/or fractures which mayhave been pathways for rising mineralizingsolutions.

Aerial photographs are normally taken be-tween mid-morning and mid-afternoon whenthe sun is high and shadows have a minimaleffect. Low sun angle photographs containshadow areas which, in areas of low relief, canreveal subtle relief and textural patterns notnormally visible in high sun elevation photo-graphs. There are disadvantages in using lowsun angles, especially in high relief areas,where loss of detail in shadow and rapid lightchange conditions occur very early and verylate in the day.

Most exploration studies involve multi-image interpretation. Often satellite imagesare examined first (at scales from 1:1M, upto 1:250,000) for the regional view. If high

altitude photography is available (1:120,000to 1:60,000) then this may also be examined.Target areas are selected from the above studiesand, in conjunction with literature reviews, itis then possible to select areas for detailedphotogeological studies.

Aerial photographs are mainly used byvarious ordnance surveys for the productionof topographic maps using photogrammetry.Photogrammetry is described in detail byAllum (1966). Some additional uses of aerialphotography are regional geological mapping(1:36,000–1:70,000), detailed geological map-ping (1:5000–1:20,000), civil engineering forroad sites, dam sites, etc., open pit managementfor monthly calculations of volumes extractedduring mining, soil mapping, land use andland cover mapping, agricultural applications,forestry applications, water resources applica-tions (e.g. irrigation, pollution, power, drinkingwater, and flood damage), manufacturing andrecreation, urban and regional planning, wet-land mapping, wildlife ecology, archaeology,and environmental impact assessment.

6.5 SUMMARY

In remote sensing we are concerned with thecollection of information about an area with-out being in contact with it, and this can beachieved using satellites or aeroplanes carryingelectronic scanners or sensors or photographicand TV cameras.

Some satellite systems producing data use-ful for mineral exploration include NASA’sLandsat and Terra (ASTER) and the FrenchSPOT. Each has its advantages (section 6.3).The first Landsat satellite was launched in1972 and SPOT 1 in 1986. Satellites are solarpowered and transmit data to their homestation in digital form which enables the geo-logist to manipulate, combine, and comparethis data with geological, geochemical, andgeophysical data which has itself been ex-pressed as digital images. Satellite imagery isused in structural investigations, in hydro-geology, to provide basic geological maps, todetect hydrothermal alteration, and to producemaps of regional and local fracture patternsthat may have controlled mineralisation orhydrocarbon accumulation.

Page 143: Introduction to mineral exploration

126 M.K.G. WHATELEY

Aerial photography has been used for muchlonger than satellite imagery and is importantto the geologist for the production of topo-graphical maps (photogrammetry) as well as inthe making of geological maps (photogeology).In areas of good exposure aerial photographsyield much valuable geological informationand even in areas of only 5% outcrop theamount of information they provide is often in-valuable to the explorationist; some of this isdiscussed in section 6.4.6.

6.6 FURTHER READING

Good textbooks which cover most aspectsof remote sensing in geology are Siegal andGillespie (1980) and Drury (2001). The classicbook about photogeology is Allum (1966),which despite its age is still worth read-ing. Detailed information on all aspects ofremote sensing is covered by Lillesand et al.(2004). Recent developments in remote sens-ing appear monthly in the InternationalJournal of Remote Sensing published by Taylor& Francis. A good introduction to digitalimage processing is given by Niblak (1986), andtechniques for image processing and classifi-cation are explained well by Schowengerdt(1983).

Page 144: Introduction to mineral exploration

7Geophysical Methods

John Milsom

7.1 INTRODUCTION

Exploration geophysicists use measurementsof physical quantities made at or above theground surface or, more rarely, in boreholesto draw conclusions about concealed geology.Lines may have to be surveyed and cleared,heavy equipment may have to be brought onsite, and detectors and cables may have to bepositioned, so geophysical work on the groundis normally rather slow. Airborne geophysicalsurveys, on the other hand, provide the quick-est, and often the most cost-effective, ways ofobtaining geological information about largeareas. In some cases, as at Elura in central NewSouth Wales (Emerson 1980), airborne indica-tions have been so clear and definitive thatground follow-up work was limited to definingdrill sites, but this is unusual. More often,extensive ground geological, geochemical, andgeophysical surveys are required to prioritizeairborne anomalies.

For a geophysical technique to be useful inmineral exploration, there must be contrasts inthe physical properties of the rocks concernedthat are related, directly or indirectly, to thepresence of economically significant minerals.Geophysical anomalies, defined as differencesfrom a constant or slowly varying background,may then be recorded. Anomalies may takemany different forms and need not necessarilybe centered over their sources (Fig. 7.1). Ideally,they will be produced by the actual economicminerals, but even the existence of a strongphysical contrast between ore minerals and thesurrounding rocks does not guarantee a signifi-cant anomaly. The effect of gold, which is both

dense and electrically conductive, is negligiblein deposits suitable for large-scale miningbecause of the very low concentrations (seeCorbett 1990). Diamonds (the other present-day “high profile” targets) are also present indeposits in very low concentration and have,moreover, no outstanding physical properties(see Chapter 17). In these and similar cases,geophysicists must rely on detecting associ-ated minerals or, as in the use of seismic reflec-tion to locate offshore placers, and magneticsand electromagnetics to locate kimberlites(see Macnae 1995), on defining favorableenvironments.

Geophysical methods can be classed as eitherpassive (involving measurements of naturallyexisting fields) or active, if the response of theground to some stimulus is observed. Passivemethods include measurements of magneticand gravity fields, naturally occurring alpha andgamma radiation, and natural electrical fields(static SP and magnetotellurics). All other elec-trical and electromagnetic techniques, seismicmethods, and some downhole methods thatuse artificial radioactive sources are active.

In the discussion that follows, the generalprinciples of airborne surveys are consideredbefore describing the individual methods indetail. The final sections are concerned withpracticalities, and particularly with the role ofnonspecialist geologists in geophysical work.

7.2 AIRBORNE SURVEYS

Magnetic, electromagnetic, gamma-ray, andmore recently gravity, measurements do not

Page 145: Introduction to mineral exploration

128 J. MILSOM

require physical contact with the ground andcan therefore be made from aircraft. Inevitably,there is some loss of sensitivity, since detectorsare further from sources, but this may evenbe useful in filtering out local effects frommanmade objects. The main virtue of airbornework is, however, the speed with which largeareas can be covered. Surveys may be flowneither at a constant altitude or (more com-monly in mineral exploration) at a (nominally)constant height above the ground. Aircraftare often fitted with multiple sensors andmost installations include a magnetometer(Fig. 7.2).

Airborne surveys require good navigationalcontrol, both at the time of survey and later,when flight paths have to be plotted (recov-ered). Traditionally, the pilot was guided by anavigator equipped with maps or photo-mosaicsshowing the planned line locations. Coursechanges were avoided unless absolutely neces-sary and in many cases the navigator’s main jobwas to ensure that each line was at least begunin the right place. Although infills were (andare) required if lines diverged too much, a linethat was slightly out of position was preferredto one that continually changed direction andwas therefore difficult to plot accurately.

Low level navigation is not easy, since eventhe best landmarks may be visible for only afew seconds when flying a few hundred metersabove the ground, and navigators’ opinions ofwhere they had been would have been very in-adequate bases for geophysical maps. Trackingcameras were therefore used to record images,either continuously or as overlapping frames,on 35 mm film. Because of the generally smallterrain clearance, very wide-angle (“fish-eye”)lenses were used to give broad, although dis-torted, fields of view. Recovery was done dir-ectly from the negatives and even the mostexperienced plotters were likely to misidentifyone or two points in every hundred, so rigorouschecking was needed. Fiducial numbers andmarkers printed on the film provided the essen-tial cross-references between geophysical data,generally recorded on magnetic tape, and flightpaths. Mistakes could be made at every stage inprocessing, and errors were distressingly com-mon in aeromagnetic maps produced, and pub-lished, when these methods were being used.

(b)

(a)

(d)

(c)

Magnetic

100

10

5

5

5

−5

−5

Gravity

Total field

directio

n

Electromagnetic(CWEM)

Electromagnetic(VLF)

%

%

nT

g.u.

FIG. 7.1 Single body geophysical anomalies.A massive sulfide orebody containing accessorymagnetite or pyrrhotite might produce themagnetic anomaly (a), the gravity anomaly (b),the Slingram electromagnetic anomaly (c), andthe VLF anomaly (d; see also Fig. 7.15). Valuesshown on the vertical axes are typical, but theranges of possible amplitudes are very large forsome of the methods.

Page 146: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 129

FIG. 7.2 Airborne geophysical installation. The aircraft has been modified for combined magnetic and TEMsurveys. The boom projecting from above the cockpit is the forward mounting of the electromagnetictransmitter loop, which extends from there to the wingtips and then to the tail assembly. The magnetometersensor is housed in the fibreglass “stinger” at the rear, which places it as far as possible from magnetic sourcesin the aircraft.

Visual navigation in savannah areas wasdifficult enough, since one clump of trees looksmuch like another when seen sideways on,but plotting from vertical photographs in suchareas was often easy because tree patterns arequite distinctive when viewed from above.Visual navigation and flight-path recoverywere both very difficult over jungles or desertsand were, of course, impossible over water.If there were no existing radio navigationalaids, temporary beacons might be set up, butsince several were required and each had tobe manned and supplied, survey costs wereconsiderably increased. Self-contained Doppleror inertial systems, which could be carried en-tirely within the aircraft, produced significantimprovements during the period from about1975 to 1990, but airborne work has now beencompletely transformed by the use of globalpositioning satellites (GPS). GPS systems pro-vide unprecedented accuracy in monitoringaircraft position and have largely eliminatedthe need for tracking cameras, although theseare still sometimes carried for back-up and

verification. Velocities can be estimated withgreat accuracy, making airborne gravimetry,which requires velocity corrections, usable forthe first time in mineral exploration.

Survey lines must not only be flown in theright places but at the right heights. Groundclearance is usually measured by radar alti-meter and is often displayed in the cockpitsby limit lights on a head-up display. The extentto which the nominal clearance can be main-tained depends on the terrain. Areas too ruggedfor fixed-wing aircraft may be flown by helicop-ter, but even then there may be lines that canonly be flown downhill or are too dangerous tobe flown at all. Survey contracts define allow-able deviations from specified heights and lineseparations, and lines normally have to bereflown if these limits are exceeded for morethan a specified distance, but contracts alsocontain clauses emphasizing the paramountimportance of safety. In a detailed survey theradar altimeter data can be combined with GPSestimates of elevation to produce a high resolu-tion digital terrain model (DTM) which may

Page 147: Introduction to mineral exploration

130 J. MILSOM

prove to be one of the most useful of the surveyproducts (see section 9.2).

The requirement for ever smaller groundclearances and tighter grids is making demandson pilots that are increasingly difficult to meet.In the future some airborne surveys, and espe-cially aeromagnetic surveys, may be flown bypilotless drones.

7.3 MAGNETIC SURVEYS

Magnetic surveys are the quickest, and oftenthe cheapest, form of geophysics that can pro-vide useful exploration information. A fewminerals, of which magnetite is by far the mostcommon, produce easily detectable anomaliesin the Earth’s magnetic field because the rockscontaining them become magnetized. Themagnetization is either temporary (induced)and in the same direction as the Earth’s field,or permanent (remanent) and fixed in directionwith respect to the rock, regardless of folding orrotation.

Since magnetite is a very minor constituentof sediments, a magnetic map generally recordsthe distribution of magnetic material in the un-derlying crystalline basement. Even sedimentsthat do contain magnetite have little effect onairborne sensors, partly because the fields fromthe randomly oriented magnetite grains typicalof sediments tend to cancel out and partly be-cause the fields due to thin, flat-lying sourcesdecrease rapidly with height. Because the smallpieces of iron scrap that are ubiquitous inpopulated areas strongly affect ground magnet-ometers, and also because ground coverage isslow, most magnetic work for mineral explora-tion is done from the air. Line separations havedecreased steadily over the years and may nowbe as little as 100 m. Ground-clearance mayalso be less than 100 m.

Magnetic surveys are not only amongst themost useful types of airborne geophysics butare also, because of the low weight and sim-plicity of the equipment, the cheapest. Thestandard instrument is now the high sensitiv-ity cesium vapor magnetometer. Proton mag-netometers are still occasionally used but thecesium magnetometer is not only a hundredtimes more sensitive but provides readings

every tenth of a second instead of every secondor half second. Both cesium and proton instru-ments are (within limits) self-orienting and canbe mounted either on the aircraft or in a towedbird. Sensors on aircraft are usually housedin specially constructed nonmagnetic booms(stingers) placed as far from the main aircraftsources of magnetic field as possible (Fig. 7.2).Aircraft fields vary slightly with heading andmust be compensated by systems of coils andpermanent magnets. The magnitude of anyresidual heading error must be monitored on aregular basis.

The Earth’s magnetic field is approximatelythat of a dipole located at the Earth’s centerand inclined at about 10 degrees to the spinaxis. Distortions covering areas hundreds ofkilometers across can be regarded as due toa small number of subsidiary dipoles locatedat the core–mantle boundary. The practicalunit of magnetic field for survey work is thenanotesla (nT, sometimes also known as thegamma). At the magnetic poles the field isabout 60,000 nT and vertical, whereas at theequator it is about 30,000 nT and horizontal(Fig. 7.3). Slow variations in both magnitudeand direction, which must be taken intoaccount when comparing surveys made morethan a few months apart, are described, togetherwith large-scale variations with latitude andlongitude, by a complicated experimentallydetermined formula, the International Geom-agnetic Reference Field (IGRF). This is gener-ally a reasonable approximation to the regionalfield in well-surveyed areas where the controlon its formulation is good, but may be unsatis-factory in remote areas.

As well as long-term variations in fieldstrength, there are cyclical daily (diurnal)changes which are normally of the order of 20–60 nT in amplitude. Variations are low at nightbut the field begins to increase at about dawn,reaches a peak at about 10 a.m., declines rap-idly to a minimum at about 4 p.m., then risesmore slowly to the overnight value. Althoughthis pattern is repeated from day to day, thechanges are not predictable in detail and mustbe determined by actual measurement. Diurnaleffects are especially important in airbornesurveys, which now typically use measuringprecisions of the order of 0.1 nT and contour

Page 148: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 131

30°

30°

60°

60°0°60°120°180° 120°

70o

70°

30,000

40,000

40,000

40,000

60,00060,000

50,000

50,0

00

50,000

60,0

00

60,000

60°

50°40

°30°20°

10°0°

10°20°30°40°

80° 80°

50°60°

70°

80°

80°

60°

FIG. 7.3 The Earth’s magnetic field. Solid lines are contours of constant dip, dashed lines are contours ofconstant total field strength (in nT).

intervals of 5 nT or less. During such surveys,fixed ground magnetometers provide digitalrecords of readings taken every few minutes.Corrections are made either by using theserecords directly or by adjustments based onsystems of tie lines flown at right angles tothe main lines of the survey, or (preferably) bya combination of the two. Tie lines can bemisleading if they cross flight lines in regionsof steep magnetic gradients, and especially ifthe terrain is rugged and the aircraft was atdifferent heights on the two passes over anintersection point.

During periods of sunspot activity, streamsof high energy particles strike the upper atmos-phere and enormously increase the ionosphericcurrents, causing changes of sometimes hun-dreds of nanotesla over times as short as 20minutes. Corrections cannot be satisfactorilymade for these “magnetic storms” (which arenot related to meteorological storms and mayoccur when the weather is fine and clear) andwork must be delayed until they are over.

Magnetized bodies produce very differentanomalies at different magnetic latitudes(Fig. 7.4), the regional changes in dip being farmore important from an interpretation pointof view than the regional changes in absolutemagnitude. Because of the bipolar nature ofmagnetic sources, all magnetic anomalies haveboth positive and negative parts and the nettotal magnetic flux involved is always zero.This is true even of vertically magnetizedsources at a magnetic pole, but the negativeflux is then widely dispersed and not obviouson contour maps. Magnetic maps obtained nearthe poles are therefore more easily interpretedthan those obtained elsewhere, and reduction-to-pole (RTP) processing has been developedto convert low and middle latitude maps totheir high latitude equivalents. If, however (asis usually the case with strongly magnetizedbodies), there are significant permanent mag-netizations in directions different from thepresent Earth field, RTP processing will distortand displace anomalies. It is far better that

Page 149: Introduction to mineral exploration

132 J. MILSOM

I = 90°H = 60,000 nT

(a)

(b)

(c)

(d)

(e)

(f)

I = 0°H = 35,000 nTMI = 90°

I = 0°H = 35,000 nT

To equator To pole

I = 60°H = 55,000 nT

I = 60°H = 55,000 nT

X

YY Y

YYY

FIG. 7.4 Variations in shape ofmagnetic anomalies with latitudeand orientation. (a–c) The variationwith magnetic inclination of theanomaly produced by the body Xwhen magnetized by inductiononly. (d) The anomaly at theequator if, in addition to theinduced magnetization, there is apermanent magnetization in thevertical direction. This situation iscommon with vertical intrusions(dikes and volcanic pipes) in lowlatitudes. (e,f) The variation in theanomaly produced in mid-latitudesby an inductively magnetized sheetwith varying dip. All profiles areoriented north–south and all bodiesare assumed to be “2D,” i.e. tohave infinite strike length at rightangles to the profile.

Page 150: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 133

interpreters become familiar with, and under-stand, the anomalies typical of the latitudesin which they work than that they rely onprocesses that can significantly degrade theirdata.

Magnetic anomalies are caused by magnet-ite, pyrrhotite, and maghemite (loosely describ-able as a form of hematite with the crystal formof magnetite). Magnetite is by far the com-monest. Ordinary hematite, the most abundantore of iron, only rarely produces anomalieslarge enough to be detectable in conventionalaeromagnetic surveys. Because all geologicallyimportant magnetic minerals lose their mag-netic properties at about 600°C, a temperaturereached near the base of the continental crust,local features on magnetic maps are virtuallyall of crustal origin. Magnetic field variationsover sedimentary basins are often only a fewnanotesla in amplitude, but changes of hun-dreds and even thousands of nanotesla are com-mon in areas of exposed basement (see Fig. 4.4).The largest known anomalies reach to morethan 150,000 nT, which is several times thestrength of the Earth’s normal field.

Because magnetite is a very common acces-sory mineral but tends to be concentrated inspecific types of rock, magnetic maps contain awealth of information about rock types andstructural trends that can be interpreted qual-itatively. Image processing techniques such asshaded relief maps are increasingly being usedto emphasize features with pre-selected strikesor amplitudes. Quantitative interpretation in-volves estimating source depths, shapes, andmagnetization intensities for individual anom-alies, but it would seldom be useful, or evenpracticable, to do this for all the anomalies in anarea of basement outcrop. Magnetization inten-sities are not directly related to any economic-ally important parameter, even for magnetiteores, and are rarely of interest. Automaticmethods such as Werner deconvolution (whichmatches observed anomalies to those producedby simple sources) and Euler deconvolution(which uses field gradients) generally focus ondepth determination. Direct determination ofgradients is becoming commonplace, especiallyin exploration for kimberlites. In airborneinstallations horizontal gradients can be meas-ured using nose-and-tail or wingtip sensors.

7.4 GRAVITY METHOD

The gravity field at the surface of the Earth isinfluenced, to a very minor extent, by densityvariations in the underlying rocks. Rock den-sities range from less than 2.0 Mg m−3 for softsediments and coals to more than 3.0 Mg m−3

for mafic and ultramafic rocks. Many ore min-erals, particularly metal sulfides and oxides,are very much denser than the minerals thatmake up the bulk of most rocks, and orebodiesare thus often denser than their surroundings.However, the actual effects are tiny, generallyamounting even in the case of large massivesulfide deposits to less than 1 part per millionof the Earth’s total field (i.e. 1 mgal, or 10 ofthe SI gravity units or g.u. that are equal to10−6 m s−2). Gravity meters must therefore beextremely sensitive, a requirement which tosome extent conflicts with the need for themalso to be rugged and field-worthy. They meas-ure only gravity differences and are subject todrift, so that surveys involve repeated refer-ences to base stations. Manual instruments arerelatively difficult to read, with even experi-enced observers needing about a minute foreach reading, while the automatic instrumentsnow becoming popular require a similar timeto stabilize. Because of the slow rate of cover-age, gravity surveys are more often used tofollow up anomalies detected by other methodsthan to obtain systematic coverage of largeareas. The potential of the systematic approachwas, however, illustrated by the discovery ofthe Neves Corvo group of massive sulfidedeposits (Fig. 7.5) following regional gravitysurveys of the Portuguese pyrite belt on 100and 200 meter grids (Leca 1990). Airborne grav-ity is technically difficult and, although avail-able in a rather primitive form as early as 1980,is still not widely used. However, BHP-Billitoninvested several million dollars in developingthe Falcon airborne gradiometer system speci-fically for use in their own mineral explorationprograms and has reported several successessince beginning operations in 1999. The BellGeospace Full Tensor Gradiometer (FTG),which is now available commercially, givescomparable results, equivalent roughly toground stations recorded on a 150 m grid, with0.2 mgal noise.

Page 151: Introduction to mineral exploration

134 J. MILSOM

FIG. 7.5 Gravity anomalies over the Neves Corvo group of orebodies in the Portuguese pyrite belt. The mainbodies, indicated by crosses, coincide with local culminations in a regional gravity high associated with highdensity strata. Under these circumstances it is difficult to separate the anomalies from their background fortotal mass or other analysis. (Data from Leca 1990.)

0 500 1000 m

Corvo

Graca

Zambujal

NevesNeves

535

535

530

530

535

540

Gravity tends to increase towards the poles.The difference between the polar and equato-rial fields at sea level amounts to about 0.5% ofthe total field, and corrections for this effect aremade using the International Gravity Formula(IGF). Gravity also varies with elevation. The0.1 g.u. (0.01 mgal) effect of a height changeof only 5 cm is detectable by modern gravitymeters and corrections for height must be madeon all surveys. Adding the free-air correction,which allows for height differences, and sub-tracting the Bouguer correction (which appro-ximates the effect of the rock mass betweenthe station and the reference surface by thatof a uniform flat plate extending to infinity inall directions) to the latitude-corrected gravityproduces a quantity known as the Bouguer

anomaly or Bouguer gravity. Provided that thedensity of the topography has been accuratelyestimated, a Bouguer gravity map is usually agood guide to the effect of subsurface geologyon the gravity field, although additional cor-rections are needed in rugged terrain. Althoughmuch can be achieved with careful processing(e.g. Leaman 1991), terrain corrections can beso large, and subject to such large errors, as toleave little chance of detecting the small anom-alies produced by local exploration targets.

Geophysical interpretations are notoriouslyambiguous but the gravity method does pro-vide, at least in theory, a unique and unambigu-ous answer to one exploration question. If ananomaly is fully defined over the ground sur-face, the total gravitational flux it represents

Page 152: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 135

is proportional to the total excess mass of thesource body. Any errors or uncertainties aredue to the difficulty, under most circum-stances, of defining background, and are notinherent in the method. The true total massof the body is obtained by adding in the massof an equivalent volume of country rock, sosome knowledge of source and country-rockdensities is needed.

Very much less can be deduced about eitherthe shapes or the depths of sources. A variety ofrule of thumb methods exist, all relying on therough general relationship between the lateralextent of an anomaly and the depth of itssource (e.g. Milsom 2002). A deeper body will,other things being equal, give rise to a broader(and flatter) anomaly. Strictly speaking, onlyhomogeneous spherical bodies can be approxi-mated by the point sources on which most ofthe rules are based, and the widths of realanomalies obviously depend on the widths oftheir sources, but rough guides are obtainedto the depths of centers of mass in many cases.Also, the peaks of gravity anomalies are gen-erally located directly above the causativebodies, which is not the case with many of theother geophysical methods. However, all inter-pretation is subject to a fundamental limita-tion because the field produced by a given bodyat a given depth can always also be produced bya laterally more extensive body at a shallowerdepth. No amount of more detailed survey orincreased precision can remove this ambiguity.

Full quantitative interpretations are usuallymade by entering a geological model into acomputer program that calculates the corres-ponding gravity field, and then modifying themodel until there is an acceptable degree offit between the observed and calculated fields.This process is known as forward modeling.Several packages based on algorithms publishedby Cady (1980) are available for modelingfields due to 2D bodies (which have constantcross-section and infinite strike extent) and to21/2D bodies, which are limited in strike length.Full three-dimensional modeling of bodies ofcomplex shape is now becoming more com-mon as desktop computers become faster andmemory sizes increase. The reliability of anyinterpretation, no matter how sophisticatedthe technique, depends, of course, on the valid-ity of the input assumptions. For an example

of the use of modeling in gold exploration seesection 14.5.4.

7.5 RADIOMETRICS

Natural radioactive decay produces alpha par-ticles (consisting of two neutrons and twoprotons bound together), beta particles (high-energy electrons), and gamma rays (very highfrequency electromagnetic waves which quan-tum theory allows us to treat as particles).Alpha and beta radiations are screened out byone or two centimeters of solid rock, and even alittle transported soil may conceal the alphaand beta effects of mineralisation. Gamma raysare more useful in exploration but even theyhave ranges of only one or two meters in solidmatter. The traditional picture of a radiometricsurvey is of the bearded prospector, with orwithout donkey, plodding through the desertand listening hopefully to clicking noises com-ing from a box on his hip. The Geiger counterhe used was sensitive only to alpha particlesand is now obsolete, but tradition dies hard.Modern gamma rays detectors (scintillometers)have dials or digital readouts, but groundsurvey instruments can usually also be set to“click” in areas of high radioactivity. For anexample of a carborne survey see Fig. 4.8.

Radiometric methods are geophysical odd-ities because the measurements are of countrates, which are subject to statistical rules,rather than of fields with definite, even if diur-nally variable, values. Survey procedures muststrike a balance between the accuracy obtain-able by observing for long periods at each sta-tion and the speed of coverage demanded byeconomics. Statistics can be improved by usingvery large detector crystals, and these are es-sential for airborne work, but their cost risesdramatically with size. The slow speeds neededto obtain statistically valid counts in the airgenerally necessitate helicopter installations.

Gamma ray energies can be measured byspectrometers, allowing different sources ofradiation to be identified. Terrestrial radia-tion may come from the decay of 40K, whichmakes up about 0.01% of naturally occurringpotassium, or from thorium or uranium. Radia-tion arises not only from decay of these long-lived primeval elements but also from unstable

Page 153: Introduction to mineral exploration

136 J. MILSOM

uranium in the potassium window. It is usualto also record total count but discrimination isimportant, as was shown by the discovery ofthe Yeeleerie uranium deposit in Western Aus-tralia in a salt lake where the radiometricanomaly had initially been attributed to potashin the evaporites. Interpretational difficultiesare introduced in the search for uranium by thefact that 214Bi lies below radon in the decaychain. Radon is a gas and the isotope concernedhas a half-life of several days, allowing con-siderable dispersion from the primary source.

The unpopularity of nuclear power, and theavailability of uranium from dismantled nu-clear bombs, made exploration for uraniummuch less attractive, and the importance ofradiometric methods declined accordingly.Much of the radiometric work being under-taken at present is for public health purposesand in such applications the monitoring ofalpha particles from radon gas using alpha cupsand alpha cards may be as useful as the detec-tion of gamma rays. Gamma-ray surveys do,however, have geological applications in locat-ing alteration zones in acid and intermediateintrusions and in the search for, and evalua-tion of, phosphate and some placer deposits.Another, and developing, application is in air-borne soils mapping, since soils derived fromdifferent rocks have different radiometric sig-natures. False-color maps, produced by assign-ing each of the three primary colors to oneof the three main radioactive source elements,can supplement magnetic maps as geologicalmapping tools and may reveal features suchas old river channels that have explorationsignificance. The addition of a gamma-raysensor (typically, a 256-channel spectrometer)to an aeromagnetic system, at an increase incost per line kilometer of the order of 10–20%,is therefore becoming routine.

7.6 RESISTIVITY

The electrical properties of continuous mediaare characterized by their resistivities, whichare the resistances of meter cubes. Rockresistivities vary widely but are normallywithin the range from 0.1 to 1000 ohm-meters.Because most minerals are insulators, these

10,000

Cou

nts

1000

100

10

1.0 2.0

Energy (MeV)

Thorium(2.615 MeV)

Uranium(1.76 MeV)

Potassium(1.46 MeV)

FIG. 7.6 Typical natural gamma-ray spectrumrecorded at ground level. The peaks due to specificdecay events are superimposed on a background ofscattered radiation from cosmic rays and higherenergy decays. Note the logarithmic vertical scale.

daughter isotopes of lighter elements, formedduring decay and usually themselves unstableand producing further offspring. Gamma-rayphotons with energies above 2.7 MeV can onlybe of extraterrestrial (usually solar) origin, but a2.615 MeV signal is produced by a thoriumdaughter. The most prominent uranium (actu-ally 214Bi) signal is at 1.76 MeV and the singlepeak associated with potassium decay is at1.46 MeV. These and other peaks are superim-posed on a background of scattered radiation(Fig. 7.6), but the relative importance of thethree main radioelements in any source can beestimated by analyzing spectra using rathersimple techniques based on the count rateswithin windows centerd on these three energylevels. Corrections must be made for the effectsof scattered thorium radiation in the uraniumwindow and for the effect of both thorium and

Page 154: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 137

I

II

I

V

VV

V I

V

a

a

aa

x

na

L L

a a a

(a) Wenner

Exact Ideal dipole ‘a’

andX = x /LY = y /L

(c) Schlumberger

(e) Dipole–dipole

(d) Gradient

where K

(b) Two-electrode (pole–pole)

VI

ρα

L2 – �2

2�VI

= πρα L2

aVI

K=

1 − X[y 2 + (1 − X )2]3 2

1+X[y 2 + (1 + X )2]3 2

(n + 1)(n + 2)a

= 2 +

ρα

L2

2�VI

=ρα

VI

= 2 aρα

88

y

2�

2L

n

Ideal dipole ‘2�’ π

VI

=ρα π

π

π

π= 2 aπ

FIG. 7.7 Arrays and geometric conversion factors for resistivity and IP surveys: (a) Wenner array; (b) pole–polearray; (c) Schlumberger array; (d) gradient array; (e) dipole–dipole array.

resistivities are commonly controlled by rockporosity and by the salinity of the pore waters.However, clay minerals are electrically polar-ized and rocks containing them are highly con-ductive when even slightly moist.

A few minerals, notably graphite and thebase metal sulfides (except sphalerite), conductby electron flow and can reduce rock resistivityto very low values if present in significantamounts. Even so, straightforward measure-ment of direct-current resistivity is seldomused by itself in the search for base metals.More common uses are in estimating over-burden thicknesses and in determining theextents of deposits of various bulk minerals.Lenses of clean, and therefore resistive, gravelscan be found in clays and, conversely, chinaclay deposits can be found in granites or re-deposited as ball clays in bedrock depressions.

The two fundamental requirements of anyresistivity survey are the introduction ofcurrent and the measurement of voltage but,because of contact effects at electrodes, sub-surface resistivity cannot be estimated usingthe same pair of electrodes for both purposes.Instead, two current and two voltage electrodesare used in arrays which are usually, but notnecessarily, linear. For convenience, current isusually supplied through the outer electrodesbut the geometrical factors of Fig. 7.7 can alsobe used to calculate the averaged or apparentresistivities if the two inner electrodes are usedfor this purpose.

Power may be provided by motor generatorsor rechargeable batteries. Generators and volt-meters may be separated or combined in singleunits that record resistance values directly. Toreduce polarization effects at the electrodes,

Page 155: Introduction to mineral exploration

138 J. MILSOM

and also to compensate for any natural currentsand voltages, directions of current flow arereversed periodically, usually at intervals of theorder of a second. Voltage electrodes that donot polarize can be made by immersing copperrods in saturated copper sulfate solution con-tained in porous pots. Contact with the groundis made by solution that leaks through thebases of the pots. These electrodes are messyand inconvenient and are used only when, as ininduced polarization surveys (section 7.8), theyare absolutely essential.

For studying lateral variations in resistiv-ity, the Wenner, pole–pole, and gradient arrays(Fig. 7.7a,b,d) are the most convenient. InWenner traversing, the leading electrode ismoved one electrode interval along the line foreach new station and each following electrodeis moved into the place vacated by its neighbor.With the gradient array, which requires verypowerful current generators, the outer elec-trodes are fixed and far apart and the innerelectrodes, separated by only a few meters, aremoved together. The pole–pole array also hasthe advantage of requiring only two movingelectrodes but the very long cables needed tolink these to the fixed electrodes “at infinity”may be inconvenient.

Resistivity surveys are also used to investi-gate interfaces, such as water tables or bedrocksurfaces, that are approximately horizontal.Current can be sent progressively deeper intothe ground by moving electrodes further apart.Resistivities can thus be estimated for pro-gressively deeper levels, although, inevitably,resolution decreases as electrode separationsincrease. The Wenner array is often used, asis the Schlumberger array (Fig. 7.7c), whichis symmetrical but has outer electrodes verymuch further apart than the inner ones. All fourelectrodes must be moved when expanding theWenner array but with the Schlumberger array,within limits, the inner electrodes can be leftin the same positions. Traditionally, the resultsobtained with either array were interpretedby plotting apparent resistivity against arrayexpansion on log–log paper and comparing thegraphs with type curves. A single sheet ofcurves was sufficient if only two layers werepresent (Fig. 7.8), but a book was needed forthree layers and a library would have been

needed for four! Multiple layers were thereforeusually interpreted by matching successivecurve segments using two-layer curves togetherwith auxiliary curves that constrained the two-layer curve positions. This technique is nowlargely obsolete, having been replaced by inter-active computer modeling, but the insightsprovided by type curves are still valuable. Moresophisticated computer programs now allowthe results obtained from multiple Wennertraverses at different electrode separationsalong single lines to be inverted to cross-sections approximating the actual subsurfacedistribution of resistivity. The fieldwork ismuch more time-consuming than simplesingle-expansion depth sounding, but the extraeffort is almost always justified by the im-proved results.

7.7 SPONTANEOUS POLARIZATION (SP)

Natural currents and natural potentials existand have exploration significance. In particu-lar, sulfide orebodies may produce negativeanomalies of several hundred millivolts. It wasoriginally thought that these potentials weremaintained by oxidation of the ore itself, but itis now generally agreed that the ore acts as apassive conductor, focussing currents producedby the oxidation-reduction reactions that takeplace across the water table. For a voltage tobe produced, the conductor must straddle thewater table and the method, although quick andsimple, requiring only a high-impedance volt-meter, some cables, and a pair of nonpolarizingelectrodes, is often rejected because economic-ally exploitable bodies do not necessarily pro-duce anomalies even when highly conductive.

7.8 INDUCED POLARIZATION (IP)

Flow of electric current in a rock mass cancause parts of it to become electrically polar-ized. The effect is almost negligible in sand-stones, quite marked in clays, in which smallpore spaces and electrically active surfacesimpede ionic flow, and can be very strong at thesurfaces of grains of electronic conductors suchas graphite and metallic sulfides. If current

Page 156: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 139

ρ2 + ρ1

Wenner“cross”

Schlumberger“cross”

Wenner: a /h (= 2L /3h)

ρα / ρ1 = 1

ρ2 – ρ1ρ ρρ ρ

ρ α/ρ

1

Schlumberger: L /h

k = 1.0

k =

k = −1.0 −0.9 −0.8

−0.7

−0.6

−0.5

−0.4

−0.3

−0.2

−0.1

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.80.9

0.0

a/h

= 1

(W

enne

r)“Cross” for Wenneror Schlumberger

field curve plottedusing total array

length (2L) asordinate

L/h

= 1

(S

chlu

mbe

rger

)

L/h

= 1

0

a/h

= 1

0

ρ

ρρ α

α

α ρ

FIG. 7.8 Two-layer type curves for Wenner and Schlumberger depth sounding. Although the curves forthese two arrays are not actually identical, the differences are so small as to be undetectable at this scale,and are less than the inevitable errors in measurement. The curves are plotted on 2 × 3 cycle log–log paper.For interpretation, field curves plotted on similar but transparent paper are laid over the type curves andthen moved parallel to the axes until a “match” is obtained. The value ρ1 of the upper layer resistivity isthen given by the point where the ρa/ρ1 = 1 line cuts the field curve vertical axis and the depth, h, to theinterface by the point where the a/h = 1 line (Wenner) or L/h = 1 line (Schlumberger) cuts the field curvehorizontal axis. The value ρ2 of the lower layer resistivity is determined using the value of k correspondingto the best-matching type curve.

flow ceases, the polarization cells discharge,causing a brief flow of current in the reversedirection. The effect can be measured in severaldifferent ways, all of which can be illustratedby considering a simple square wave.

A square-wave voltage applied to the groundwill cause currents to flow, but the currentwaveform will not be perfectly square and will,moreover, be different in different parts of thesubsurface. Its shape in any given area can bemonitored by observing the voltage developed

between two appropriately placed nonpolariz-ing electrodes. Attainment of the theoreticalsteady voltage, V0, is delayed by polarization forsome time after current begins to flow. Further-more, when the applied voltage is terminated,the voltage drops rapidly not to zero but to amuch lower value, Vp, and then decays slowlyto zero (Fig. 7.9a). Time-domain IP results arerecorded in terms of chargeability, defined intheory as the ratio of Vp to V0, but in practiceit is impossible to measure Vp and voltages

Page 157: Introduction to mineral exploration

140 J. MILSOM

FIG. 7.9 Idealized IP effects in(a) the time domain and (b) thefrequency domain. Not to scale.The voltage applied at the currentelectrodes may be several ordersgreater than the voltage observedacross the voltage electrodes, andVp will usually be at least twoorders of magnitude smallerthan V0.

Voltage observed atvoltage electrodes

Voltage applied at current electrodes (reduced vertical scale)

(a) (b)

Vo

M = Vp /Vo= 100 × Vp /(Vo − Vp)

PFE = 100 × ( ρd.c. – ρh.f.)/ρh.f.

Vp

Vp

ρ ρ ρ

Vo − Vp

are measured after one or more pre-determinedtime delays. The ratios of these to V0 are meas-ures of the polarization effect and are usu-ally quoted in millivolts per volt. Some earlyinstruments measured areas under the decaycurves and results were quoted in milliseconds.

If, as in Fig. 7.9b, the voltage is applied for avery short time only, V0 is never reached and,since resistance is proportional to voltage in allof the array formulae (Fig. 7.7), a low apparentresistivity is calculated. Calculations involv-ing the relationship between this value and thedirect-current resistivity are the basis of fre-quency domain IP. The difference between thetwo divided the by the high frequency resistiv-ity and expressed as a percentage gives themost commonly used parameter, the percentfrequency effect (PFE). In practice, because ofthe need to reverse the current at intervalsto eliminate the effects of natural potentials(see section 7.5), the “direct current” resistiv-ity is actually measured at a low alternatingfrequency, typically 0.25 Hz. Nor can very highfrequencies be used without electromagneticinduction affecting the results, so the “high”frequency may be as low as 4 Hz. Becausemeasurement of IP effects involves arbitrarychoices (of time delays or of frequencies), itis not normally possible to relate resultsobtained with different makes of equipmentquantitatively, and even instruments of thesame make may record slightly different valuesin the same places.

The asymmetry of the observed voltagecurve implies frequency-dependent phase dif-

ferences between it and the primary signal.Advances in solid-state circuitry and precisecrystal-controlled clocks now allow thesedifferences to be measured and plotted asfunctions of frequency. This is phase IP.

Time, frequency, and phase techniques allhave their supporters. Time-domain worktends to require dangerously high voltageand current levels, and correspondingly power-ful and bulky generators. Frequency-domainequipment can be lighter and more portable butthe measurements are more vulnerable toelectromagnetic noise. Phase measurements,which span a range of frequencies, allow thisnoise (electromagnetic coupling) to be estim-ated and eliminated, but add complexity tofield operations.

Coupling can be significant in IP work of anytype and especially severe if cables carryingcurrent pass close to cables connected to volt-age electrodes. Keeping cables apart is easywith the dipole–dipole array (Fig. 7.7e), whichis therefore very popular.

Low frequency resistivity is one of theparameters measured in frequency IP, and canbe calculated in time-domain IP provided thatthe input current is recorded. IP surveys arethus also resistivity surveys. In the frequency-domain it is common practice to calculate ametal factor by dividing the PFE by the resist-ivity and, since this would be a rather smallquantity, multiplying by a factor of the orderof ten thousand. Metal factors emphasize re-gions that are both chargeable and conductiveand can thus be useful for detecting massive

Page 158: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 141

FIG. 7.10 An IP pseudo-section,showing a “pant’s-leg” anomalydue to a shallow chargeable body.

V1 V3V2

0 9

0 3

2 1

n = 1

n = 1 Near-surfacechargeable material

n = 2

n = 2

n = 3

n = 3

n = 4

0 6 1 8

2 0

6 6

4 9

6 6

6 1

8 1

9 2

0 7

3 73 43 95 8

4 14 0

0 7

0 4

0 6

0 8

0 9

0 3

0 4

6 9

6 3

I

sulfides, but may very effectively concealorebodies of other types. A clear chargeabilityanomaly produced by disseminated copper orecan appear less attractive on a metal factor plotthan an adjacent zone of conductive alteration.The moral is that resistivity and IP data shouldeach be respected for the information that theyindividually provide and should not be con-fused into a single composite quantity.

Gradient arrays (Fig. 7.7d) are often used forIP reconnaissance, with results presented asprofiles. Dipole–dipole results may also bepresented in this way, with separate profilesfor different values of the ratio (n) of theinterdipole to intradipole distance. Increasing nincreases the penetration (at the expense, in-evitably, of resolution), and an alternative formof presentation known as the pseudo-section isoften used (Fig. 7.10). Pseudo-sections givesome indications of the depths of chargeable orconductive bodies but can be very misleading.Indications of dip are notoriously ambiguousand it is quite possible for the apparent dip of abody on a pseudo-section to be in the oppositedirection to its actual dip. The classic “pant’sleg” anomaly, shown in Fig. 7.10 and appar-ently caused by a body dipping both ways at45 degrees from the surface, is usually due toa small source near to one specific dipole posi-tion. Every measurement that uses that posi-tion will be anomalous. Such problems arenow addressed by the use of two-dimensionalinversions similar to those used in resistivitywork, which produce depth sections plottedagainst true depth scales. The reliability of theinterpretations can be improved by using high-density acquisition along closely spaced lines.However, the fact remains that the arrays mostcommonly used for IP surveys, and especially

the dipole–dipole array, are poorly adapted toprecise target location. Model studies canprovide a guide to the position of a chargeablemass but it may still require several drillholesto actually find it.

IP is important in base metal explorationbecause it depends on the surface area of theconductive mineral grains rather than theirconnectivity and is therefore especially sens-itive to disseminated mineralisation whichmay produce no resistivity anomaly. In theory,a solid mass of conductive sulfides would givea negligible IP response, but real massive con-ductive ores seem always to be sufficientlycomplex to respond well. Since both massiveand disseminated deposits can be detected, IP isvery widely used, even though it is rather slow,requires moderately large field crews, and isconsequently relatively expensive. The needfor actual ground contact also causes problems,especially in areas of lateritic or caliche over-burden. IP also suffers from the drawbacks in-herent in electrical methods, in responding tographite and barren pyrite (the initials have,rather unkindly, sometimes been said to standfor “indicator of pyrite”) but not to sphalerite.Galena also may produce little in the way of ananomaly, especially in “colloidal” deposits.

There is now much interest in the possibilityof using phase/frequency plots to discriminatebetween different types of chargeable mater-ials. It seems that curve shapes are dictatedby grain size rather than actual mineral typebut even this may allow sulfide mineralisationto be distinguished from graphite because ofits generally smaller grains. Phase and multi-frequency surveys are now commonplace, oftendeliberately extending to frequencies at whichelectromagnetic effects are dominant.

Page 159: Introduction to mineral exploration

142 J. MILSOM

FIG. 7.11 Schematic of anelectromagnetic prospectingsystem. The transmitter (Tx) andreceiver (Rx) coils are horizontaland co-planar. They may also bedescribed as “vertical dipoles,”referring to the magnetic fieldproduced. The system is said tobe “maximum-coupled” becausethe primary field is at right anglesto the plane of the receiver coilwhere it passes through it. Inground surveys this system issometimes denoted by the Swedishterm Slingram. The currentsinduced in the conductor generatean alternating magnetic field thatopposes the primary (transmitted)field, and hence produces anegative anomaly maximum(see Fig. 7.1c). (Drawing based onGrant & West 1965.)

RxTx

Primary magnetic field Secondary magnetic field

Induced electric currents (eddy currents)

7.9 CONTINUOUS-WAVE ELECTROMAGNETICS

Electric currents produce magnetic fields.Conversely, voltages are induced in conductorsthat are exposed to changing magnetic fields,and currents, limited by circuit resistivitiesand self-inductances, will flow in any closedcircuits present. These induced currents pro-duce secondary magnetic fields. Electromag-netic surveys, which measure these fields, useeither continuous, usually sinusoidal, waves(CWEM) or transients (TEM). Transmitters areusually small coils through which electric cur-rents are passed, producing magnetic fields thatdecrease as the inverse cube of distance. Othermethods use long, grounded, current-carryingwires or extended current loops with dimen-sions comparable to the areas being surveyed.Receivers are almost always small coils. Phy-sical contact with the ground is not requiredand electromagnetic methods can thereforebe used from aircraft. Figure 7.11 shows, inschematic form, a system of the type used inexploration for massive sulfide orebodies, us-ing transmitter and receiver loops with theirplanes horizontal and axes vertical. Because

the magnetic field from the transmitter cutsthe plane of the receiver coil at right angles,such systems are said to be maximum-coupled.

In CWEM surveys, alternating currents arepassed through transmitter loops or wires atfrequencies that usually lie between 200 and4000 Hz. There will, in general, be phase differ-ences between the primary and secondaryfields, and anomalies are normally expressed interms of the amplitudes of the secondary fieldcomponents that are in-phase and in phase-quadrature (i.e. 90 degrees out of phase) withthe primary field. For any given body, the am-plitudes (response functions) of the in-phaseand quadrature anomalies are determined by aresponse parameter that varies with conduct-ivity and frequency. For a target consisting ofa simple conducting loop the parameter isequal to the frequency multiplied by the self-inductance and divided by the resistance(Grant & West 1965). In Fig. 7.12 the responseof such a loop to the system of Fig. 7.11 isshown plotted against response parameter. Ifthe loop is made of relatively resistive material,induced currents are small, quadrature signalsare also small (being, in the limit, proportional

Page 160: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 143

FIG. 7.12 Peak responses recordedby a “Slingram” horizontal-loope.m. prospecting system over avertical conducting loop. The sizeof the anomaly is determined bythe response parameter, which isa function of the frequency, f, theself-inductance, L, and theresistance, R, of the loop. Naturalconductors respond in morecomplicated but generally similarways.

Loop response parameter

In-phase response

α2

1 + α2

1.0

1.0 10

=

1002πfL

R

0.01 0.1

0.2

0.4

0.6

0.8

In-p

hase

Quadrature response

Quadratu

re

Res

pons

e α1 + α2

πα

αα

αα

to frequency and inversely proportional toresistivity), and in-phase signals are evensmaller. For a good conductor, on the otherhand, the induced currents are large but almostin phase with the transmitted field. Quadraturesignals are then again small but in-phase signalsare large. The quadrature signal reaches a max-imum at the point where the quadrature andin-phase responses are equal. Similar but morecomplex relationships apply to real geologicalbodies, complicated still further if these havesignificant magnetic permeability. It is evidentthat the in-phase/quadrature ratio, which forthe loop target is equal to the response para-meter, provides crucial information on conduc-tivity. For steeply dipping tabular bodies theeffects of conductivity and thickness are diffi-cult to separate and results are often interpretedin terms of a conductivity-thickness product.In principle, readings at a single frequency aresufficient to determine this, but readings atseveral frequencies improve the chances ofdistinguishing signal from noise.

The depth of penetration achievable with anelectromagnetic system is determined in mostcircumstances by the source–receiver spacing,but there is also an effect associated with thefrequency-dependent attenuation of electro-magnetic waves in conducting media. This isconveniently expressed in terms of the skindepth in which the field strength falls to aboutone-third of its surface value (Fig. 7.13). If re-

ceiver and transmitter are widely separated,penetration will be skin depth limited.

The induced currents circulating in each partof a conductor produce fields which act onevery other part, and computed solutions, ex-cept in very simple cases, have to be obtainedby successive approximations. Considerablecomputer power is needed, even for simpletwo-dimensional bodies, and an alternative isto use physical scale models. A model will bevalid if the ratio of the skin depths in the modeland in the real geology is the same as the ratioof their lateral dimensions. In principle the per-meabilities and dielectric constants could bescaled, but if only the conductivities and lateraldimensions are varied, field instruments canbe used in the laboratory, coupled to miniaturereceiver and transmitter loops. Aluminumsheets may be used to model conductive over-burden and copper sheets to represent sulfideorebodies. Physical models are convenient forthree-dimensional studies, being easily con-figured to represent the complicated combina-tions of orebodies and formational conductorsfound in the real environment.

In early airborne systems, transmitter coilswere mounted on the aircraft and the receivercoils were towed in aerodynamic birds. Prob-lems due to the almost random coupling of thereceiver to the primary field were avoided byrecording only the quadrature signals. Thisled to considerable arguments as to whether

Page 161: Introduction to mineral exploration

144 J. MILSOM

FIG. 7.13 Variation of skin depthwith resistivity and frequency inhomogenous media. The signalamplitude falls to about one third(actually 1/e = 1/2.718) of itsoriginal value in one skin depth.100010

30

300

10

50

500

10,000100

100

ρ (ohm-meters)

1250 Hz

5 kHz20 kHz 80 kHz

Ski

n de

pth

(m)

Skin depth = 500 /f

ρ

ρ

natural conductors existed of such quality thatthe quadrature anomalies would be neglig-ible. The question never seems to have beenformally answered, there being commercialvested interests on both sides, but it does seemto be generally agreed that some good targetsmay produce only very small phase shifts.In any case, as noted above, the in-phase toquadrature ratio is an important interpretationparameter. For CWEM surveys the transmitterand receiver coils are now rigidly mountedeither on the aircraft frame or in a single largetowed bird. Birds are now used almost exclu-sively from helicopters, as civil aviation au-thorities enforce increasingly stringent safetyregulations on fixed wing aircraft.

Originally almost all airborne systems usedthe vertical-loop coaxial maximum-coupledcoil configuration, but multi-coil systems arenow general. Because the source–receiverseparations in airborne systems are inevitablysmall compared to the distances to the con-ducting bodies (even though surveys with sen-sors only 30 m above the ground are becomingcommonplace), anomalies generally amount toonly a few parts per million of the primaryfield. Inevitably, there is some degree of flexurein even the most rigid of coil mountings andthis produces noise of similar amplitude. Insome systems flexure is monitored electronic-ally and corrections are made automatically.

7.10 TRANSIENT ELECTROMAGNETICS (TEM)

One solution to the problem of variable source–receiver coupling is to work in the time-

domain. A magnetic field can be suddenly col-lapsed by cutting off the current producingit, and the magnetic effects of the currentsinduced in ground conductors by this abruptchange can be observed at times when no pri-mary field exists. In the case of a fixed-wing air-craft, the transmitter coil may extend from thetailplane to the nose via the wingtips (Fig. 7.2).The receiver coil may be mounted in a towedbird, the variation in coupling to the primaryfield being irrelevant since this field does notexist when measurements are being made.Helicopter TEM is becoming steadily morepopular, with a number of new systems com-peting in the market. Induced currents attenu-ate quite rapidly in poorly conductive countryrock but can persist for considerable periodsin massive sulfide orebodies (and often ingraphitic shales). Signals associated with thesecurrents were originally observed at four delay-times only (Fig. 7.14), but the number of chan-nels has steadily increased, up to 256 in somerecent instruments.

The TEM method was first patented byBarringer Research of Toronto as the INducedPUlse Transient system (INPUT) and was usedalmost exclusively for airborne surveys. Thesuccess of INPUT (and the skills of its patentlawyers) delayed development of ground-basedtransient systems outside the communist blocfor some years, but by 1968 a Russian instru-ment was being used for surveys in regionsof high-conductivity overburden aroundKalgoorlie. The SIROTEM system was thendeveloped in Australia and, at about the sametime, the Crone PEM system was designed in-dependently in North America. Subsequently

Page 162: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 145

FIG. 7.14 INPUT Mk III transientresponse over the Timminsorebody. Poor quality, largevolume surficial conductors causelarge excursions in Channel 1which die away rapidly at longerdelay times. The orebody responsepersists through all four channels.Modern TEM systems may have asmany as 256 channels but the basicprinciples remain unchanged.(Data from Barringer Research Inc.,Toronto.)

Channel 1

Channel 2

Channel 3

Channel 4

another system, the EM37, was produced byGeonics in Canada. A different approach totransients is found in UTEM systems, whichuse transmitter currents with precisely tri-angular waveforms. In the absence of groundconductivity, the received UTEM signal, pro-portional to the time-derivative of the mag-netic field, is a square-wave. In most TEMsystems (but not UTEM), the transmitter loopcan also be used to receive signals since meas-urements are made after the primary currenthas been terminated. Even if separate loops areused, the absence of primary field at measure-ment time allows the receiver coil to be placedwithin the transmitter loop, which is impract-ical in CWEM because of the very strongcoupling to the primary field.

The Fourier theorem states that any transi-ent wave can be regarded as the sum of a largenumber of sine and cosine waves, and TEMsurveys are therefore equivalent to multi-frequency CWEM surveys. Although firstdeveloped almost entirely for the detection ofmassive sulfide ores, their multi-frequencycharacter allows calculation of conductivity–depth plots that can be used in studies ofporphyry systems, in detection of stratiformorebodies and kimberlites, and even in thedevelopment of water resources. Interpreta-tions of TEM depth-soundings typically useone-dimensional “smooth-model” inversions.

Transient methods are also superficiallysimilar to time-domain IP, and in IP also thereis a time–frequency duality. The most obviousdifference between the electromagnetic and IPmethods is that currents are applied directlyto the ground in IP surveys whereas in most

electromagnetic work they are induced. Morefundamentally, IP systems either sample atdelay times of from 0.1 to 2 seconds (the initialdelay being introduced specifically to avoidelectromagnetic noise) or work at frequenciesat which electromagnetic effects are negligible.Modern IP units are designed to work over wideranges of frequencies to obtain conductivityspectra, and so necessarily, and often deliber-ately, do record some electromagnetic effects,but it is quite possible to avoid working inregions, of either frequency or time-delay,where both effects are both significant.

7.11 REMOTE SOURCE METHODS (VLF ANDMAGNETOTELLURICS)

Life would be much easier for EM field crewsif they had only to make measurements anddid not also have to transmit primary signals.Efforts have therefore been made to usevirtually all the various existing forms of back-ground electromagnetic radiation for geophy-sical purposes. The most useful are militarytransmissions in the 15–25 kHz range (termedvery low frequency or VLF by radio engineers,although very high by geophysical standards)and natural radiation in the 10 Hz to 20 kHzrange produced by thunderstorms (audio-frequency magnetotellurics or AMT). In bothAMT and VLF work, the electromagneticwavefronts are regarded as essentially planar,and in neither is it possible to measure phasedifferences between primary and secondaryradiation. However, differences between thephases of various magnetic and electrical

Page 163: Introduction to mineral exploration

146 J. MILSOM

In-phase(dip angle)

Quadrature

To transmitte

r

Primary

PrimaryPrimary

Dip angleDip angle

H

E

Secondary SecondaryPower

Magnetic

FieldSecondary

magnetic field

Secondaryfield

FIG. 7.15 Dip angle anomaly in the VLF magnetic field. Over homogenous ground the VLF magnetic vectoris horizontal. The largest anomalies are produced when, as in this case, the conductor is elongated in thedirection towards the VLF transmitter (which may be thousands of kilometers away).

components of the waves can be measured andused diagnostically.

When using VLF radiation, the magneticcomponents are generally the most importantif the conductors dip steeply (Fig. 7.15), whilethe electrical components provide informationon the conductivities of flat-lying near-surfacelayers. Because of the high frequencies used,even moderate conductors respond stronglyprovided they are large enough and orientedin directions that ensure good coupling to the

primary field, and the VLF method is especiallyuseful in locating steeply dipping fracturezones.

Magnetotelluric signals offer a wide fre-quency range but poor consistency in sourcedirection and power, and there has been in-creasing use of local artificial sources thattransmit over the same frequency range. Incontrolled-source magnetotellurics (CSAMT),the source is usually a grounded wire severalkilometers long. Ideally, it should be located

Page 164: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 147

M

20

10

30

40

50

60

70

D S B

FIG. 7.16 Seismic reflection profile across a buried submarine channel. Vertical scale is two-way reflectiontime (TWT), in milliseconds. Ten milliseconds TWT is approximately equivalent to 7.5 m of water or 10 mof soft sediments. The reflector indicated by “S” is the sea floor and “B” is a strongly reflecting “basement”surface. “M” is a multiple of the seafloor, created by wavefronts that have made two trips between thereflector and the sea surface. “D” is created by the wave that has traveled directly through the water fromthe source to the receiver. (Reproduced with the permission of Dr D.E. Searle and the Geological Survey ofQueensland.)

far enough from the survey area for plane-waveapproximations to apply, since complex correc-tions are needed where this is not the case.Unfortunately, because distances in the AMTequations are not absolute but are expressed interms of wavelengths, near-field correctionshave to be applied to at least the low frequency,long wavelength data in almost all surveys.For an example of CSAMT use see Fig. 16.3.

7.12 SEISMIC METHODS

Seismic methods dominate oil industry geo-physics but are comparatively little used inmineral exploration, partly because of theirhigh cost but more especially because mostorebodies in igneous and metamorphic rocks

lack coherent layering. One obvious applica-tion is in the search for offshore placers. Sub-bottom profiling using a sparker or boomersource and perhaps only a single hydrophonedetector can allow bedrock depressions, andhence areas of possible heavy mineral accumu-lation, to be identified. Subsea resources ofbulk minerals such as sands and gravels maybe similarly evaluated. The images producedcan be very striking (Fig. 7.16) but, being scaledvertically in two-way reflection time ratherthan depth, need some form of velocity controlbefore they can be used quantitatively.

Where ores occur in sedimentary rocks thathave been only gently folded or faulted, seismicsurveys can be useful. However, reflectionwork onshore is slow and expensive becausedetectors (geophones) have to be positioned

Page 165: Introduction to mineral exploration

148 J. MILSOM

Sin i3 = V1/V3

Sin 2ic = V2/V3

S

i2

i2

ic

2ic

2ic

2ic

ic

V1

V3

V2

Head wave

Head wave

FIG. 7.17 Principles of seismicrefraction for a three layer case. lfVl < V2 < V3, critical refraction canoccur at both interfaces, producingplanar wavefronts known as “headwaves.” Near to the shot point (S)the direct wave will arrive first,but it will soon be overtaken by thewave that travels via the firstinterface and, eventually, by thewave that travels via the second.

slowly, back to the surface (Fig. 7.17). Becausehorizontal velocities estimated from plots ofdistance against time are used in calculationsas if they were vertical velocities, calibrationagainst borehole or other subsurface data isdesirable, the more so since an interface maybe hidden by others if the velocity contrast orunderlying layer thickness is small, and will becompletely undetectable if velocity decreaseswith depth. Despite these limitations, there aremany cases, especially in the assessment ofdeposits of industrial mineral, where refractionsurveys can be useful.

7.13 GROUND RADAR

Ground penetrating radar (GPR) is a relativelynew addition to the geophysical armoury, butreports of mineral exploration applicationsare beginning to appear (e.g. Francké & Yelf2003). GPR results are displayed as images verysimilar to those used for seismic reflectionwork, but penetration is limited to a few tens ofmeters at the best, and may stop at the watertable, which is usually a very strong reflector.In some cases the mapping of the water table,which is typically a surface at which the costsof extracting bulk minerals increase dramatic-

individually by hand and sources may need tobe buried. Also, because of the variability ofnear-surface weathered layers, results requiresophisticated processing and even then areusually less easy to interpret than marine data.The use of reflection in onshore exploration forsolid minerals other than coal is consequentlyrare, although Witwatersrand gold reefs (seesection 14.5.4), flat-lying kimberlite sills (seesection 17.2), and some deep nickel sulfidebodies have all been investigated in this way(Eaton et al. 2003).

Refraction methods use simpler equipmentand need less processing than reflectionmethods but can succeed only if the subsurfaceis regularly layered and velocity increases withdepth at each interface. Generally, only the firstarrivals of energy at each geophone are used,minimizing processing but limiting applica-tions to areas with no more than four signi-ficant interfaces at any one locality. Depthsto interfaces can be estimated because, exceptclose to the shot point, where the direct route isthe quickest, the first arrivals will have beencritically refracted. Critically refracted rays canbe pictured as traveling quite steeply from thesurface to the refractor at the overburden velo-city, then along the refractor at the velocityof the underlying layer, and then steeply, and

Page 166: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 149

16

8

–8

0

15010050 200

A’A

met

res

meters

FIG. 7.18 Radar profile across a sand dune in Namibia. The subhorizontal reflector AA′ is rockhead. Horizontallines are at intervals of 50 ns TWT. Calculated time shifts (static corrections) have been applied to individualtraces to adjust the record to the topographic profile. The vertical scale on the left shows elevations in metersrelative to an off-dune ground-surface datum, assuming a radar wave velocity of 0.16 m ns−1. The data wereobtained using a 100 MHz radar signal, an antenna spacing of 1 m, and a 0.5 m step size. (Image reproduced bycourtesy of Dr C. Bristow, Birkbeck College, University of London.)

ally, is, by itself, sufficient to justify the use ofradar. GPR can also be used onshore to locateriver paleochannels beneath overburden andto define favorable locations within them forplacer deposits. Sedimentary layering can oftenbe imaged within such channels, and also insand dunes (Fig. 7.18), which may include bedsrich in heavy minerals.

7.14 BOREHOLE GEOPHYSICS

Most geophysical techniques can be modifiedfor use in boreholes but neither gravity normagnetic logging is common and seismic vel-ocity and radiometric logs, both very importantin the oil industry, are little used in mineralexploration. In downhole resistivity logging, asin the surface equivalents, four electrodes mustbe used, but usually at least one is outside thehole. In normal logs, one current and one volt-age electrode, a few tens of centimeters apart,are lowered downhole, with the other twoat infinity on the surface. The limitations ofdirect current resistivity in base metal ex-ploration apply downhole as much as on thesurface and more useful results may be ob-tained with IP or electromagnetics. Electrodepolarization can be reduced in downhole IPsurveys by using lead (Pb) electrodes, and canbe eliminated, with some practical difficulty,using porous pots. In electromagnetic work,

small coils are used for both transmitting anddetecting signals, which may be continuouswaves or transients. Transient (TEM) tech-niques are now generally considered the moreeffective, with a proven ability to detect“missed target” massive sulfides (Bishop &Lewis 1996, see Fig. 15.7).

When conductive ore has been intersected,the mise-à-la-masse method can be used. Onecurrent electrode is positioned down the holeand the other is placed on the surface wellbeyond the area of interest. If the ground wereelectrically homogenous (and its surface flat),equipotentials mapped at the surface using onefixed and one mobile voltage electrode wouldcontour as circles centered above the downholeelectrode. If, however, the downhole electrodeis placed at an ore intersection, there are depar-tures from this pattern indicating the strikedirection and extent of continuous mineralisa-tion. Equipotentials are actually diverted awayfrom good conductors that are not in electricalcontact with the electrode. Mise-à-la-masseis thus a powerful tool for investigating thecontinuity of mineralisation between inter-sections in a number of boreholes.

Because mineral exploration holes are norm-ally cored completely, the mineral industry hasbeen slower than the oil industry to recognizethe advantages of geophysical logging. How-ever, the high cost of drilling makes it essentialto obtain the maximum possible information

Page 167: Introduction to mineral exploration

150 J. MILSOM

from each hole, and the electrical character-istics of the rocks surrounding a borehole canoften be better guides to the significance ofan ore intersection, or of the presence of orenearby, than the information contained in afew centimeters of core.

7.15 GEOPHYSICS IN EXPLORATIONPROGRAMS

A logically designed exploration program pro-gresses through a number of stages, fromregional reconnaissance to semi-detailedfollow-up and thence to detailed evaluation.The geophysical component will tend to passthrough the same stages, reconnaissance oftenbeing airborne and the final evaluation perhapsinvolving downhole techniques. The exactroute followed will depend on the nature of thetarget, but a typical base metal or kimberliteexploration program would have, as its firststage, the assembling of all the geophysicaldata already available. Lease conditions may insome cases require the integration of certaintypes of survey data into national databases,and most countries have at least partial mag-netic and gravity coverage and there may havebeen local surveys of these or other types. Withgravity and magnetics it is usually worthwhilegoing beyond the published maps, which oftencontain errors, to the original data.

For sulfide ores or kimberlites, the next stagemight involve a combined airborne magneticand electromagnetic survey, using the smallestpossible terrain clearance and very closelyspaced lines. Semi-regional magnetic mapswould be produced and, with luck, a number ofinteresting electromagnetic anomalies mightbe located. Some sources might be obviousand irrelevant (e.g. powerlines, pipelines,corrugated-iron barns) but any other anomalieswould have to be checked on the ground, withmagnetic and electromagnetic surveys to loc-ate the anomaly precisely, geological mappingto determine rock type, and perhaps a gravitysurvey if density differences are probablebetween ore and country rock. Other electricalmethods, such as IP and even resistivity, mayalso be used to define drilling targets. Groundgeophysics might also be used in areas that had

not been shown to be anomalous from the airbut which appeared geologically or geochemic-ally promising. Once a drilling program hasbegun, not only can the new information be fedback into the geophysical interpretation butthe holes can be used for electrical or radiome-tric logging or, if there are ore intersections,for mise-à-la-masse. Work of this sort may wellcontinue into the development stage.

Not all geophysical exploration follows thispattern. Sometimes only one or two methodshave any chance of success. Seismic tech-niques, whether reflection or refraction, tendto be used alone, being applicable where, as inthe search for placer deposits, other methodsfail, and virtually useless in dealing with thecomplex geologies encountered in hard rockmining. Lateral thinking has also producedsome quite surprising specialized applicationsof geophysical methods, e.g. phosphate gradescan in some cases be estimated by radiometriclogging of uncored boreholes.

7.16 INTEGRATION OF GEOLOGICAL ANDGEOPHYSICAL DATA

Making good use of data from a variety of verydifferent sources is the most testing part ofan exploration geologist’s work, requiring the-oretical knowledge, practical understanding,and a degree of flair and imagination. All thesewill be useless if information is not available inconvenient and easily usable forms. The mosteffective way of recognizing correlations is tooverlay one set of data on another (see sections5.1.8. & 9.2), and in a paper-based explorationprogram decisions on map scales and mapboundaries have to be made at an early stageand adhered to thereafter. Computer-basedgeographical information systems (GIS) offergreater flexibility, although most require a con-siderable investment of time and effort on thepart of the users before their advantages can berealized (discussed in detail in section 9.2).

Presenting different types of data togetheron a single map is generally useful if they areshown as separate entities that can still beclearly distinguished. Combining data of differ-ent types before they are plotted can be danger-ous. This approach is common in exploration

Page 168: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 151

geochemistry, but its extension to geophysicshas not been notably successful. Displays ofmathematically combined gravity and mag-netic data have generally succeeded only indisguising important features while producingno new insights. Allowing computers to cross-correlate geophysics with geology or geochem-istry is even less likely to give meaningfulresults. Correlation is essential but should bedone in the brain of the interpreter.

Of the various possible correlations, thosemost likely to be useful in the early stages ofan exploration program are between geology,air photographs, and magnetic maps. Struc-tural dislocations such as faults and shears arelikely to be indicated magnetically and theintersections of magnetic lineaments havebeen fruitful exploration targets in many areas.Gravity and radiometric data, along withinfrared or radar imagery and soils geochem-istry, may also help in obtaining the best pos-sible understanding of the geology. Geophysicsprovides independent information on the thirddimension (depth), and may be especially use-ful in preparing regional cross-sections.

The inherent ambiguity of all geophysicalsolutions makes it essential to use all possibleadditional information. For there to be even achance of a model being correct, it must satisfyall the known geological and geophysicalconstraints. Consequently, it must also becontinually refined, not only up to but alsothroughout any drilling program. If a worth-while target has been defined but the hole hasproved unsuccessful, it is not enough to simplyblame geophysics. There is a reason for any fail-ure. It may be obvious, as in the case of an IPanomaly produced by barren pyrite, but untilthe reason has been established, a valid targetremains to be tested.

7.17 THE ROLE OF THE GEOLOGIST INGEOPHYSICAL EXPLORATION

A mineral exploration program will usually bethe responsibility of a small team, and whetherthis should include a geophysicist is itself adecision that requires geophysical advice.Factors to be considered are the types of targetssought (and therefore the types of methods

appropriate), the likely duration of the projectand, of course, the availability of personnel. If afull-time geophysicist cannot be justified butsome geophysical techniques are to be used,there may be company geophysicists who canallocate a percentage of their time. Alternat-ively, consultants can be retained and/or one ormore of the team geologists can be given re-sponsibility for day to day geophysical matters,including supervision of contractors and possi-bly the conduct of some surveys. It is usuallybest if one (and preferably only one) geologist isresponsible for routine geophysics, but he orshe must be able to call on more geophysicallyexperienced help when needed and the freedomto seek such help must not be unduly restrictedby short-term cost considerations. Geophysicalsurveys are often among the most expensiveparts of an exploration program and muchmoney can be wasted by a few apparentlytrivial, but wrong, decisions.

Project geologists should not be expected todecide, unaided, on methods, instrument pur-chases, and overall geophysical explorationstrategy. In companies with their own geo-physicists it would be unusual for them to haveto do so, but where these are not available it isall too common for consultants to be called inonly when things have already gone irretriev-ably wrong. The instruments purchased, andused, the contracts let, and the contractorsselected may all have been totally unsuitablefor the work required.

What, then, can reasonably be expected ofa project geologist whose only experience ofgeophysics may be a few dimly rememberedlectures on a university course several years inthe past, possibly accompanied by some miser-able days in the field pressing the keys on a pro-ton magnetometer or trying to level a gravitymeter? First, and perhaps most important, is adegree of humility and a willingness to seekhelp. Second, a variety of geophysical textsare required and a readiness to read them all(beware of relying on any single authority).Third, a realistic attitude is needed. Explora-tion programs have limited and usually pre-defined budgets, and geophysical instrumentsare neither magic wands that reveal everythingabout an area, nor pieces of useless electroniccircuitry inflicted on hard-working geologists

Page 169: Introduction to mineral exploration

152 J. MILSOM

specifically to complicate their lives. In con-sidering geophysics, geologists must ask them-selves a number of questions. What informationis needed? Is there a realistic expectation thatgeophysical methods will provide it? What in-formation will they not provide? Is this infor-mation essential and can any incompletenessbe tolerated? Is geophysics affordable and (evenif the answer is yes) would the money be betterspent on something else? Team geologists maynot be able to answer all these questions them-selves but they need to ask them, and to getanswers. The necessary approaches can beillustrated by specific examples:1 An exploration area is some hundreds ofsquare kilometers in extent and is coveredby regional aeromagnetic maps. Are theseadequate or should a new, detailed survey beplanned? (Seek geophysical advice, but notfrom a contractor who may be hoping forwork.) Is it reasonable to suppose that thetargets will be magnetic or that there will becritical structural or lithological informationin the magnetic data? Will such data only beobtained by high sensitivity or gradient surveysor will a 1 nT magnetometer do? If aeromag-netic work is to be done, is there a case foradding other sensors (electromagnetic, radio-metric), bearing in mind the inevitable in-creases in cost? Is the area topographicallysuitable for fixed-wing survey or will a heli-copter have to be used? How much help willhave to be bought in for contract supervisionand interpretation, and can this be accom-modated within the existing budget?2 A lease of a few square kilometers is to beexplored for massive sulfides. SP surveys havelocated some mineralized bodies in the regionbut have missed others. Bearing in mind thatother, more expensive methods will also haveto be used if the area is to be fully evaluated,is more SP worthwhile? Should it be kept inreserve (the equipment is cheap) in case thereare field hands spare for a few days? Or, havinggone to the trouble of gridding and pegginglines, should not every possible method be usedeverywhere?

Geophysical work will involve surveying,and may also require line clearing and cutting.At this point the geologist responsible for geo-physics may well come into conflict with thegeologist responsible for geochemistry, compli-

cating their lives even further. Geochemists,once beyond the reconnaissance stream sedi-ment stage, also take samples along peggedlines, but like to be able to plot locations at re-gular intervals on planimetric base maps. Plani-metrically regular spacing (secant-chaining)is also desirable for geophysical point data(gravity, magnetic, and SP), but in resistivity,IP, and electromagnetic surveys it is the actualdistances between coils or electrodes that mustbe constant and these may differ significantlyfrom plan distances. Whether lines are secant-or slope-chained will depend on how much,of what, is being measured, but this apparentlytrivial problem can be a major cause of frictionbetween team members.

The final requirements of all explorationgeologists, and particularly of those who in-volve themselves in geophysics, are serendipityand an awareness of the possible. Geophysicalsurveys always provide some information,even though it may not seem directly relevantat first, and many mineral deposits have beendiscovered by what may be considered luck butat least required geologists with their witsabout them. In the 1960s, radiometric surveyswere flown in the Solomon Islands primarilyfor detecting phosphates. A ground follow-upteam sent to Rennel Island expected to findphosphatic rock, of which there was very little,but did recognize that the source of theanomaly was uraniferous bauxite, in mineablequantities.

7.18 DEALING WITH CONTRACTORS

Exploration geologists may be involved withgeophysical contractors at many differentstages, and will often need the support andadvice of an experienced geophysicist who iscommitted to them and not to the contractor.The most important stage is when the contractis first being drafted. Geophysical consultantsfind it frustrating (but also sometimes veryprofitable) to be asked to sort out contracts,already written and agreed, which are either soloose that data of almost any quality must beaccepted or which specify procedures that actu-ally prevent useful data being acquired. Projectgeologists will share in the frustration but notin the profit!

Page 170: Introduction to mineral exploration

7: GEOPHYSICAL METHODS 153

Survey parameters such as line separationand line orientation, and all specifications re-lating to final reports and data displays, shouldbe decided between the project geologistsand their geophysical adviser. More technicalquestions, such as tolerances (navigationaland instrumental), instrument settings, flyingheights in airborne surveys, and data controland reduction procedures, should be left to thegeophysicist. Geologists who set specificationswithout this sort of help, perhaps relying on thecontractor’s advice and goodwill, are takinggrave risks. Goodwill is usually abundant dur-ing initial negotiations, and contractors will,after all, only stay in business if they keep theirclients reasonably happy, but as deadlines drawcloser, and profit margins come to look tighter,it becomes more and more tempting to rely onthe letter of the contract.

Once a contract has been signed, the workhas to be done. The day-to-day supervision canusually be left to the geologists on the spot,who will find a comprehensive and tightlywritten contract invaluable. The greater thedetail in which procedures are specified, theeasier it is to see if they are being followedand to monitor actual progress. Safety and theenvironment are becoming more importantevery year, and although the contract may(and should) specify that care for these are theresponsibility of the contractor, the companycommissioning the work will shoulder muchof the blame if things go wrong. Random andunannounced visits to a field crew, especiallyearly in the morning (are they getting up ata reasonable hour, without hangovers, and dothey have precise and sensible plans for theday’s work?) or at the end of the day (have theyworked a full day and are their field notes forthat day clear and intelligible?), are likely tobe more useful than any number of formalmeetings with the party chief. Evening visitsare particularly useful, revealing whether datareductions and paperwork are keeping pacewith the field operations and whether theevenings are being passed in an alcoholic haze.

A geophysical contractor can reasonably beexpected to acquire data competently, with dueregard to safety and environmental protection,reduce it correctly, and present results clearly.Many also offer interpretations, but these maybe of limited value. Few contractors are likely

to have anything approaching the knowledge ofa specific area and its geological problems thatis possessed by the members of an explorationteam. Moreover, mining companies are oftenreluctant to provide contractors, who may nextweek be working for a competitor, with re-cently acquired data. Since such informationcan be vital for good interpretation, contrac-tors’ reports are often, through no fault of theirown, bland and superficial. They may only beworth having if the method being used is sonew or arcane that only the operating geophysi-cists actually understand it.

7.19 FURTHER READING

A number of textbooks cover the geophysicaltechniques used in mineral exploration in somedetail. Of these, the most comprehensive isApplied Geophysics (Telford et al. 1990). Abriefer but still excellent coverage is providedby An Introduction to Geophysical Explora-tion (Kearey et al. 2002), while Principles ofApplied Geophysics (Parasnis 1996) has astrong mineral exploration bias, derived fromthe author’s own background in Scandinavia.Interpretation Theory in Applied Geophysics(Grant & West 1965) is now hard to find butalso hard to beat as a summary of the theoret-ical backgrounds to most geophysical methods.Practical Geophysics II (van Blaricom 1993)and Geophysics and Geochemistry at theMillenium (Gubins 1997) provide a wealthof largely North American case histories andpractical discussions but both tend to dis-integrate if taken into the field. A third editionof Practical Geophysics was produced in 1998,on CD-ROM only. Practical aspects of smallscale ground-based geophysical surveys andsome elementary interpretational rules arediscussed in Field Geophysics (Milsom 2002).Valuable case histories are to be found inmany journals, but especially in Geophysicsand Exploration Geophysics, published bythe Society of Exploration Geophysicists andthe Australian Society of Exploration Geo-physicists respectively. The same two organ-izations publish newsletter journals (TheLeading Edge and Preview) that provide up-to-the-minute reviews and case histories. ThreeAustralian compilations, on orebodies at Elura

Page 171: Introduction to mineral exploration

154 J. MILSOM

ical techniques have been applied to specificbase metal deposits. Hoover et al. (1992)provide similar information with a WesternHemisphere emphasis.

(Emerson 1980), Woodlawn (Whitely 1981), andthroughout Western Australia (Dentith et al.1995), provide useful, although now somewhatdated, case histories where multiple geophys-

Page 172: Introduction to mineral exploration

8Exploration

GeochemistryCharles J. Moon

Geochemistry is now used in virtually everyexploration program, if only to determine thegrade of the material to be mined. Howeverexploration geochemistry has evolved from itsearly origins in assaying, to using the chem-istry of the environment surrounding a depositin order to locate it. This particularly applies tothe use of surficial material, such as soil, till,or vegetation, that can be used in areas wherethere is little outcrop. The object is to define ageochemical anomaly which distinguishes thedeposit from enhancements in background andnonsignificant deposits. This chapter explainshow geochemistry may be employed in thesearch for mineral deposits. Further details ofthe theory behind exploration geochemistryare given in the exhaustive but dated Rose et al.(1979) and Levinson (1980).

Exploration geologists are likely to be moredirectly involved in geochemistry than withgeophysics which is usually conducted by con-tractors and supervised by specialist geophy-sicists. A geochemical program can be dividedinto the following phases:1 Planning;2 Sampling;3 Chemical analysis;4 Interpretation;5 Follow-up.

The field geologist will probably carry outphases 1, 2, 4, and 5, while analysis is normallyperformed by a commercial laboratory.

8.1 PLANNING

The choice of the field survey techniqueand the analytical methods depends on the

commodity sought and its location. In the sameway as geological and grade–tonnage modelsare generated (seesection 4.1.3) modeling canbe extended to include geochemical factors,summarized in Barton (1986). Thus the geolo-gist will start with a knowledge of the elementsassociated with a particular deposit type, anidea of the economic size of the deposit tobe sought, the mineralogical form of the ele-ments, and the probable size of the elementalanomalies around it. The outline of a deposit isdefined by economic criteria and the mineablematerial is surrounded by lower concentra-tions of the mined elements which are how-ever substantially enriched compared withunmineralized rock. This area of enrichment isknown as the primary halo, by analogy withthe light surrounding the outline of the moon,and the process of enrichment as primary dis-persion. In addition ore-forming processes con-centrate or deplete elements other than thosemined. For example, massive sulfide depositsoften contain substantial arsenic and goldin addition to the copper, lead, and zinc forwhich they are mined. A summary of typicalelemental associations is shown in Table 8.1.

The geologist’s problem is then to adaptthis knowledge of primary concentration to theexploration area. The geochemical response atthe surface depends on the type of terrain andespecially on the type of material covering thedeposit as shown in Fig. 8.1. The response in anarea of 2 m of residual overburden is very differ-ent from that of an area with 100 m deep cover,or if the overburden has been transported. Alsoelements behave differently in the near-surfaceenvironment from that in which the depositformed. For example, in cases where copper,

Page 173: Introduction to mineral exploration

156 C.J. MOON

TABLE 8.1 Elemental associations and associated elements (pathfinders) useful in exploration. (Largely fromRose et al. 1979 with some data from Beus & Grigorian 1977, p. 232, and Boyle 1974.)

Type of deposit Major components Associated elements

Magmatic depositsChromite ores (Bushveld) Cr Ni, Fe, MgLayered magnetite (Bushveld) Fe V, Ti, PImmiscible Cu–Ni–sulfide (Sudbury) Cu, Ni, S Pt, Co, As, AuPt–Ni–Cu in layered intrusion (Bushveld) Pt, Ni, Cu Sr, Co, SImmiscible Fe–Ti–oxide (Allard Lake) Fe, Ti PNb–Ta carbonatite (Oka) Nb, Ta Na, Zr, PRare–metal pegmatite Be, Li, Cs, Rb B, U, Th, rare earths

Hydrothermal depositsPorphyry copper (Bingham) Cu, S Mo, Au, Ag, Re, As, Pb, Zn, KPorphyry molybdenum (Climax) Mo, S W, Sn, F, CuSkarn–magnetite (Iron Springs) Fe Cu, Co, SSkarn–Cu (Yerington) Cu, Fe, S Au, AgSkarn–Pb–Zn (Hanover) Pb, Zn, S Cu, CoSkarn–W–Mo–Sn (Bishop) W, Mo, Sn F, S, Cu, Be, BiBase metal veins Pb, Zn, Cu, S Ag, Au, As, Sb, MnSn–W greisens Sn, W Cu, Mo, Bi, Li, Rb, Si, Cs, Re, F, BSn–sulfide veins Sn, S Cu, Pb, Zn, Ag, SbCo–Ni–Ag veins (Cobalt) Co, Ni, Ag, S As, Sb, Bi, UEpithermal precious metal Au, Ag Sb, As, Hg, Te, Se, S, CuSediment hosted precious metal (Carlin) Au, Ag As, Sb, Hg, WVein gold (Archaean) Au As, Sb, WMercury Hg, S Sb, AsUranium vein in granite U Mo, Pb, FUnconformity associated uranium U Ni, Se, Au, Pd, AsCopper in basalt (L. Superior type) Cu Ag, As, SVolcanic-associated massive sulfide Cu Cu, S Zn, AuVolcanic-associated massive sulfide Zn, Pb, Cu, S Ag, Ba, Au, As

Zn–Cu–PbAu–As rich Fe formation Au, As, S SbMississippi Valley Pb–Zn Zn, Pb, S Ba, F, Cd, Cu, Ni, Co, HgMississippi Valley fluorite F Ba, Pb, ZnSandstone-type U U Se, Mo, V, Cu, PbRed bed Cu Cu, S Ag, Pb

Sedimentary typesCopper shale (Kupferschiefer) Cu, S Ag, Zn, Pb, Co, Ni, Cd, HgCopper sandstone Cu, S Ag, Co, NiCalcrete U U V

lead, and zinc are associated in volcanic-associated massive sulfide deposits, zinc isnormally more mobile in the surface environ-ment than copper and much more so thanlead. Lead is more likely to be concentratedimmediately over the deposit, as it is relativelyinsoluble, whereas zinc will move or dispersefrom the deposit. This process of movementaway from the primary source is termed sec-ondary dispersion and it can also be effected

by mechanical movement of fragments undergravity, movement as a gas, or diffusion of theelements in the form of ions as well as move-ment in solution.

The background levels of an element inrocks and soils also have to be consideredwhen trying to find secondary dispersion fromdeposits. All elements are present in every rockand soil sample; the concentration will dependon the mode of formation of the rock and the

Page 174: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 157

2–5 m

Residual

Overburden

Rock

Ice

Glacial

Till 2

Till 1

Rock

>50 m

Deep Tropical

Overburden

Rock

FIG. 8.1 Sketch showing dispersion through majortypes of overburden.

largely be determined by the overburden con-ditions but the density and exact nature ofthe sample will be based either on previousexperience in the area or, if possible, on anorientation survey.

8.1.1 Orientation surveys

One of the key aspects of planning is to evalu-ate which techniques are effective for the com-modity sought and in the area of search. Thisis known as an orientation survey. The bestorientation survey is that in which a variety ofsampling methods is tested over a prospect ordeposit of similar geology to the target and insimilar topographical conditions to determinethe method which yields the best results. Achecklist for an orientation study is givenbelow (Closs & Nichol 1989):1 Clear understanding of target deposit type;2 Understanding of surficial environment ofthe search area;3 Nature of primary and secondary dispersionfrom the mineralisation;4 Sample types available;5 Sample collection procedures;6 Sample size requirements;7 Sample interval, orientation, and arealdensity;8 Field observations required;9 Sample preparation procedures;10 Sample fraction for analysis;11 Analytical method required;12 Elemental suite to be analyzed;13 Data format for interpretation.

The relatively small cost involved in under-taking an orientation survey compared withthat of a major geochemical survey is alwaysjustified and whenever possible this approachshould be used. However, if a physical orienta-tion survey is not possible, then a thoroughreview of the available literature and discus-sion with geochemical experts is a reasonablealternative option. Of particular use in orien-tation studies are the models of dispersion pro-duced for various parts of the world; CanadianShield and Canadian Cordillera (Bradshaw1975), Western USA (Lovering & McCarthy1977), and Australia (Butt & Smith, 1980).

Particularly useful discussion on orientationprocedures is provided by Thomson (1987) andCloss and Nichol (1989).

process forming the soil. Indications of back-ground levels of elements in soils are givenin Table 8.2, as are known lithologies withelevated concentrations which might providespurious or non-significant anomalies duringa survey. These background levels can be ofuse in preparing geochemical maps whichcan be used to infer lithology in areas of pooroutcrop. The reader is advised to get some ideaof the background variation over ordinary rockformations from a geochemical atlas, such asthose for England and Wales (Webb et al. 1978),Alaska (Weaver et al. 1983), the former WestGermany (Fauth et al. 1985), and Europe(Salmimen et al. 2004).

The basis of a geochemical program is a sys-tematic sampling program (Thomson 1987) andthus decisions must be made in a cost-effectivemanner as to the material to be sampled, thedensity of sampling, and the analytical methodto be employed. Cost/benefit ratios should beconsidered carefully as it may be that a slightlymore expensive method will be the only effect-ive technique. The material to be sampled will

Page 175: Introduction to mineral exploration

158 C.J. MOON

TABLE 8.2 Background concentration of trace elements and utility in geochemical exploration. (Largely afterRose et al. 1979, Levinson 1980.)

Element

AntimonyArsenic

Barium

BerylliumBoronBismuthCadmiumChromium

Cobalt

Copper

FluorineGoldLeadLithiumManganese

Mercury

Molydenum

NickelPlatinum

Rare earthsSeleniumSilverTelluriumThalliumTinTungsten

Uranium

VanadiumZinc

Zirconium

Use in exploration

PathfinderPathfinderespecially for AuPanned concen-trations in VMSsearchOccasional useFor borates

Zn depositsChromite in pannedconcs.Widely used

Most surveys

F depositsGold depositsWide useTin depositsScavenges Co, Zn,AgWide use aspathfinderWide use

Wide useDifficult todetermine

Little useDifficult to useDifficult to useEpithermal AuPanned concentratesScheelite fluorescesin UVDetermine in water

Little useMost surveys

Little use

Typicalbackgroundconcentrationin soils (ppm)

110

300

500 ppb30500 ppb100 ppb45

10

15

3001 ppb1520300

50 ppb

3

171 ppb

La 30300 ppb100 ppb?10 ppb200 ppb?101

1

5535

270

Enrichedlithologies

Ironstones

Sandstones

GranitesGranitesGranitesBlack shalesUltramafics

Ultramafics

Basic igneous

Alkaline igneousBlack shalesSandstonesGranitesVolcanics

Black shale

UltramaficsUltramafics

Beach sandsBlack shales

Acid intrusives

GranitesGranites

Phosphorites

UltramaficsBlack shales

Alkaline igneous

Surficial mobility

LowMobile; FescavengedLow often barite

High if not in berylModerateLowHighLow

Moderate MnscavengedpH > 5 low, elsemoderateHighLowLowModerateModerate, high atacid pHHigh

Moderate to highpH > 10Low, scavengedVery low

Very lowHighHigh; Mn scavengedLowLowVery lowVery low

Very high; organicscavengedModerate?High; scavenged byMnVery low

Page 176: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 159

After the orientation survey has been con-ducted the logistics of the major survey needto be planned. The reader should devise achecklist similar to that of Thomson (1987).1 Hire field crew with appropriate experienceand training;2 Obtain base maps and devise simple samplenumbering scheme;3 Designate personnel for communicationwith the laboratory;4 Arrange quality control of laboratory;5 Arrange for data handling and interpretation;6 Organize archive of samples and data;7 Liaise with other project staff (e.g. geophysi-cists) and arrange reporting to management.

Reporting of geochemical surveys is import-ant and readers should consult the short book-let by Bloom (2001). This provides a wealth ofinformation as well as the basis for fulfillinglegal requirements in Canada.

8.2 ANALYSIS

As the geologist generally sees little of theprocess of analysis, which is usually done atsome distance from the exploration project,analytical data tend to be used uncritically.While most laboratories provide good qualitydata they are usually in business to make aprofit and it is up to the geologist to monitorthe quality of data produced and investigatethe appropriateness of the analytical methodsused.

8.2.1 Accuracy and precision

The critical question for the geologist is howreproducible the analysis is and how repres-entative of the “correct” concentration theconcentration is, as shown in Fig. 8.2. Thereproducibility of an analysis is termed theprecision and its relation to the expected orconsensus value the accuracy. For most pur-poses in exploration geochemistry it is vitallyimportant that an analysis is precise but theaccuracy is generally not so crucial, althoughsome indication of the accuracy is needed. Atthe evaluation stage the analyses must be pre-cise and accurate. The measurement of accur-acy and precision requires careful planning andan understanding of the theory involved.

A number of schemes have been devised butthe most comprehensive is that of Thompson(1982). Precision is measured by analyzingsamples in duplicate whereas accuracy requiresthe analysis of a sample of known composi-tion, a reference material. The use of duplicatesamples means that precision is monitoredacross the whole range of sample composi-tions. Reference materials can be acquiredcommercially but they are exceedingly expen-sive (>US$100 per 250 g) and the usual practiceis to develop in-house materials which arethen calibrated against international referencematerials. The in-house reference materialscan be made by thoroughly mixing and grind-ing weakly anomalous soils from a variety ofsites. The contents should be high enough togive some indication of accuracy in the anti-cipated range but not so high as to requirespecial treatment (for copper, materials in therange 30–100 ppm are recommended). If com-mercial laboratories are used then the referencematerials and duplicates should be includedat random; suggested frequencies are 10% forduplicates and 4% for reference materials. Ifthe analyses are in-house then checks can bemade on the purity of reagents by runningblank samples, i.e. chemicals with no sample.Samples should if possible be run in a differentand random order to that in which they arecollected. This enables the monitoring of sys-tematic drift. In practice this is often not easy

Precise, accurate

n

n

Imprecise, accurate

n

n

Concn

Precise, inaccurate Imprecise, inaccurate

FIG. 8.2 Schematic representation of precision andaccuracy assuming normal distribution of analyticalerror.

Page 177: Introduction to mineral exploration

160 C.J. MOON

and it is not desirable to operate two number-ing systems, so some arrangement should bemade with the laboratories to randomize thesamples within an analytical batch. The resultscan then easily be re-sorted by computer.

When laboratory results become availablethe data should be plotted batch by batch toexamine the within- and between-batch effects.The effects observable within a batch are pre-cision and systematic instrumental drift. Pre-cision should be monitored by calculating theaverage and difference between duplicates on achart such as is detailed by Thompson (1982).The precision for most applications should beless than about 20%, although older methodsmay reach 50%. If precision is worse than anti-cipated then the batch should be re-analyzed.Systematic drift can be monitored from thereference materials within a batch and thereagent blanks. Between-batch effects representthe deviation from the expected value of thereference material and should be monitored bycontrol charts. Any batch with results from ref-erence materials outside the mean ±2 standarddeviations should be re-analyzed; commerciallaboratories normally do this at no furthercharge. Long-term monitoring of drift can showsome interesting effects. Coope (1991) was ableto demonstrate the disruption caused by themoving of laboratories in a study of the gold ref-erence materials used by Newmont Gold Inc.

8.2.2 Sample collection and preparation

Samples should be collected in nonmetalliccontainers to avoid contamination. Kraft paperbags are best suited for sampling soils andstream sediments because the bags retain theirstrength if the samples are wet and the samplescan be oven dried without removing them fromtheir bags. Thick gauge plastic or cloth are pre-ferred for rock samples. All samples should beclearly labeled by pens containing nonmetallicink.

Most sample preparation is carried out in thefield, particularly when it involves the collec-tion of soils and stream sediments. The aim ofthe sample preparation is to reduce the bulk ofthe samples and prepare them for shipment.Soils and stream sediments are generally driedeither in the sun, in low temperature ovens, orfreeze dried; the temperature should be below

65°C so that volatile elements such as mercuryare not lost. Drying is generally followed bygentle disaggregation and sieving to obtainthe desired size fraction. Care should be takento avoid the use of metallic materials and toavoid carryover from highly mineralized tobackground samples. Preparation of rocks andvegetation is usually carried out in the labor-atory and care should be taken in the selec-tion of crushing materials. For example, in arock geochemical program a company search-ing for volcanic-associated massive sulfidesfound manganese anomalies associated witha hard amphibolite. They were encouraged bythis and took it as a sign of exhalative activity.Unfortunately further work showed that themanganese highs were related to pieces of man-ganese steel breaking off the jaw crusher andcontaminating the amphibolite samples. Otherless systematic variation can be caused bycarryover from high grade samples, for examplenot cleaning small grains of mineralized veinmaterial (e.g. 100,000 ppb Au) will cause signi-ficant anomalies when mixed with background(1 ppb Au) rock. Contamination can be elimin-ated by cleaning crushing equipment thor-oughly between samples and by checking thisby analyzing materials such as silica sand.

The Bre-X scandal, in which alluvial goldwas added to drill pulps before sending thesamples for analysis (see section 5.4), empha-sizes the importance of recording the methodsof sample preparation and controlling access tothe samples.

8.2.3 Analytical methods

Most analysis is aimed at the determination ofthe elemental concentrations in a sample andusually of trace metals. At present it is impos-sible to analyze all elements simultaneouslyat the required levels, so some compromiseshave to be made (Fig. 8.3). In exploration forbase metals it is usual to analyze for the ele-ments sought, e.g. copper in the case of a copperdeposit, and as many useful elements as pos-sible at a limited extra cost. With modern tech-niques it is often possible to get 20–30 extraelements, including some that provide littleextra information but a lot of extra data for in-terpretation. The major methods are as shownin Table 8.3 but for detail on the methods used

Page 178: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 161

Other methods

Neutron activation analysis

X-ray fluorescence

Atomic absorption spectrophotometry

Inductively coupled mass spectrometry

Fire assay preconcentration various finishes

Inductively coupled emission spectrometry

H

Li

Na

K

Rb

Cs

Fr

Be

Mg

Ca

Sr

Ba

Ra

Sc

Y

La

Ac

Ti

Zr

Hf

V

Nb

Ta

Cr

Mo

W

Mn

Tc

Re

Fe

Ru

Os

Co

Rh

Ir

Co

Rh

Ir

Ni

Pd

Pt

Cu

Ag

Au

Zn

Cd

Hg

B

Al

Ga

In

Tl

C

Si

Ge

Sn

Pb

N

P

As

Sb

Bi

O

S

Se

Te

Po

F

Cl

Br

I

At

Ne

Ar

Kr

Xe

Rn

He

Ce

Th

Pr

Pa

Nd

U

Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

FIG. 8.3 Summary periodic table showing cost-effective methods of elemental analysis at levels encounteredin background exploration samples.

TABLE 8.3 Summary of the main methods used in exploration geochemistry.

Method Capital cost $ Multielement Precision Sample type Cost per comments sample(approx.) ($US)

Colorimetry 8000 No Poor Solution 2–10 Good for field useand W, Mo

Atomic absorption 60,000 No Good Solution 1–5 Cheap and precisespectrophotometryX-ray fluorescence 200,000 Yes Good Solid 20 Good for

refractoryelements

ICP–ES 150,000 Yes Good Solution 10 Good fortransition metals

ICP–MS 150,000– Yes Good Solution 10 Good for heavy1000,000 metals. High

resolutioninstruments verylow detectionlimits

Page 179: Introduction to mineral exploration

162 C.J. MOON

the reader should consult Fletcher (1981, 1987)and Thompson and Walsh (1989).

The differences between the methods shownare cost, the detection limits of analysis, speedof analysis, and the need to take material intosolution. Most general analysis in developedcountries is carried out by inductively coupledplasma emission spectrometry (ICP–ES), oftenin combination with inductively coupledplasma mass spectrometry (ICP–MS), or X-rayfluorescence (XRF). All three methods requirehighly sophisticated laboratories, pure chem-icals, continuous, nonfluctuating power sup-plies, and readily available service personnel,features not always present in developingcountries. In less sophisticated environments,high quality analysis can be provided by atomicabsorption spectrophotometry (AAS), whichwas the most commonly used method indeveloped countries until about 1980. Anothermethod which is widely used in industry isneutron activation analysis (NAA) but its useis restricted to countries with cheap nuclearreactor time, mainly Canada.

Precious metals (gold and platinum groupelements) have been extremely difficult to de-termine accurately at background levels. Theboom in precious metal exploration has, how-ever, changed this and commercial laboratoriesare able to offer inexpensive gold analysis atgeochemical levels (5 ppb to 1 ppm) using solv-ent extraction and AAS, ICP–ES or alternat-ively NAA on solid samples. For evaluation themethod of fire assay is still without equal: inthis the precious metals are extracted into asmall button which is then separated from theslag and determined by AAS, ICP–ES, or ICP–MS. The analysis of precious metals is differentfrom most major elements and base metals inthat large subsamples are preferred to overcomethe occurrence of gold as discrete grains. Typic-ally 30 or 50 g are taken in contrast to 0.25–1 gfor base metals; in Australia, 8 kg are oftenleached with cyanide to provide better samp-ling statistics (further details in section 16.1.2).

Elements which occur as anionic speciesare generally difficult to measure, especiallythe chloride, bromide, and iodide ions whichserve as some of the ore transporting ligands.Although some of these elements can bedetermined by ICP–ES or XRF, the most usefulmethod is ion chromatography.

Isotopic analysis is not yet widely used inexploration although some pilot studies, suchas that of Gulson (1986), have been carriedout. The main reason for this is the difficultyand cost of analysis. Although high resolutionICP sourced mass spectrometers are findingtheir way into commercial laboratories and arethe main hope for cheap analysis, they are notyet routine.

The choice of analytical method will aimat optimizing contrast of the main target ele-ment. For example, it is little use determiningthe total amount of nickel in an ultramaficrock when the majority of nickel is in olivineand the target sought is nickel sulfides. Itwould be better in this case to choose a reagentwhich will mainly extract nickel from sulfidesand little from olivine. In soils and streamsediments, optimum contrast for base metals isnormally obtained by a strong acid (e.g. nitric +hydrochloric acids) attack that does not dis-solve all silicates, and an ICP–ES or AAS finish.Most rock analysis uses total analysis by XRFor by ICP–ES/ICP–MS following a fusion ornitric–perchloric–hydrofluoric acid attack.

8.3 INTERPRETATION

Once the analytical data have been receivedfrom the laboratory and checked for precisionand accuracy, the question of how the data istreated and interpreted needs to be addressed.As the data are likely to be multi-element andthere are likely to be a large number of samplesthis will involve the use of statistical analysison a computer. It is recommended that the dataare received from the analyst either in the formof a floppy disk, CD-ROM, or over the Internet.Re-entering data into a computer from a papercopy is expensive and almost certain to intro-duce major errors. Normally the data will betransferred into an electronic database (seesection 9.1) to allow easy access or, in the caseof small data sets stored in a spreadsheet.

8.3.1 Statistics

The object of geochemical exploration is todefine significant anomalies. In the simplestcase these are the highest values of the ele-ment sought but they could be an elemental

Page 180: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 163

association reflecting hydrothermal alterationor even element depletion. Anomalies aredefined by statistically grouping data andcomparing these with geology and samplinginformation. Normally this grouping will beundertaken by computer and a wide variety ofstatistical packages are available; for micro-computers, one of MINITAB, SYSTAT, andSTATISTICA is recommended at a cost of$US200–500 each.

The best means of statistically grouping datais graphical examination using histograms andbox plots (Howarth 1984, Garrett 1989). Thisis coupled with description using measures ofcentral tendency (mean or median) and of stat-istical dispersion (usually standard deviation).It would be expected that if data are homo-genous then they will form a continuousnormal or, more likely, log-normal distribu-tion but if the data fall into several groupsthen they will be multi-modal as shown in theexample in Fig. 8.4. These are a set of copper

determinations of soil samples from the DaisyCreek area of Montana and have been describedin detail by Sinclair (1991). The histogramshows a break (dip) in the highly skewed dataat 90 and 210 ppm. This grouping can be wellseen by plotting the data with a probabilityscale on the x axis (Fig. 8.5); log-normal dis-tributions form straight lines and multi-modalgroups form straight lines separated by curves(for full details see Sinclair 1976). The DaisyCreek data show thresholds (taken at the upperlimit, 99%, of the lower subgroups) which canbe set at 100 ppm to divide the data into twogroups and at 71 and 128 ppm if three groupsare used. The relationship of the groups toother elemental data should then be examinedby plotting and calculating the correlationmatrix for the data set. In the Daisy Creekexample there is a strong correlation (0.765)between copper and silver (Fig. 8.6) which re-flects the close primary association betweenthe two elements, whereas the correlationbetween copper and lead is low (0.12) reflect-ing the spatial displacement of galena veinsfrom chalcopyrite. If geological or samplingdata are coded it will be possible to comparethese groups with that information.

The spatial distribution of the data mustthen be examined by plotting elemental datausing the intervals that delineate the groups orby plotting the data by percentiles (percentageof data sorted into ascending order); 50, 75, 90,

100

80

60

40

20

010 110 210 310 410 510

Fre

quen

cy

500 - 980

Copper (ppm)

(a)

40

30

20

10

010 100 1000

Fre

quen

cy

Log(10)copper (ppm)

(b)

FIG. 8.4 Arithmetic and log10 transformedhistograms of Daisy Creek copper data. Note thepositive skew of the arithmetic histogram. Forfurther discussion see text and Sinclair 1991.

Log

(ppm

)

0.80

1.28

1.76

2.24

2.72

99 95 85 70 50 30 15 5 1

Cumulative (%)

l

l

ll

ll

ll

ll

ll

ll

ll

ll

l

ll

lCu

A : 12.6%

B : 87.4%

FIG. 8.5 Probability plot of the Daisy Creek copperdata. The data have been subdivided into two lognormal groups.

Page 181: Introduction to mineral exploration

164 C.J. MOON

95 are recommended for elements that areenriched (see Fig. 9.8).

Geologists are used to thinking in termsof maps and the most useful end product tocompare geochemical data with geology andgeophysics is to summarize the data in mapform using a gridding or GIS package suchas Geosoft Oasis Montaj, ArcView, or MapInfo(see section 9.2). However, some care shouldbe taken with the preparation of these maps asit is extremely easy to prejudice interpretationand many maps in the literature do not reflectthe true meaning of the data but merely theeasiest way of representing it. If the data reflectthe chemistry of an area, such as a catchment,it is best if the whole area is shaded with anappropriate color or tone, rather than contour-ing (e.g. Figs 8.8, 9.8). Data which essentiallyrepresent a point can be plotted by posting thevalue at that point or if there are a lot of data

Lead

(pp

m)

1200

1000

800

600

400

200

0

1400

0 200 400 600 800 1000

Copper (ppm)

0

0.5

1

1.5

2

Silv

er (

ppm

)(a)

(b)

FIG. 8.6 Scatter plots of Daisy Creek data showing(a) the strong correlation of copper and silver and(b) the lack of correlation of lead and copper.

++++++++++++++++++++++++++++++++++++++++

++++++++++++++++++++++++++++++++

+++++++++++++++++++

++++++++++++++++++++++++++

++++++++++++++++++++++

+++++++++++++++++++++++++++

+++++++++++++++

++++++++++++++++++++++++++++++++++++++++++++++++

0 50 m

N

CuPbAg

FIG. 8.7 Contour plot of Daisy Creek data showingthe highest anomalous populations for copper(>128 ppm), lead (>69 ppm), and silver (>0.55 ppm).The relative size and relative positions of theanomalous areas reflect the primary zoning in theunderlying prospect. (From Sinclair 1991.)

by representing them by symbols, which areeasier to interpret. Typical symbols are givenby Howarth (1982). A more usual method ofpresentation if the samples were collected ona grid is to present the data as a contour plotusing the intervals from a histogram (Figs 8.7,8.4). Plots are best colored from blue to greento yellow to red – cold to hot, as an indicationof low to high concentrations.

The combined statistical analysis and ele-ment mapping is used to define areas of inter-est which need to be investigated in detail,using detailed geology, topography samplinginformation, and, probably, a return to the site.These areas can be classified as to their sug-gested origin, seepage, and precipitation fromlocal groundwaters, from transported material,or probable residual anomalies. The aim mustbe to provide a rational explanation for thechemistry of all the areas of interest and notmerely to say that they are of unknown origin.

8.4 RECONNAISSANCE TECHNIQUES

The application of these techniques has beendiscussed in section 4.2.2 but the actual

Page 182: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 165

TABLE 8.4 Application of the main exploration geochemical methods.

Planning Drainage Lake Water Soil Deep Rock Vegetation Gasscale sediments sediments overburden

Reconnaissance 1:10,000 to Common Glacial Arid Common Laterite Glacial Good outcrop Forests1:100,000

Detailed 1:2500 to Common Common Forests Arid1:10,000

Drilling >1:1000 Common

Common, commonly used. Other techniques useful in areas noted.

method to be used depends on the origin of theoverburden. If it is residual, i.e. derived fromthe rock underneath, the problem is relativelystraightforward but if the overburden is exotic,i.e. transported, then different methods must beused. A summary of the application of variousgeochemical techniques is given in Table 8.4.

8.4.1 Stream sediment sampling

The most widely used reconnaissancetechnique in residual areas undergoing activeweathering is stream sediment sampling, theobject of which is to obtain a sample that isrepresentative of the catchment area of thestream sampled. The active sediment in thebed of a river forms as a result of the passage ofelements in solution and in particulate formpast the sampling point. Thus a sample can beregarded as an integral of the element fluxes.The simplicity of the method allows the rapidevaluation of areas at relatively low cost. Inter-pretation of stream sediment data is carried outby comparing the elemental concentrations ofcatchments, as there is only a poorly defined re-lationship between the chemistry of the streamsediment and the chemistry of the catchmentfrom which it is derived. The simple mass bal-ance method put forward by Hawkes (1976),that relates the chemistry of anomalous soilsto that of stream sediments, works in tropicalareas of intense weathering but requires modi-fication by topographical and elemental factorsin other areas (Solovov 1987, Moon 1999).

In stream sediment sampling the wholestream sediment or a particular grain size ormineralogical fraction of the sediment, such asa heavy mineral concentrate, can be collected.In temperate terrains maximum anomaly/

background contrast for trace metals isobtained in the fine fraction of the sediments asthis contains the majority of the organic mater-ial, clays, and iron and manganese oxides. Thecoarser fractions including pebbles are gener-ally of more local origin and depleted in traceelements. Usually the size cut-off is taken at80 mesh (<177 µm). However, the grain size giv-ing best contrast should be determined by anorientation survey. For base metal analysis andgeochemical mapping a 0.5 kg sample is suf-ficient but a much larger sample is requiredfor gold analysis due to the very erratic distri-bution of gold particles. A number of authors(Gunn 1989, Hawkins 1991, Akçay et al. 1996)have collected 8–10 kg of −2 mm material andcarefully subsampled this using methods suchas automated splitting (section 10.1.4).

The most common sample collectionmethod is to take a grab sample of activestream sediment at the chosen location. Thisis best achieved by taking a number of sub-samples over 20–30 m along the stream and atdepths of 10–15 cm to avoid excessive iron andmanganese oxides. Care must be taken, if a finefraction is required, that sufficient sample iscollected. In addition contamination shouldbe avoided by sampling upstream from roads,farms, factories, and galvanized fences. Anyold mine workings or adits should be carefullynoted as the signature from these will masknatural anomalies. Most surveys use quite adense sample spacing such as 1 km−2 but thisshould be determined by an orientation survey.The main mistake to avoid is sampling largerivers in which the signature from a depositis diluted.

An alternative approach used by geologicalsurveys, for example the British Geological

Page 183: Introduction to mineral exploration

166 C.J. MOON

>163

73 – 163

64 – 73

40 – 64

21 – 40

<21

Geological boundarySample site•

0 1 km

A

B

Copper ppm C

Sea

N

FIG. 8.8 Stream sediment survey of part of the Gairloch area, Wester Ross, NW Scotland. The survey,which was designed to follow up known mineralisation (A), detected further anomalies along strike (B) anddisseminated mineralisation at C. The analytical method was ICP–ES following a HNO3–HClO4 digestion.

massive sulfide mineralisation and the copperdata is shown in the figure. The area sampledconsists of interbedded units of amphiboliteand mica schist, within which a deposit hadbeen drilled at point A. The area suffers veryhigh rainfall and much precipitation of iron andmanganese; as a result dispersion trains arevery short, of the order of 200 m, as shown byan orientation survey. Thus the regional surveywas conducted at very close-spaced intervals.However, the deposit is clearly detected andother prospects were defined along strikefrom the deposit (point B) and disseminatedchalcopyrite was found at C.

Survey, is to sieve the samples in the fieldusing a minimum of water. The advantagesof this are that it is certain that enough finefraction is collected and samples are easier tocarry in remote areas. It also seems that thesamples are more representative but the costefficiency of this in small surveys remains tobe demonstrated. A full review of the use ofregional geochemical surveys is given in Haleand Plant (1994) and useful discussion inFletcher (1997).

An example of a stream sediment survey isshown in Fig. 8.8. The object was to define fur-ther mineralisation in an area of Besshi style

Page 184: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 167

Heavy mineral concentrates have beenwidely used as a means of combating excessivedilution and of enhancing weak signals. Themethod is essentially a quantification of thegold panning method, which separates grainson the basis of density differences. Panningin water usually separates discrete mineralswith a density of greater than 3. Besides pre-cious metals, panning will detect gossanousfragments enriched in metals, secondary oreminerals such as anglesite, and insolubleminerals such as cassiterite, zircon, cinnabar,baryte, and most gemstones, including dia-mond. The mobility of each heavy mineral willdepend on its stability in water, for examplesulfides can only be panned close to theirsource in temperate environments whereasdiamonds will survive transport for thousandsof kilometers.

The samples collected are usually analyzedor the number of heavy mineral grains arecounted. The examination of the concentratecan prove very useful in remote areas wherelaboratory turnaround is slow and the cost ofrevisiting the area high, as it will be possibleto locate areas for immediate follow-up. Themajor problem is that panning is still some-thing of an art and must be practised for a fewdays before the sampler is proficient. The dia-gram and photograph in Figs 8.9 and 8.10 showpanning procedures. Useful checks on panningtechnique can be made by trial runs in areasof known gold or by adding a known numberof lead shot to the pan and checking their re-covery. It is usual to start from a known samplesize and finish with a concentrate of knownmass. However, the differences in panningtechnique mean that comparisons betweendifferent surveys are usually impossible.

8.4.2 Lake sediments

In the glaciated areas of northern Canada andScandinavia access to rivers is difficult on footbut the numerous small lakes provide an idealreconnaissance sampling medium as they areaccessible from the air. A sample is taken bydropping a heavy sampler into the lake sedi-ment and retrieving it. The sampling density ishighly varied and similar to stream sediments.Productivity is of the order of ten samples perflying hour. Essentially data are interpreted in a

similar way to that from stream sediments andcorrections should be made for organic matter.Lake sediment is only useful if glacial materialis locally derived and is ineffective in areasof glaciolacustrine material such as parts ofManitoba. Further details are contained in thereviews by Coker et al. (1979) and Davenportet al. (1997).

8.4.3 Overburden geochemistry

In areas of residual overburden, overburdensampling is generally employed as a follow-upto stream sediment surveys but may be usedas a primary survey for small licence blocksor more particularly in areas of exotic over-burden. Perhaps the key feature is that it isonly employed when land has been acquired.The results are generally plotted at scales from1:10,000 to 1:1000.

The method of sampling depends on thenature of the overburden; if the chemistry ofthe near-surface soils reflects that at depth thenit is safe to use the cheap option of samplingsoil. If not, then samples of deep overburdenmust be taken. The main areas where surfacesoils do not reflect the chemistry at depth areglaciated areas, where the overburden has beentransported from another area, in areas of wind-blown sand, and in areas of lateritic weatheringwhere most trace metals have been removedfrom the near-surface layers.

Surface soil sampling

The simplest sampling scheme is to take near-surface soil samples. The major problem is todecide which layer of the soil to sample, asthe differences between the layers are oftengreater than that between sites. The type ofsoil reflects the surface processes but in generalthe most effective samples are from a zone ataround 30 cm depth formed by the downwardmovement of clays, organic material, and ironoxides, the B horizon indicated in Fig. 8.11, orthe near-surface organic material (A horizon).This downward movement of material is re-sponsible for the depletion and concentrationof trace elements causing variations that maybe greater than that over mineralisation. It isessential that the characteristics of the soilsampled are recorded and that an attempt is

Page 185: Introduction to mineral exploration

168 C.J. MOON

Step 5 Step 6

Step 4b

1

3

24

56

Step 3 Step 4a

Step 1 Step 2

FIG. 8.9 The stages of panning a concentrate from sediment. Often the sediment is sieved to ≤2 mm althoughthis may eliminate any large nugget. Note the importance of mixing the sediment well so that the heavyminerals stay at the bottom of the pan. For an experienced panner the operation will take about 20 minutes.(Modified from Goldspear 1987.)

Page 186: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 169

FIG. 8.10 Panned concentratesampling in northern Pakistan. Thewater is glacial so edge of riversampling is essential.

made to sample a consistent horizon. If differ-ent horizons are sampled anomalies will reflectthis as shown in Fig. 8.12.

The usual method of soil collection in tem-perate terrains is to use a soil hand auger, asshown in Fig. 8.13. This allows sampling to adepth of the order of 1 m, although normallysamples are taken from around 30 cm andmasses of around 100–200 g collected for basemetal exploration. In other climatic terrains,particularly where the surface is hard or wherelarge samples (500 g to 2 kg) are required forgold analysis, then small pits can be dug. Thearea of influence of a soil sample is relativelysmall and should be determined during anorientation survey. The spacing is dependenton the size of the primary halo expected to

occur across the target and the type of over-burden, but a rule of thumb is to have at leasttwo anomalous samples per line if a targetis cut. Spacing in the search for veins may beas little as 5 m between samples but 300 mbetween lines, but for more regularly dis-tributed disseminated deposits may be as muchas 100 m by 100 m. Sample spacing may also bedictated by topography: in flat areas or wherethe topography is subdued then rectilineargrids are the ideal choice, but in mountainousareas ridge and spur sampling may be the onlyreasonable choice.

A typical example of overburden samplingis shown in Figs 8.14a–c. The area is on theeastern side of the Leinster Granite in Irelandand the target was tin–tungsten mineralisation

Page 187: Introduction to mineral exploration

170 C.J. MOON

FIG. 8.13 Typical surface soilsampling in temperate terrainsusing hand augers over theCoed-y-Brenin copper prospect,Wales.

FIG. 8.11 Soil profile showing the effect of sampletexture and soil horizons on copper content. Notethe bedrock copper content of 0.15% Cu. Lettersrefer to soil horizons and subdivisions, from the topL, A, B, C, and R (rock). (From Hoffman 1987.)

FIG. 8.12 Effect of sampling differing soil horizonswith very different metals contents during a soiltraverse.

10 20 50 100 200 500 1000 5 6 7 8

0.0

0.3

0.6

0.9

OrganicSand/silt/clay

Sand

Sand/silt

Cu (ppm) pH

Dep

th (

m)

LH

AE

BF

BM

BM

C1

C2

C1

R

R

R

Sand/silt/clay

Sand/silt/clay

Sand/silt/clay

Sand

9

m

ppm

a

bc

Page 188: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 171

+

+

+

+

+

++

+++

++

+

++

++

++

++

+

++

++ +

++ +

++ +

++ +

+

+

++

+

+ ++

++

+

++ +

+

+

+ +

++

+

++

+

+

+

+

+

++

+

+

+

+

N

Granite

Palaeozoic sediments

Soil traverse

Height above sea level (m)

+ + Microgranite

Microgranite (inferred)

(a)

0 2 km

Ballinglen

Tinahely

Aughrim

1100

Deep overburden tungstenanomalies (>20 ppm W)

River

0

300

0

183

305

183

183

183

183305

183

FIG. 8.14 (a) Regional geochemistry of the Ballinglen tungsten prospects, Co. Wicklow, Ireland. Detaileddispersion through overburden and soil response are shown in b and c. (From Steiger & Bowden 1982.)

associated with microgranite dykes, which arealso significantly enriched in arsenic and lessercopper and bismuth. Regional base of over-burden sampling (usually 2–5 m depth), whichwas undertaken as the area has been glaciated,clearly delineates the mineralized dykes (Fig.8.14a). Surface soil sampling was also effective

in delineating the dykes using arsenic and tosome extent copper and Fig. 8.14b demon-strates the movement of elements through theoverburden. Tungsten is very immobile andmerely moves down slope under the influenceof gravity whereas copper and arsenic are moremobile and move further down slope. The

Page 189: Introduction to mineral exploration

172 C.J. MOON

0–1010– 2020– 3030– 4040– 5050– 6060–70

0–5050–100

100–150150– 200200– 250250– 300300–350

0–55–10

10–1515–2525– 3535– 4545– 55

0

1

2

3

4

Dep

th (

m)

0

1

2

3

4

Dep

th (

m)

300 400 500 600 700 800

0

1

2

3

4

Dep

th (

m)

(b)

Copper ppm

Arsenic ppm

Tungsten ppm

FIG. 8.14 (b) Dispersion through overburden along the soil traverse shown in Fig. 8.14(a). Note the differentmobilities of tungsten, copper, and arsenic.

of their nickel contents, as nickel is mobile,but possible if their multi-element signaturesare used (e.g. Ni, Cr, Cu, Zn, Mo, Mn) or the Pt,Pd, Ir content (Travis et al. 1976, Moeskops1977, Smith 1977).

Drilling to fresh rock, often 50–60 m, hasbeen widely used to obtain a reasonable signa-ture despite the costs involved. Mazzucchelli(1989) suggests that costs are of the order of$A10 m−1 and could easily reach $A1M km−2

for detailed grids. He recommends surfacesampling to outline areas for follow-up work.

Transported overburden

In areas of transported overburden, such asglaciated terrains of the Canadian Shield, sandydeserts, or gravel-covered areas in the Andes,sampling problems are severe and solutions tothem expensive.

density of soil samples on the surface traverseis clearly greater than needed. The reader maycare to consider the maximum sample spacingnecessary to detect two anomalous samples inthe traverse.

A special form of residual overburden isfound in lateritic areas such as in AmazonianBrazil or Western Australia where deep weath-ering has removed trace elements from thenear-surface environment and geochemicalsignatures are often very weak. One techniquethat has been widely used is the examination ofgossans. As discussed in section 5.1.6, gossanshave relic textures which allow the geologistto predict the primary sulfide textures presentat depth. In addition the trace element signa-ture of the primary mineralisation is preservedby immobile elements. For example, it is diffi-cult to discriminate gossans overlying nickeldeposits from other iron-rich rocks on the basis

Page 190: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 173

Lightweight drills are cheaper and easier tooperate but results are often ambiguous, as it isnot easy to differentiate the base of overburdenfrom striking a boulder. The use of heavyequipment in most glaciated areas is restrictedto the winter when the ground is frozen. Thedifficulties of finding the source of glacial dis-persion trains for small targets at Lac de Gras,Canada, are discussed in section 17.2.

In sandy deserts water is scarce, most move-ment is mechanical, and most fine materialis windblown. Thus the −80 mesh fraction ofoverburden is enriched in windblown materialand is of little use. In such areas either a coarsefraction (e.g. 2–6 mm), reflecting locally derivedmaterial, or the clay fraction reflecting ele-ments moved in solution is used (Carver et al.1987). One of the most successful uses ofthis approach has been in the exploration forkimberlites in central Botswana. Figure 8.16shows the result of a regional sampling pro-gram which discovered kimberlite pipes inthe Jwaneng area. Samples were taken on a0.5 km grid, heavy minerals separated fromthe +0.42 mm fraction and the number ofkimberlite indicator minerals, such as picro-ilmenite, counted. The anomalies shown aredisplaced from the suboutcrop probably dueto transport by the prevailing northeasterlywinds.

Recent studies in Chile and Canada haveexamined the use of various weak extractionsto maximize the signature of deposits coveredby gravel and till. Figure 8.17 shows the resultof a survey over the Gaby Sur porphyry cop-per deposit. The weak extractions, includingdeionized water, show high contrast anomaliesat the edge of the deposit, probably generatedby the pumping of groundwater leaching thedeposit (Cameron et al. 2004).

8.4.4 Hydrogeochemistry

Hydrogeochemistry uses water as a samplingmedium. Although water is the most widelyavailable material for geochemistry, its useis restricted to very specific circumstances.The reasons for this are that not all elementsshow equal dissolution rates, indeed many areinsoluble, contents of trace elements are verylow and have been difficult to measure until re-cently, being highly dependent on the weather

As

ppm

400

300

200

100

0

Bi p

pm

1.25

Cu

ppm

10009008007006005004003002001000

Distance (m)1100

(c)

1.00

0.75

0.50

0.25

0.00

50

25

0

FIG. 8.14 (c) Soil traverse across the mineralizeddyke. Note the sharp cut-off upslope and dispersiondownslope.

In glaciated terrains overburden rarelyreflects the underlying bedrock and seepages ofelements are only present where the over-burden is less than about 5 m thick. In addi-tion the overburden can be stratified withmaterial of differing origins at different depths.If the mineralisation is distinctive, it is oftenpossible to use boulder tracing to follow theboulders back to the apex of the boulder fan,as in the case of boulders with visible gold,sulfides, or radioactive material. Generallyhowever the chemistry of the tills must beexamined and basal tills which are usually oflocal origin sampled. Figure 8.15 shows a typ-ical glacial fan in Nova Scotia. Usually basaltill sampling can only be accomplished bydrilling; the most common methods are lightpercussion drills with flow-through samplersor heavier reverse circulation or sonic drills.

Page 191: Introduction to mineral exploration

174 C.J. MOON

l

l

l

ll

l

l

ll

lll

ll

l

ll

l

ll

l

l

lll

l

l

ll

lll

llll

ll

ll

l

l

l

ll

ll

lll

l

l

l

l

ll

ll

l

l

ll

l

l

l

l

lll

lll

ll

llll

l l

ll

l

ll

llll

l

ll

l

l

l

lllllll

l

llll

l

l

ll

l

l

ll

ll

ll

l

l

l

ll

ll

l

ll

l

l

l

ll

ll

ll

l

l

l

l

ll

lllll

ll

l

l

ll

l

lll

lll

ll

ll

l

llll

ll

ll

l

ll

ll

lll

ll

l

l

l

lll

ll

l

l

l

llll

lllll

l

lllll

ll

lll

ll

ll

lll

llll

lllllll

lllll l l

llll

llll

l l

l

ll

ll

ll

lll

ll

ll

lll

ll

lll

l

l

ll

lll

l

l

llll

l

l

ll

l

l

ll

lll

l

ll

ll

l

ll

l

l

ll

l

l

ll

l

ll

l

l

ll

ll

l

l l

lll

ll

ll

l

ll

l

ll

lll

ll

l

l

l

ll

l

ll

lll

l

l

ll

l

ll

l

l

ll

l

l

ll

lll

l

ll

l

l

ll

l

l

l

l

l l

l

l

l

l

l

ll

ll

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

l

l

l

l

ll

ll

ll

l

l

l

l

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l l

l

l

l

ll

l

l

lll ll

l

ll

lll

ll

l ll

ll

l

ll

ll

l

lll

lllllll

lll

l l l l

ll l

l

l

lll

l l

l

ll

ll

l

l

ll

l

l

l

l

l

l

l

l

ll

lll

l

l

ll

ll

l

l

l

l

l

ll

l

l

ll

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l l

l

l

l

ll

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

Sample locations

Grandiorite

Meguma Group

Sn (ppm)

LakeGeorge

JordanLake

GreatPubnico

Lake

Shelburne

YAR

MO

UTH

CO

.

SH

ELB

UR

NE

CO

.

DIGBY C

OUNTY

YARMOUTH COUNTY

SHELBURNE CO.

QUEENS CO.

100–299

300–599

600–899

900–1199

1200–1499

1500+

East Kemptville Tin Mine

l

M

M

DCg

DCg

DCg

DCg

DCg

DCg

M

MM

M

M

N

ICE

0 10 km

FIG. 8.15 Glacial dispersion train from the East Kemptville tin deposit, Nova Scotia, Canada. Basal tillsamples. (From Rogers et al. 1990.)

Page 192: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 175

> 5> 10> 20> 40

Ilmenite counts

DK2 Kimberlite

N

24° 45' E

24° 30' S

DK6

DK5 DK

3 DK1

DK9DK

4

DK2

DK7

WIN

D

0 10 km

FIG. 8.16 Dispersion of ilmenite around the Jwanengkimberlites, Botswana. Ilmenite is one of the heavyminerals used to locate kimberlites. (From Lock1985.)

and easily contaminated by human activity.In general where surface waters are presentit is far more reliable and easier to take a streamsediment sample. The possible exception iswhen exploring for fluorite as fluoride iseasily measured in the field using a portablesingle ion electrode. Where hydrogeochemistrybecomes useful is in exploration of arid areaswith poor outcrop. In this terrain water wellsare often drilled for irrigation and these wellstap deep aquifers that can be used to explore inthe subsurface. This approach has been used inuranium exploration although it has not beenvery successful as the uranium concentrationis dependent on the age of groundwater, theamount of evaporation, and its source, whichare extremely difficult to determine. For gooddiscussions with case studies see Taufen (1997)and Cameron et al. (2004).

8.4.5 Gases

Gases are potentially an attractive medium tosample as gases can diffuse through thick over-burden; in practice most surveys have met withdiscouraging results, although there has beena recent resurgence in interest due to morerobust sampling methods. A number of gases

have been used of which mercury has been themost successful. Mercury is the only metallicelement which forms a vapor at room tempera-ture and it is widely present in sulfide deposits,particularly volcanic-associated base metaldeposits. The gas radon is generated during thedecay of uranium and has been widely usedwith some success. More recently the enrich-ment of carbon dioxide and depletion of oxygencaused by weathering of sulfide deposits hasbeen tested, particularly in the western USA(Lovell & Reid 1989).

In general results have been disappointingbecause of the large variations in gas concen-tration (partial pressure) caused by changes inenvironmental conditions, particularly changesin barometric pressure and rainfall. The moresuccessful radon and mercury surveys havesought to overcome these variations by inte-grating measurements over weeks or months.Gaseous methods work best in arid areas asdiffusion in temperate climates tends to beovershadowed by movement in groundwater.The need to “see through” thick overburdencoupled with improvements in analytical andsampling methodologies has prompted a re-assessment of gas geochemistry and the tech-nique is likely to be used more in the future.

8.4.6 Vegetation

Vegetation is used in two ways in explorationgeochemistry. Firstly the presence, absence orcondition of a particular plant or species canindicate the presence of mineralisation or a par-ticular rock type, and is known as geobotany.Secondly the elemental content of a particularplant has been measured, this is known asbiogeochemistry. Biogeochemistry has beenused more widely than geobotany and hasfound particular application in the forest re-gions of northern Canada and Siberia wheresurface sampling is difficult, but it should onlybe used with caution. A comprehensive treat-ment of the subject can be found in Brooks(1983) and the biogeochemical part has beenrevised in Dunn (2001).

Geobotany

One of the pleasurable, although all tooinfrequent, parts of exploration is to identify

Page 193: Introduction to mineral exploration

176 C.J. MOON

This condition is particularly amenable toremote sensing from satellites. Most gardenersappreciate that plants have favorite soil condi-tions and it is possible to make an estimateof bedrock based on the distribution of plantspecies. Unfortunately this method tends to berather expensive and is little used.

Biogeochemistry

Plants require most trace elements for theirsurvival and take these through their roots,transpiring any residues from their leaves andconcentrating most trace elements in their

FIG. 8.17 Comparison of different partial extractions on soil samples across the Gaby Sur porphyry copperdeposit, Chile. (From Cameron et al. 2004.)

plants, in particular flowers associated withmineralisation. The most famous of these isthe small mauve copper flower of the ZambianCopper Belt, Becium homblei. In general thisplant requires a soil copper content of 50–1600 ppm Cu to thrive, conditions that arepoisonous to most plants (Reedman 1979). Thistype of plant is known as an indicator plant.Unfortunately recent research has demon-strated that most indicator plants will flourishunder other conditions and are not veryreliable.

A more reliable indicator of metal in soils isthe stunting of plants or yellowing (chlorosis).

GravelOxide

SulfidePorphyry

Granodiorite andmetasedimentary

rocks

0

Cu

hydr

oxyl

amin

e (p

pb)

1000

2000

3000

0

Na

deio

nize

d w

ater

(pp

b)

6000

2000

8000

4000

516,000 517,000 518,000 519,000 520,000 521,000 m

0

Cu

aqua

reg

ia (

ppb)

20,000

40,000

60,000

80,000

100,000

0Cu

amm

oniu

m a

ceta

te (

ppb)

100

300

500

200

400

0

Cu

enzy

me

leac

h (p

pb)

20

60

100

40

80

0

Cu

deio

nize

d w

ater

(pp

b)

40

80

20

60

Page 194: Introduction to mineral exploration

8: EXPLORATION GEOCHEMISTRY 177

recent growth. Unfortunately for the geochem-ist the rate of uptake and concentration of ele-ments is highly dependent on the species andthe season. In general sampling is conductedon one plant species and one part of the plant,usually first and second year leaves or twigs.The variation in concentration with seasonis so severe that sampling must be restrictedin time. In general around 500 g of sample arecollected and ashed prior to analysis.

The main advantage of biogeochemistry isthat tree roots often tap relatively deep watertables which in glacial areas can be belowthe transported material and representativeof bedrock. A similar effect can be obtainedby sampling the near-surface humic horizonswhich largely reflect decayed leaves and twigsfrom the surrounding trees (Curtin et al. 1974).

8.5 FOLLOW-UP SAMPLING

Once an anomaly has been found during re-connaissance sampling and a possible sourceidentified, it is necessary to define that sourceby more detailed sampling, by highlightingareas of elemental enrichment, and eliminat-ing background areas until the anomaly isexplained and a bedrock source, hopefully adrilling target, proven. If the reconnaissancephase was stream or lake sampling, this willprobably involve overburden sampling of thecatchments. In the case of soil or overburdensampling it will mean increasing the density ofsampling until the source of an anomaly isfound, proving the source by deep overburdensampling, and sampling rocks at depth or atsurface when there is outcrop. A typical sourcecan be seen in Fig. 8.14. The regional tungstenanomaly was followed up by deep overburdensampling such as that shown in Fig. 8.14b andthe suboutcrop of the dyke defined. The dip ofthe dyke could be estimated from the occur-rence of disseminated sulfides, which wouldalso respond to induced polarization, and adrilling target outlined.

8.5.1 Rock geochemistry

Rock sampling is included in the techniquesfor follow-up because although it has beenapplied with some success in regional re-

connaissance, it is really in detailed work,where there is good outcrop or where there isdrill core, that this technique becomes mosteffective.

On a regional basis the most successfulapplications have been in the delineation ofmineralized felsic plutons and of exhalativehorizons. Full details are provided in reviewsby Govett (1983, 1989) and Franklin (1997).Plutons mineralized in copper and tungstenare usually enriched in these elements but in-variably show high variability within a pluton.Tin mineralisation is associated with highlyevolved and altered intrusives and these areeasily delineated by plotting on a K–Rb versusRb–Sr and Mg–Li diagram or examing thegeochemistry of minerals, such as micas. It isrecommended that 30–40 samples are takenfrom each pluton to eliminate local effects.Studies in the modern oceans indicate haloes ofthe order of kilometers around black smokerfields and these have been detected around anumber of terrestrial massive sulfide deposits,notably by the manganese content in theexhalative horizon.

Some care needs to be taken in the collec-tion of rock samples. In general 1 kg samplesare sufficient for base metal exploration butprecise precious metal determination requireslarger samples, perhaps as large as 5 kg, ifthe gold is present as discrete grains. Surfaceweathering products should be removed witha steel brush. Rock geochemistry depends onmulti-element interpretation and computer-based interpretation with careful subdivisionof samples on the basis of lithology.

Mine scale uses

One of the major applications of rock geo-chemistry is in determining the sense of topand bottom of prospects and detecting altera-tion that is not obvious from visual exam-ination. Geochemists from the former SovietUnion have been particularly active in develop-ing an overall zonation sequence which theyhave used to construct multiplicative indica-tions of younging for a variety of deposit types(Govett 1983, 1989). Rock geochemistry hasbeen widely applied in the search for volcanic-associated massive sulfide deposits in Canada,especially in combination with downhole

Page 195: Introduction to mineral exploration

178 C.J. MOON

electromagnetic methods (see section 15.3.8).In particular rock geochemistry has detectedthe cone-shaped alteration zones beneaththem, which are enriched in Mg and depletedin Na, Ca and sometimes in K, and recognizeddistal exhalite horizons which are enriched inmore mobile elements such as arsenic, anti-mony, and barium (for a good case history seeSeverin et al. 1989). Another major applicationis in the exploration of sulfide-rich epithermalgold deposits, which often show elementzonation (Heitt et al. 2003). The more volatileelements such as As, Sb and Hg are concen-trated above the major gold zones whereas anycopper is at depth. The successful applicationof these methods on a prospect demands adetailed knowledge of the subsurface geology.

8.6 SUMMARY

Exploration geologists are more likely to bedirectly involved in geochemical surveys thanin geophysical ones and they should thereforehave considerable knowledge of this techniqueand its application in mineral exploration. Ageochemical program naturally commenceswith a planning stage which involves choiceof the appropriate field survey and analyticalmethods suitable for use for a particular areaand the commodity sought. Individual areaswill present very different problems, e.g. areaswith deep or shallow overburdens, and there-fore orientation surveys (section 8.1.1) are verydesirable. If possible such a survey should beone in which a variety of sampling methods aretested over a deposit of similar geology to thatof the target and in similar topographical con-ditions to determine which produces the bestresults. This done, the logistics of the majorsurvey can be planned.

The analysis of samples collected duringgeochemical surveys is generally carried out bycontract companies and the quality of the datathey produce must be monitored, and for thisreason the geologist in charge must be familiarwith sample collection and preparation, ana-lytical methods, and the statistical interpreta-tion of the data obtained.

The material which will be sampled dependsupon whether the overburden is residual ortransported. The most widely used technique

in residual areas suffering active weathering isstream sediment sampling which is a relativelysimple method and allows the rapid evaluationof large areas at low cost. Panning to collectheavy mineral concentrates can be carried outat the same time, thus enhancing the valueof this method. In glaciated areas of difficultaccess lake sediment rather than stream sedi-ment surveys may be used.

Overburden geochemistry is often a follow-up to stream sediment surveys in areas of re-sidual and transported overburden. The methodof sampling is very varied and dictated by thenature of the overburden. When the overbur-den is transported (glaciated terrains or sandydeserts) sampling problems are severe and solu-tions to them expensive.

The use of hydrogeochemistry, gases, andvegetation is restricted but rock geochemistryis increasing in importance both in surface andunderground work as well as by utilizing drillcore samples.

8.7 FURTHER READING

The literature on exploration geochemistryis large and reasonably accessible. Geoche-mistry in Mineral Exploration by Rose et al.(1979) remains the best starting point. A goodgeneral update on soil sampling is providedby Fletcher et al. (1987) with reviews ofmost other geochemical techniques in Gubins(1997). The Handbook of Exploration Geo-chemistry series, although expensive, providescomprehensive coverage in the volumes so farpublished: chemical analysis (Fletcher 1981),statistics and data analysis (Howarth 1982),rock geochemistry (Govett 1983), lateritic areas(Butt & Zeegers 1992), arctic areas (Kauranneet al. 1992), and stream sediments (Hale &Plant 1994). Readers are strongly advised to trythe practical problems in Levinson et al. (1987).

Discussion of techniques above can be foundin Explore, the newsletter of the Associationof Exploration (now Applied) Geochemistswhich also publishes the more formal Geo-chemistry: Exploration, Environment, Analy-sis. The reader will also find articles of interestin Journal of Geochemical Exploration,Applied Geochemistry, and Exploration andMining Geology.

Page 196: Introduction to mineral exploration

9Mineral Exploration Data

Charles J. Moon and Michael K.G. Whateley

One of the major developments in mineralexploration has been the increased use of com-puterized data management. This has been usedto handle the flow of the large amounts of datagenerated by modern instrumentation as wellas to speed up and improve decision making.This chapter details some of the techniquesused to integrate data sets and to visualize thisintegration. Two types of computer packageshave evolved to handle exploration and devel-opment data: (i) Geographical InformationsSystems (GIS) for early stage exploration data,usually generic software developed for othernongeologic applications, discussed in section9.2, and (ii) mining-specific packages designedto enable mine planning and resource calcula-tions, discussed in section 9.3.

What must be emphasized is that the qualityof data is all important. The old adage “rubbishin and rubbish out” unfortunately still applies.It is essential that all data should be carefullychecked before interpretation, and the besttimes to do this are during entry of the data intothe database and when the data are collected.A clear record should also be maintained ofthe origin of the data and when and who editedthe data. These data about data are known asmetadata.

9.1 DATA CAPTURE AND STORAGE

9.1.1 Theory

In order to integrate data, they must be avail-able in an appropriate digital form. If the dataare in paper form they require conversion, a

process known as digitization. Even if dataare derived as output from digital instruments,such as airborne magnetometers or downholeloggers, the data may need conversion to adifferent format.

Computers do not know how to classify geo-logical objects so a format for storing data mustbe defined. This format will be determined bythe type, relationship, attributes, geometry,and quality of data objects (for further detailssee Bonham-Carter 1994). An example ofsimple geological map data based on Fig. 9.1is: (i) type – geological unit; (ii) relationship –contacts; (iii) attributes – age, lithology; (iv)shape – polygon. Two main components can beseparated out for all data types: (i) a spatialcomponent dependent on location (e.g. samplelocation for a point sample with x, y ± z com-ponents); and (ii) an attribute component notdependent on location but linked to the spatialcomponent by a unique identifier (e.g. samplenumber or drillhole number and depth). Figure9.1 shows typical geological objects with differ-ing dimensions: lithological units as polygons,samples as points, faults as lines, structuresas points with orientation, drillholes as points(vertical) or lines (inclined).

9.1.2 Spatial data models

There are two major methods of representingspatial data, raster and vector. In the vectormodel the spatial element of the data is repres-ented by a series of coordinates, whereas in theraster model space is divided into regular pixels,usually square. Each model has advantages anddisadvantages summarized in Table 9.1 but the

Page 197: Introduction to mineral exploration

180 C.J. MOON & M.K.G. WHATELEY

Verticaldrillhole

Inclineddrillhole

Structuralreading(dip)

Fault

Samplelocation

Lithologicalunit(polygon)

FIG. 9.1 Typical geological objectswith differing dimensions.

key factors in deciding on a format are resolu-tion and amount of storage required. Figure 9.2shows a simple geological map in vector andraster format. The raster method is commonlyused for remote sensing and discussed in sec-tion 6.2.2, whereas the vector method is usedfor drillholes and geological mapping. Mostmodern systems allow for integration of thetwo different types as well as conversion fromone model to another, although raster to vectorconversion is much more difficult than thatfrom vector to raster.

In a simple (two-dimensional) vector model,points are represented by x and y coordinates,lines as a series of connected points (known asvertices), and polygons as a series of connectedlines or strings. This simple model for polygonsis known as a spaghetti model and is thatadopted by computer aided drawing (CAD)packages. For more complex querying andmodeling of polygons, the relationship betweenadjoining polygons must be established andthe entire space of the study area subdivided.This is known as the topological model. In this

TABLE 9.1 Typical geological features and their representation and quality.

Relationship Typical attributes Geometry Typical quality

Geological unit Contacts Lithology, age, name Polygon ±25 mFault Contacts Type, name Line ±25 mStructural reading Type, orientations Point ±5 mSample location Number, lithology, assay Point ±5 mInclined drillhole Number, date, driller Line ±0.1 mVertical drillhole Number, date, driller Point ±0.1 mDrillhole sample Depth, lithology, assay Point or line ±0.5 m

(interval)

Page 198: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 181

N

Vector 100 m cells 250 m cells

10km

Mylor beds

Gramscatho group

Granite

Quartz porphyry dykesSea

Legend

FIG. 9.2 Vector and raster representation of lithologies from a geological map. The area in SW Cornwall,England and the original map was compiled at 1:250,000 scale based largely on a map of the British GeologicalSurvey. Note the impact of using different cell sizes in a raster map. The lithology of the rasterized cell istaken from the lithology with the largest area within the cell.

model, polygons are formed by the use ofsoftware as a mesh of lines, often known asarcs, that meet at nodes. Another variation ofthis model often used for height data is that ofthe triangular irregular networks (TIN) and isused to visualize digital elevation surfaces orconstruct digital terrain models (DTM). TheTIN model is similar to the polygons used inore resource and reserve calculation (see sec-tion 10.5).

9.1.3 Storage methods

The simplest solution for data storage is thatof the flat file method in which each point hasassociated x, y (and z) coordinates, as well asattributes. This is the familiar style of anaccounting ledger or table and implementedelectronically in a spreadsheet, such as Lotus1-2-3 and Microsoft Excel. In this format(Table 9.2) attributes are stored in columns orfields, and rows, that are known as tuples. Thefeatures of this type of storage are that: (i) allthe data are represented in the table; (ii) any cellin the table must have a single value, replic-ate samples require additional tuples; (iii) no

TABLE 9.2 Advantages and disadvantages of rasterand vector formats. (After Bernhardsen 1992.)

Raster Vector

Data collection Rapid SlowData volume Large SmallData structure Simple ComplexGraphical treatment Average GoodGeometrical accuracy Low HighArea analysis Good AverageGeneralization Simple Complex

duplicate tuples are allowed; and (iv) tuples canbe rearranged without changing the nature ofthe relation (Bonham-Carter 1994). The flat filemethod is however an inefficient way to storedata as a minor change, for example, a changeto the name of a lithology in Table 9.3 (which ispart of the table associated with Figs 9.4 & 9.5),requires a global search and change of all exam-ples. Data are more efficiently stored and editedin a relational database in which the data arestored as a series of tables linked by uniquekeys, such as sample numbers. The flat file

Page 199: Introduction to mineral exploration

182 C.J. MOON & M.K.G. WHATELEY

TABLE 9.3 Part of the data associated with Fig. 9.3 stored as a flat file. Note that a change (e.g. name ofgeological unit) would require editing of all instances.

Area Perimeter Modgeol_ Unit Lithology Age

12,983 711 2 Basic volcanics basic volcanics Middle Devonian2648,177 8049 3 Upper Devonian slates slates Upper Devonian

14,476 866 4 Basic volcanics basic volcanics Middle Devonian977,190 10,015 5 Sea none None

7062,567 27,324 6 Upper Devonian slates slates Upper Devonian69,818,680 231,598 7 Mid Devonian slates slates Middle Devonian

16,556 598 8 Dolerite dolerite Upper Devonian11,447,985 51,973 9 Basic volcanics basic volcanics Middle Devonian

415,067 5493 10 Mid Devonian slates slates Middle Devonian8805 566 11 Basic volcanics basic volcanics Middle Devonian6650 468 12 Mid Devonian limestone limestone Middle Devonian

860,230 5309 13 Basic volcanics basic volcanics Middle Devonian9946 437 14 Dolerite dolerite Upper Devonian

47,900 958 15 Dolerite dolerite Upper Devonian36,817 980 16 Dolerite dolerite Upper Devonian15,053 492 17 Dolerite dolerite Upper Devonian

297,739 3253 18 Basic volcanics basic volcanics Middle Devonian39,071 972 19 Dolerite dolerite Upper Devonian

156 86 20 Sea none None

format can be converted to a relational databaseby a process known as normalization. Table 9.4shows an example of this process, a unit num-ber (finalgeol) which is a simplified versionof modgeol, lithology number (lith#) and age#have been added as the first step in this process.The next step is (Table 9.5) to break Table 9.3into its components. The first table of Table 9.5link the polygons with a finalgeol. The secondlinks finalgeol and with lith# and age# (Table9.5b). The third part of Table 9.5 links lith#with lithology name (Table 9.5c) and the fourthage# with age (Table 9.5d). This decompositioninto their normalized form allows easy editing,e.g. an error in the formation name can becorrected with a single edit.

The flat file and relational database systemsare mature technologies and used extensivelyfor 2D GIS and geometric models. It is howeverbecoming increasingly difficult for them tomanage 3D information with complex topo-logical relationships. Difficulties arise in con-verting the complex data types used inportraying these objects being represented intorelational tables with links. This problem isovercome in an object-oriented (OO) databaseapproach in which conversion is unnecessary

as the database stores the objects directly, com-plete with topological and other informationlinks. For this reason, OO database systems arebeing increasing used in Internet informationdelivery where complex multimedia objectsneed to be retrieved and displayed rapidly(FracSIS 2002).

9.1.4 Corporate solutions

As large amounts of money are invested in col-lecting the data, it is crucial that the data aresafely archived and made available to thosewho need them as easily as possible. Integrityof data is paramount for any mining or explora-tion company, both from a technical andlegal viewpoint (acQuire 2004). However thisintegrity has often been lacking in the past andmany organizations have had poor systemsgiving rise to inconsistencies, lost data, anderrors. Increasingly, in the wake of incidentssuch as the Bre-X fraud (see section 5.4), bothindustry and government departments requirehigher levels of reporting standards. Relationaldatabases provide the means by which data canbe stored with correct quality control pro-cedures and retrieved in a secure environment.

Page 200: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 183

TA

BLE

9.4

Tab

le w

ith

lith

# an

d ag

e# in

sert

ed t

o al

low

th

e ta

ble

to b

e n

orm

aliz

ed a

s sh

own

in T

able

9.5

.

Are

aPe

rim

eter

Mod

geol

_Fi

nalg

eol

Uni

tLi

th#

Lith

olog

yA

ge#

Age

12,9

8371

12

6B

asic

vol

can

ics

6ba

sic

volc

anic

s2

Mid

dle

Dev

onia

n26

48,1

7780

493

8U

pper

Dev

onia

n s

late

s4

slat

es3

Upp

er D

evon

ian

14,4

7686

64

6B

asic

vol

can

ics

6ba

sic

volc

anic

s2

Mid

dle

Dev

onia

n97

7,19

010

,015

510

Sea

0n

one

0N

one

7062

,567

27,3

246

8U

pper

Dev

onia

n s

late

s4

slat

es3

Upp

er D

evon

ian

69,8

18,6

8023

1,59

87

5M

id D

evon

ian

sla

tes

4sl

ates

2M

iddl

e D

evon

ian

16,5

5659

88

9D

oler

ite

8do

leri

te3

Upp

er D

evon

ian

11,4

47,9

8551

,973

96

Bas

ic v

olca

nic

s6

basi

c vo

lcan

ics

2M

iddl

e D

evon

ian

415,

067

5493

105

Mid

Dev

onia

n s

late

s4

slat

es2

Mid

dle

Dev

onia

n88

0556

611

6B

asic

vol

can

ics

6ba

sic

volc

anic

s2

Mid

dle

Dev

onia

n66

5046

812

7M

id D

evon

ian

lim

esto

ne

5li

mes

ton

e2

Mid

dle

Dev

onia

n86

0,23

053

0913

6B

asic

vol

can

ics

6ba

sic

volc

anic

s2

Mid

dle

Dev

onia

n99

4643

714

9D

oler

ite

8do

leri

te3

Upp

er D

evon

ian

47,9

0095

815

9D

oler

ite

8do

leri

te3

Upp

er D

evon

ian

36,8

1798

016

9D

oler

ite

8do

leri

te3

Upp

er D

evon

ian

15,0

5349

217

9D

oler

ite

8do

leri

te3

Upp

er D

evon

ian

297,

739

3253

186

Bas

ic v

olca

nic

s6

basi

c vo

lcan

ics

2M

iddl

e D

evon

ian

39,0

7197

219

9D

oler

ite

8do

leri

te3

Upp

er D

evon

ian

156

8620

10Se

a0

non

e0

Non

e

Page 201: Introduction to mineral exploration

184 C.J. MOON & M.K.G. WHATELEY

Area Perimeter Modgeol_ Finalgeol

12,983 711 2 62648,177 8049 3 8

14,476 866 4 6977,190 10,015 5 10

7062,567 27,324 6 869,818,680 231,598 7 5

16,556 598 8 911,447,985 51,973 9 6

415,067 5493 10 58805 566 11 66650 468 12 7

860,230 5309 13 69946 437 14 9

47,900 958 15 936,817 980 16 915,053 492 17 9

297,739 3253 18 639,071 972 19 9

156 86 20 10

(a)

Finalgeol Unit Lith# Age#

1 Dartmouth Group 1 12 Meadfoot Group 2 13 Acid volcanics 7 14 Staddon Formation 3 15 Mid Devonian slates 4 26 Basic volcanics 6 27 Mid Devonian limestone 5 28 Upper Devonian slates 4 39 Dolerite 8 3

10 Sea 0 0

(b)

(c)Age# Age

0 None1 Lower Devonian2 Middle Devonian3 Upper Devonian

(d)Lith# Lithology

0 None1 sandstone and shale2 shale and sandstone3 sandstone4 Slate5 limestone6 basic volcanics7 acid volcanics8 Dolerite

TABLE 9.5 Normalized form ofTable 9.4. Note the linkagesbetween tables and how a changein lithology or age can be moreeasily edited.

dif, txt (tab delimited and fixed width formats),as well as numerous proprietary formats.

The strategy for collection and evaluation(checking) of data (Walters 1999) is often a mat-ter of company procedure. Most errors are gross

Many proprietary technical software productsprovide such storage facilities, for exampleacQuire (acQuire 2004) provides such a solu-tion for storage and reporting of data that alsointerfaces with files in text formats such as csv,

Page 202: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 185

mapping. However this situation is changingrapidly with the advent of inexpensive portabledigital assistants using a stylus for input andbetter quality displays.

Digitization

Spatial data are normally digitized either byscanning a map or by using a digitizing table.In the digitizing table method a point or line isentered by tracing the position of a puck overit relative to fine wires within the table. Theposition of the point or line is then convertedby software into the original map coordinatesfrom the position measured on the table. Trac-ing lines is very laborious and maps often haveto be simplified or linework traced to avoidconfusion during the digitizing process. Thescanning method has become much easier withthe advent of inexpensive scanners. In thismethod a georeferenced image is traced on acomputer screen using a mouse or puck. Whenthe map or scan has been digitized all lineworkshould be carefully edited. Generally this edit-ing is laborious and more prone to errors thanthe original digitization.

Attribute data

Attribute data can be captured by typing hand-written data or by scanning data that is alreadytypewritten and relatively clean. The scannedimages are then converted into characters forstorage using optical character recognitionsoftware. This software is however not perfectand the resulting characters should be carefullychecked for errors.

9.1.6 Data sources

Sources of digital data have become widespreadand much less expensive over the last 10 years.In particular there are digital topographical(ESRI 2004, Globe 2004), Landsat (Geocover2004, GLCF 2004) and geological map cover-ages of the world, albeit at small scales (GSC1995, Arcatlas 1996). In North America andAustralia most geological, as well as airbornegeophysical data, are available from the na-tional or provincial geological surveys for thecost of reproduction or freely across theInternet. The reader is advised to make a search

and can be easily filtered out. Each geologistand mining or processing engineer knows whatthe database should contain in terms of ranges,values, and units. It is a simple matter of set-ting up the validation tables to check that thedata conform to the ranges, values, and unitsexpected. A simple example would be ensuringthat the dip of drillholes is between 0 and −90degrees for surface drilling.

Documentation for the database is essentialincluding the source and method of data entryas well as their format. It is essential to main-tain adequate, long-term archiving and access.For contingencies, it should be possible to addfields and columns to the database should it benecessary to add information at a later stage.The database should also be capable of mani-pulating all data and the ability to evaluate andreview the data is essential.

Normally a single, maintained, flexible andsecure storage source is recommended. Thismeans that only the most recent copy of thedatabase is being used for evaluation at themine site with older versions archived regu-larly. Multiple access to the database, evenfrom other sites, can be achieved through webtechnology. The additional advantage of suchweb technology is that different sets of data,each containing information about a particu-lar item such as a borehole, can be availablethrough a single search interface (Whateley2002). The ultimate long-term storage howeverremains paper and a hard copy should beregularly filed of all important data sets, e.g.drillhole logs and assays.

9.1.5 Data capture

The capture of attribute data is usually relat-ively straightforward but spatial data is oftenmore difficult. Most field data are now gener-ated in digital form but any data only availableon paper will require digitization.

Field data capture

Capture of data in the field using computersis becoming routine often using ruggedizedcomputers that can be easily transported. Atpresent this approach is widely used for routinetasks such as sample collection and corelogging but is less well suited to geological

Page 203: Introduction to mineral exploration

186 C.J. MOON & M.K.G. WHATELEY

of the Internet using Google or a similar searchengine to find up to date information.

9.1.7 Coordinate systems and projections

Geologists have in the past generally managedto avoid dealing with different coordinatesystems in any detail, as the areas they weredealing with were small. The advent of GPSand computerized data management haschanged this. The plotting of real world data ona flat surface is known as projection and is theresult of the need to visualize data as a flatsurface when the shape of the earth is bestapproximated by a spheroid, a flattened sphere.For small areas the distortion is not importantbut for larger areas there will be a compromisebetween preserving area and distance relation-ships. For example, the well-known Mercatorprojection emphasizes Europe at the expense ofAfrica. The scale of the data also governs thechoice of projection. For maps of scales largerthan 1:250,000, either a national grid or aUniversal Transverse Mercator (UTM) grid isgenerally used. In the latter projection, theearth is divided into segments of 6 degreeslongitude with a value of 500,000 m E given tothe central meridian of longitude and a north-ing origin of 0 m at the equator, if north of theequator, or large number, often 10,000,000 m,if south of the equator.

There are a variety of different values in usefor the ellipsoid that approximates the shapeof the earth, known as the datum. The mostcommonly used datum for GPS work is WorldGeodetic System (WGS 1984) but the datumused on the map must be carefully checked,as the use of different datums can change coor-dinates by up to 1500 m. The reader is advisedto read about the problems in more detail intexts such as Longley et al. (2001) and Snyder(1987).

9.2 DATA INTEGRATION AND GEOGRAPHICALINFORMATION SYSTEMS

One of the major advances in technology at theearly exploration stage has been the ability tointegrate data easily. This has been driven bythe development of Geographical InformationSystems (GIS). Although GIS is usually defined

to include data storage, its main function is toallow the easy integration of data and output,usually in the form of maps (Longley et al.1999, 2001). Its development has been designedfor a wide range of applications involvingspatial data, including the design of optimumsiting of pizza delivery sites and monitoring thespread of disease, but the generic commercialsystems are applicable (with limitations) tomineral exploration data. The dominant com-mercial systems, at the time of writing, areArcView and ArcGIS (ESRI) with 35% of totalmarket share and they have wide use in geo-logical applications; MapInfo has a muchsmaller overall market share but a wide follow-ing in the geological world, particularly inAustralia as add-on programs have been devel-oped for mineral exploration. Current commer-cial systems allow the display of both vectorand raster data with varying degrees of queryingand modeling facilities. These systems are wellsuited to 2D data but, at the time of writing,only partly usable for 3D drill data includingdrillholes and underground sampling.

In addition to the complex (and expensive)fullblown GIS systems there are a number ofsimpler software packages, such as OasisMontaj (Geosoft 2004), Interdex (Visidata 2004),and Micromine (Micromine 2004), specificallydesigned for mineral exploration. Some havefeatures not easily accessed in GIS packages,such as the gridding facilities for geophysicaldata and integrated data management in OasisMontaj.

9.2.1 Arrangement of data using layers

The basis for the integration of data in a GISis their use as a series of layers. This methodis really an extension of the old light tablemethod in which maps were overlaid and theresult viewed by shining a light through them.Combining layers is much easier then splittingthem apart and a good rule is to build layersfrom the simplest components. For example,rather than having a topographical layer it isbetter to have separate layers for roads, fieldboundaries, buildings, and rivers. In the case ofa conventional geological map separate layerswould be generated for geological units (poly-gons), structural readings (points), and faults(lines).

Page 204: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 187

9.2.2 South Devon example

Perhaps the best way of explaining the methodsof handling data in a GIS is to use a workedexample. This example is typical of early stageexploration and refers to an area in southDevon, England. Prior to the mid 1980s thisarea was not thought prospective for gold andno systematic exploration had been under-taken. The British Geological Survey had longbeen active in the area and had investigated thearea of basic volcanics in the north of Fig. 9.3for its base metal potential. The initial orienta-tion samples derived from acid volcanics(266,000E, 49,000N, British National Grid)returned small grains of gold in a panned con-centrate and it was decide to sample for gold.Gold was also known from a small prospect atHopes Nose, Torquay, approximately 30 km tothe northeast of the discovery site.

The area discussed is on the southern flankof one of the major early Permian granite in-trusives, the Dartmoor Granite. This graniteintrudes an area generally of low grade marginalmarine and fluvial sediments with interbeddedvolcanics of Devonian age that were deformedduring the late Carboniferous. The Devoniansediments (and originally the granite) areunconformably overlain by red-bed sandstonesand fine sediments of Permo-Triassic age.Marine sediments of Jurassic and Cretaceousage probably originally covered the area buthave been removed by Tertiary erosion.

Simple display and querying:geological map

The basis for geological interpretation in thearea is a digitized version of the 1:50,000 scale

280,000275,000270,000265,000260,000255,000250,000

N5

km

55,000

50,000

45,000

Legend

Prospect Basic volcanics

Mid Devonian slates

Staddon Fm.

Acid volcanics

Meadfoot Gp.

Dartmouth Gp.

River

Sea

Dolerite

Upper Devonian slates

Mid Devonian limestone

FIG. 9.3 Geological map (in ArcGIS) and index map. This map shows lithologies and no structure. (Largelyafter Ussher 1912.)

Page 205: Introduction to mineral exploration

188 C.J. MOON & M.K.G. WHATELEY

55,000

50,000

45,000

280,000250,000 255,000 260,000 265,000 270,000

Prospectkm

10

Legend

River

Volcanics

275,000

FIG. 9.5 Geological map with only acid and basic volcanics selected. Compare this map with Fig. 9.3.

lithological number and lithological name inthe key (Fig. 9.4, Tables 9.3 & 9.4). In contrastto conventional maps it is easy to select indi-vidual units; in this case (Fig. 9.5) the acid andbasic volcanics have been chosen.

published geological mapping (Ussher 1912).The most important component is the distribu-tion of lithologies that were outlined by gener-ating polygons. Each polygon has a uniquenumber which is linked in a table to the

FIG. 9.4 Geological map of Fig. 9.3 showing polygon numbers.

55,000

50,000

45,000

280,000250,000 255,000 260,000 265,000 270,000 275,000

N

5km

Page 206: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 189

55,000

50,000

45,000

250,000 255,000 260,000 265,000 270,000 275,000 280,000

Prospect10

km

FIG. 9.6 Digital elevation model of the area based on the Ordnance Survey of Great Britain digital datacollected at 50 m intervals. The image is hillshaded with sun elevation of 45 degrees from 315 degrees soSW–NE striking structures are highlighted. (Source: Digimap, University of Edinburgh.)

Remote sensing data

As the geological mapping was quite old(Ussher 1912) and the oucrop poor, it waspossible to check the mapping with Landsat,air photography, and digital elevation models.Figure 9.6 shows a hill shaded image of thedigital elevation model. A number of featurescan be detected by comparison with Fig. 9.3,particularly the high ground underlain by theStaddon Formation, a coarse sandstone unit.Some mapped structures are obvious, a mappedfault at 266,000E, 51,000N and the strongN–S lineament at ~274,000E. Landsat images(Fig. 9.7) are difficult to interpret as the area isintensely farmed in small fields. Vegetationis dominant and many of the dark gray areasof Fig. 9.7 are woodland. In spite of this itwas possible to extract lineaments, which areshown in Fig. 9.10.

Geochemical data

One of the key data sets in the area is the resultof a panned concentrate sampling exercise(Leake et al. 1992). Samples were analyzed byX-ray fluorescence for trace elements and forgold by solvent extraction and atomic absorp-

tion spectrometry. Gold geochemistry is themost obvious indicator of area for follow-up.Gold show a strongly skewed distributionwith a mean of 273 ppb, median of 10 ppb, andmaximum of 5700 ppb Au (Fig. 9.8). The soft-ware used offers four options to divide the data:(i) quantiles – e.g. 20, 40, 60, 80, 100 percentilesin this case 5, 5, 24, 200, 5700 ppb; (ii) equalintervals – 1140, 2280, 3420, 4560, 5700 ppb;(iii) natural breaks (based on breaks in thehistogram) 215, 708, 1540, 3100 ppb; (iv) meanand standard deviations – 273, 1060, 1846,2633 ppb. As the distribution is so highlyskewed and 40% of the concentrations are lessthan the detection limit 10 ppb (set to 5 ppb),in this case a percentile method based on thehigher part of the distribution was used: 50, 75,90, 95 percentiles rounded to familiar numbers,10, 150, 500, 1500 (Fig. 9.8). This map picksout catchments with detectable gold as wellas differentiating those which have very highgold concentrations. These were followed upand gold grains were discovered in soil at thelocations marked by the prospect symbols.

If the log-transformed data are plotted on aprobability scale (Fig. 9.9), a threshold of 50 pbbmight be selected. This simple subdivision isshown in Fig. 9.9, although it is, at least in the

Page 207: Introduction to mineral exploration

190 C.J. MOON & M.K.G. WHATELEY55,000

50,000

45,000

250,000 255,000 260,000 265,000 270,000 275,000 280,000

Prospect

Legend 10km

FIG. 9.7 Landsat image of the area. Black and white version of a color composite of bands 7, 5, and 4. The darkgrey areas are mainly trees in river valleys and the black area is sea.

FIG. 9.8 Panned concentrate gold data in ppb. Intervals are approximately 50, 75, 90, and 95 percentiles,as discussed in the text. (Data courtesy British Geological Survey but converted into their catchments(drainage basins) by C.L. Wang.)

Page 208: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 191

55,000

50,000

45,000

250,000 255,000 260,000 265,000 270,000 275,000 280,000

Prospect

0.5–50

50–5700

RiverAu (ppb)

log10

3.59

2.26

.93

–.4

–1.74

0 10 30 50 70 90 98

–2

–1

0

1

2Legend

Au ppb

percentiles (probability scale)

5kmN

FIG. 9.9 Gold data of Fig. 9.8 replotted using a simple threshold of 50 ppb (log10 = 1.7 of the lower plot).The log10–probability plot of the Au data was made in Interdex software and should be compared with thehistogram of Fig. 9.8.

opinion of the writers, inferior to the multipleclass map of Fig. 9.8.

The panned concentrates were analyzed fora variety of elements in addition to gold andmuch more information can be deduced fromthe other elements. One surprise was the highconcentrations of tin, present as cassiterite,as the area is distant from the well-known tinmineralisation of Dartmoor and Cornwall.These high tin concentrations probably reflectheavy mineral concentration on an erosion sur-face of unknown, although probably Permianand Miocene, ages.

Data overlay and buffering

Overlaying data is one of the key methods inGIS. A typical use of the data overlay function

is to analyze the control of lineaments on thedistribution of the known gold soil anomalies.In the study area two major trends of linea-ments, NW–SE and NE–SW, have been recog-nized. Possible lineaments following theseorientations were derived from an analysis ofa Landsat TM image and digitized into twocoverages (Fig. 9.10). The relation of linea-ments with gold mineralisation can beanalyzed by calculating the distance of the goldoccurrences from the lineaments. In thisexample there does not seem to be a definitiverelationship between gold occurrences and aparticular set of lineaments.

Lineaments are often inaccurately definedand it more practical to define zones of influ-ence. The ability to generate corridors of agiven width around features or buffer around

Page 209: Introduction to mineral exploration

192 C.J. MOON & M.K.G. WHATELEY

55,000

50,000

45,000

N

5km

Legend

Prospect

Mappedfault or thrust

Landsatlineament

Lithologicaloutline

River

250,000 255,000 260,000 265,000 270,000 275,000 280,000

FIG. 9.10 Lineaments digitized from Landsat coverage by C.L. Wang and buffered at 250 m for clarity. Knowngold occurrences and mapped faults are added for comparison.

structures and geological contacts is thereforea major advance. In Fig. 9.10 buffers of 250 mare shown around the NW–SE lineaments.

Boolean methods

Once the layers to be investigated have beenoverlaid they can be used to define areas forfollow-up. The simplest method of selectingareas is to use straightforward Boolean algebra.In this approach data from layers are selectedand combined with other layers depending onthe prejudices of the geologist (Fig. 9.11). Anexample in the south Devon data is to selectcatchments which are underlain by acid vol-canics and have panned concentrate gold>1000 ppb Au based on the premise that theacid volcanics are the most favorable rock forgold deposits. The statements used by thecomputer are “geology = acid volcanics andAu > 1000 ppb”. The database is then scannedand each unit is either true (1) or false (0) forthe statement. If dolerite is considered to beof similar favorability to acid volcanics thenthe statements “(geology = acid volcanics) and(Au > 1000 ppb)” will meet these requirements

(Fig. 9.12). Although the Boolean approach canbe very effective and is simple, it assumes thatthere is a sharp break in attributes, for example,gold in panned concentrates is significant at1001 ppb but not at 999 ppb, and in the simpleuse of the method that each layer is of equalimportance. In the index overlay variety ofthis method, weights are assigned to thedifferent layers, depending on their perceivedimportance.

Complex methods

Although the Boolean approach can be effec-tive, and is easily understood, it is difficult todetect more subtle patterns, to use very largeand complex data sets, or to apply complicatedrules. More complex methods have been widelyapplied to mineral exploration data but arehighly dependent on software. Originally theprograms were expensive and difficult to usebut now many methods are available throughthe use of add-on software to Arcview andMapinfo. Of particular note is the Arc-SDMsoftware developed by the US and Canadianfederal geological surveys and available freely

Page 210: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 193

Au OR Felsite Felsite XOR Au

FelsiteAu

Felsite NOT Au Au AND Felsite

FelsiteAu

FelsiteAu FelsiteAuFIG. 9.11 Boolean algebra. Thecircles are diagrammaticrepresentations of two sets, felsiteand Au. The possible combinationsare shown. XOR is for felsite or Aubut not where they overlap.

55,000

50,000

45,000

250,000 255,000

Legend

260,000

Prospect

Lithologicaloutline

Samplecatchment

Felsite

Au >1000 ppb

10km

Intersectionfelsite andAu >1000 ppb

265,000 270,000 275,000 280,000

N

FIG. 9.12 Simple Boolean search in south Devon data for the presence of acid volcanics (felsite) and gold incatchment >1000 ppb, based on the maps of Figs. 9.3 and 9.8. The selected lithological units and catchmentsare shown by outlining and the very small area which meets both these criteria is in black and arrowed.

Page 211: Introduction to mineral exploration

194 C.J. MOON & M.K.G. WHATELEY

on the Internet (Arc-SDM 2001). The differingcomplex methods can be divided into two mainapproaches: (i) data driven, without a precon-ceived model; (ii) model driven, dependent onthe type of model used. Although the methodsdiffer, their aim is to map the probability offinding mineralisation.

Data-driven approaches

The main data driven techniques are (1)weights of evidence and (2) logistic regression,both of which require the location of mineral-isation in at least part of the area to be known,in order to use as a training set for use in theremainder of the area. The reader can findfurther details in Bonham-Carter (1994) and atthe Arc-SDM website (Arc-SDM 2001).1 Weights of evidence was pioneered in min-eral exploration use by Bonham-Carter et al.(1988). This method uses Bayesian statisticsto compare the distribution of the knownmineralisation with geological patterns thatare suspected to be important controls.

The first stage in using the method is tocalculate the probability of deposit occurrencecalculated without knowing anything of thetest pattern (prior probability). The distributionof mineralisation in the training area is thencompared with geological patterns in turn. Thecalculations for each pattern give two weights:(i) a positive weight indicating how importantthe pattern is for indicating an occurrence; and(ii) a negative value indicating how importantthe absence of the pattern is for indicatingno occurrence. The weights for the differentpatterns are then examined to see if they aresignificant and combined to calculate an over-all posterior probability. This is then used tocalculate the probability of mineralisationbased on the geological patterns.

In the example of the data from south Devon,the summary weights for different geologicalcontrols are shown in Table 9.6. The mostimportant patterns using the known goldoccurrences are the occurrence of felsite anddolerite which can be seen to dominate thefinal probability map (Fig. 9.13).2 Logistic regression uses a generalized regres-sion equation to predict the occurrence of min-eralisation. Logistic regression is distinct fromthe more familiar least squares regression in

that it uses presence or absence of geologicalpatterns, rather than a continuous variable.The significance of the various geological con-trols is tested and a model built up.

In the south Devon example (Fig. 9.14) usingdrainage catchments containing more than1000 ppb Au generated a probability p of:

Logit (p) = −3.81 + 1.47X5 − 1.74X11 +2.04X2X5 + 3.07X3X4

where X3X4 is combined felsite andDartmouth Group occurrence, X5 MeadfootGroup occurrence, X2X5 Meadfoot Group andbasic volcanics, and X11 faults. In this model,the felsite and Dartmouth Group factor is themost important and faults are a less importantcontrol.

Model-driven approaches

A number of model-driven approaches havebeen used which develop on the index overlaytheme: (1) fuzzy logic, (2) Dempster Shafermethods, and (3) neural networks.1 Fuzzy logic. This has probably been themost widely used technique. In the fuzzy logicmethod variables are converted from raw datato probability using a user-defined function.For example, if arsenic soil geochemistry is sus-pected as a pathfinder for gold mineralisation,then low As concentrations (say <100 ppm As)will be given a probability of 0 and high con-centrations (say >250 ppm As) a probability of1. Intermediate concentrations will be given aprobability of between 0 and 1 depending onthe function applied (Fig. 9.15). The variablesare then combined using algebra similar tothe Boolean functions but including sum andproduct functions. The latter two functionsare often combined together using a gamma-function to generate an overall probabilityfrom individual layers (Bonham-Carter 1994,Knox-Robinson 2000). Knox-Robinson (2000)advocates the use of methods that allow thevisualization of the degree of support of geo-logical evidence. An example of the use ofthese techniques on Pb-Zn deposits in WesternAustralia can be found in D’Ercole et al.(2000). In the south Devon example, a fuzzy“or” has been used to combine layers whichwere estimated to indicate the occurrence of

Page 212: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 195TABLE 9.6 Calculated weights and contrasts for patterns in generating a probability map. The mineralisationpattern was known gold occurrences as 1 km2 cells.

Patterns W j+ W j

− C

Landsat lineaments NE–SW 250 m buffer −0.222 0.025 −0.248Landsat lineaments NW–SE 250 m buffer 0.247 −0.039 0.285Landsat lineaments NW–SE 500 m buffer 0.298 −0.119 0.417Landsat lineaments NW–SE 750 m buffer 0.315 −0.245 0.560Landsat lineaments NW–SE 1 km buffer 0.259 −−−−−0.339 0.598Landsat lineaments NW–SE 1.25 km buffer 0.143 −0.276 0.419Landsat lineaments NW–SE 1.5 km buffer 0.052 −0.145 0.197Landsat lineaments NW–SE 1.75 km buffer 0.002 −0.009 0.011Landsat lineaments NW–SE 2 km buffer −0.003 0.023 −0.027Felsite 3.739 −−−−−0.059 3.797Dartmouth Slate 1.050 −−−−−0.618 1.668Anomalous catchments with Au >>>>> 5 ppb 0.844 −−−−−0.689 1.533Contact (Dart – Mead ) 2 km buffer 0.625 −−−−−0.741 1.366Dolerite 1.184 −−−−−0.017 1.201Thrust 250 m buffer 0.784 −−−−−0.062 0.846Contact (Mead – Stad ) 750 m buffer 0.548 −−−−−0.199 0.746Contact (Mead – Mdev ) 750 m buffer 0.605 −−−−−0.084 0.689

For each lineament set a series of buffers were calculated and an example of the results for different buffers of the NW–SElineaments is shown. Otherwise only the final results are shown. The significant features (shown in bold) were used tocalculate the final probability.W j

+, positive weight; W j−, negative weight; C, contrast.

Dart, Dartmouth Slate; Mead, Meadfoot Group; Stad, Staddon Grits; Mdev, Middle Devonian Slates.

FIG. 9.13 Posterior probability map resulting from a weights of evidence analysis of geological andgeochemical data discussed in the text. Note the general good correlation with known occurrences with theexception of that at ~2,72000E, 52,000N, which is probably of different origin. (Source: Wang 1995.)

55,000

50,000

45,000

250,000 255,000

Legend

Probability

260,000

Prospect

N

Coast

5km

0.075–0.125

0.125–0.175

0.175–0.9

No data

0–0.025

0.025–0.075

265,000 270,000 275,000 280,000

Page 213: Introduction to mineral exploration

196 C.J. MOON & M.K.G. WHATELEY

55,000

50,000

45,000

250,000 255,000

Legend Probability

260,000

Prospect

N

Coast

5km

0.03–0.10

0.11–0.40

0.41–0.52

0.00

0.01–0.02

265,000 270,000 275,000 280,000

FIG. 9.14 Probability map based on logistic regression of Au panned concentrate values >1 ppm againstgeology and structure. Note the clustering of high probabilities in the center of the map and the prospectivecatchment at ~27,000E, 45,000N. (Source: Wang 1995.)

these can be difficult to apply but a good ex-ample of an application in Canada is found inMoon (1990).3 Neural networks. This is a method in whichsoftware is trained by using the absence orpresence of mineral deposits in large mixeddata sets to find predictive geological factors.

mineralisation, as the “or” includes most ofthe control broad areas that are defined withinwhich there are higher probability zones(Fig. 9.16).2 Dempster Shafer belief functions. Theseallow the input of some idea of the degree ofconfidence in the evidence used. In practice

As (ppm)0 250

0100

1

Pro

babi

lity

FIG. 9.15 Relationship betweenarsenic geochemistry andprobability of indicatingmineralisation in a hypotheticalexample, in which <100 ppm hasno association with mineralisationand >250 ppm is associated withall gold shows. A straight linefunction is dashed and anotherfunction is shown in dots.

Page 214: Introduction to mineral exploration

9: MINERAL EXPLORATION DATA 197

55,000

50,000

45,000

250,000 255,000

Legend Probability

260,000

Prospect

N

Coast

5km

0.40–0.55

0.55–0.65

0.65–0.75

>0.75

no data

0.25–0.40

< 0.25

265,000 270,000 275,000 280,000

FIG. 9.16 Probability map resulting from a fuzzy logic analysis. This analysis highlights any layer which haspossible association with gold mineralisation and is the result of a fuzzy “or.” This should include alloccurrences. (Source: Wang 1995.)

9.3 INTEGRATION WITH RESOURCECALCULATION AND MINE PLANNINGSOFTWARE

Most resource calculations performed in explo-ration, in feasibility studies, or in routine minegrade control and scheduling use a specializedcomputer package (or packages) that deals with3D data. At the early stages of exploration, thekey features of the package will be to input drillinformation and relate this to surface features,as discussed in Chapter 5 and section 10.4.When a resource is being calculated, the abilityof the package to model the shape of geolo-gical units and calculate volumes and tonnagesbecomes more important. Subsequently infeasibility studies (see section 11.4.4) the capa-city to design underground or surface workingsand schedule production is crucial.

The package chosen will depend on thefinance available and the nature of the opera-tion, as there are packages specifically designedfor both open cut and underground mining. The

Brown et al. (2000) compared the method withfuzzy logic and weights of evidence on datafrom an area in New South Wales and foundthat it gave superior results. They claim thatalthough neural networks require a trainingset the number of occurrences required is notas onerous as that of the weights of evidencemethod.

There have been a number of other studiesof using complex methods in integrating dataand the reader should examine some of theseand decide which approach is suitable. A goodstarting point is the volume of Gubins (1997).The paper by Chinn and Ascough (1997)details the effort of Noranda geologists in theBathurst area of Canada. They used a fuzzylogic approach to define areas for follow-upbut concluded that the main advantage ofcomputer-based methods was to get theirdeposit experts to sit down together and puttheir experience on paper. The methods aretherefore not a substitute for excellent geologybut a complement.

Page 215: Introduction to mineral exploration

198 C.J. MOON & M.K.G. WHATELEY

selection of such technical software is often anemotive issue and often depends upon a per-son’s previous exposure to and their familiaritywith the software, or its impressive graphicscapability. Unfortunately software is rarelyselected for the relevance to the operation ofthe functions that it provides. There are meritsand weaknesses in each package. In addition,improvements keep appearing and a differentdecision may be reached if the selection exer-cise was repeated at a different moment intime.

Most operations have selected geologicaland mine modeling software that suits theirbudget, and geological and mining complexity.Occasionally, a mining house will standardizeon one software package, e.g. BHP Billiton usesVulcan software. The software modeling sys-tems currently available include Surpac andMinex from the Surpac Minex Group, Maptek’sVulcan, Mincom, Datamine, Mintec’sMineSight, and Gemcom. Other operationsmay have task-specific proprietary software,such as Isatis (Geovariances 2001), used for

resource and/or reserve calculations, orWhittle or Earthworks products for strategicmine planning.

9.4 FURTHER READING

This is a rapidly developing area and thereis not a single source covering the field.There are many good texts covering the use ofGIS but probably the best starting point is thegeneral textbook of Longley et al. (2001) whichis a cut-down version of the comprehensiveLongley et al. (1999). Geographic InformationSystems for Geoscientists by Bonham-Carter(1994) remains the only text aimed at geologists,although the field has advanced significantlysince the book was published.

The use of mining-specific software is notwell covered in the formal literature althoughmuch information is available on the Internet.For Spanish speakers the book by Bustillo andLopez (1997) provides a well illustrated guide toexploration geology and mine design.

Page 216: Introduction to mineral exploration

10Evaluation TechniquesMichael K.G. Whateley and Barry C. Scott

As an exploration geologist you will beexpected to be familiar with mineral depositgeology, to understand the implications ofextraction on the hydrology of the mineraldeposit area, to recognize the importance ofcollecting geotechnical data as strata controlproblems may have considerable impact onmine viability, to be able to propose suitablemining methods, and to assess the economicviability of a deposit. To be able to cope with allthese tasks you will need to be technicallycompetent particularly in sample collection,computing, and mineral resource evaluation.The basis of all geological evaluation is thesample. Poor sample collection results in un-reliable evaluation.

This chapter will present some of themethods that are used in the field to obtainrepresentative samples of the mineralized rockthat will enable the geologist to undertakemineral deposit evaluation. This includes thevarious drilling techniques, pitting and trench-ing as well as face and stope sampling. Oncethese data have been collected they are evalu-ated to determine how representative they areof the whole deposit. This is achieved usingstatistics, a subject covered in the first part ofthe chapter.

The geologist prepares and evaluates thedeposit’s geological and assay (grade) data. Inan early phase of a project, global resourceestimates using classical methods will norm-ally be adopted. Later, as more data becomeavailable, local estimates are calculated usinggeostatistical methods (often computer based).An outline of these evaluation methods is alsopresented.

Finally, to ensure that the maximumamount of information is derived from drillcore (other than the obvious grade, thickness,rock type, etc.) a basic geotechnical and hydro-geological outline is given. This informationmay be used in assessing possible dilution ofthe ore upon mining, or may identify stratacontrol or serious water problems which mayoccur during mining.

10.1 SAMPLING

10.1.1 Introduction to statistical concepts

Population and sample

Sampling is a scientific, selective processapplied to a large mass or group (a population,as defined by the investigator) in order toreduce its bulk for interpretation purposes.This is achieved by identifying a componentpart (a sample) which reflects the character-istics of the parent population within accept-able limits of accuracy, precision, and costeffectiveness. In the minerals industry theaverage grade of a tonnage of mineralized rock(the population) is estimated by taking sampleswhich are either a few kilograms or tonnes inweight. These samples are reduced to a fewgrams (the assay portion) which are analyzedfor elements of interest.

Results of analyzed samples plotted as afrequency curve are a pictorial representationof their distribution (Fig. 10.1). Distributionshave characteristics such as mid-points andother measures which indicate the spread

Page 217: Introduction to mineral exploration

200 M.K.G. WHATELEY & B. SCOTT

Fre

quen

cy

Low Variable High

X2

X1

FIG. 10.1 Two normal distributions. Distribution 1has a lower arithmetic mean (X1) than distribution 2but higher variance (i.e. a wider spread). Distribution2 has the reverse, a higher mean (X2) but a lowervariance.

tion. Mineralisation is composed of dissimilarconstituents and is heterogeneous – it is rarely(if ever) homogenous – and correct sampling ofsuch material has to ensure that all constituentunits of the population have a uniform prob-ability of being selected to form the sample,and the integrity of the sample is respected (seelater). This is the concept of random sampling.

Normal and asymmetrical distributions

This is a brief introduction to the subject of dis-tributions. The reader is referred to standardtexts such as Issaks and Srivastava (1989) formore details.

Normal distributionIn a normal distribution the distributioncurve is always symmetrical and bell shaped(Fig. 10.2). By definition, the mean of a normaldistribution is its mid-point and the areasunder the curve on either side of this valueare equal. Another characteristic of this distri-bution, or curve, is the spread or dispersion ofvalues about the mean which is measured bythe variance, or the square root of the variancecalled the standard deviation. Variance (σ 2)is the average squared deviation of all possiblevalues from the population mean:

σ 2

2

( )

=∑ −x

ni X

where σ 2 = population variance, xi = any samplevalue, X = population mean, n = number ofsamples.

of the values and their symmetry. These areparameters if they describe a population andstatistics if they refer to samples.

In any study the investigator wishes to knowthe parameters of the population (i.e. the “true”values) but these cannot be established unlessthe population is taken as the sample. This isnormally not possible as the population is usu-ally several hundred thousands or millions oftonnes of mineralized rock. A best estimate ofthese parameters can be made from samplingthe population and from the statistics of thesesamples. Indeed a population can be regarded asa collection of potential samples, probably sev-eral million or more, waiting to be collected.

It is fundamental in sampling that samplesare representative, at all times, of the popula-tion. If they are not the results are incorrect.The failure of some mineral ventures, andlosses recorded in the trading of mineral com-modities, can be traced to unacceptable sam-pling procedures due to confusion betweentaking samples that are representive of the de-posit being evaluated, and specimens whosedegree of representation is not known.

Homogeneity and heterogeneity

Homogeneity is the property that defines apopulation whose constituent units are strictlyidentical with one another. Heterogeneity isthe reverse condition. Sampling of the formermaterial can be completed by taking any groupof these units such as the most accessible frac-

Fre

quen

cy

Low Variable High

99%

95%

68%

−3 −2 −1 +1 +2 +3X

FIG. 10.2 The normal distribution: the variable has acontinuous and symmetrical distribution about themean. The curve shows areas limited and occupiedby successive standard deviations.

Page 218: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 201

Variance units are the square of the units ofthe original observation. A small variance indi-cates that observations are clustered tightlyabout the arithmetic mean whilst a large vari-ance shows that they are scattered widely aboutthe mean, and that their central clustering isweak. A useful property of a normal distribu-tion is that within any specified range, areasunder its curve can be exactly calculated. Forexample, slightly over two-thirds (68%) of allobservations are within one standard deviationon either side of the arithmetic mean and95% of all values are within ±2 (actually 1.96)standard deviations from the mean (Fig. 10.2).The mean of observations, or average, is theirtotal sum divided by the number of observa-tions, n.

Asymmetrical distributionsMuch of the data used in geology have an asym-metrical rather than a normal distribution.Usually such distributions have a preponder-ance of low values with a long tail of high val-ues. These data have a positive skew. Measuresof such populations include the mode which isthe value occurring with the greatest frequency(i.e. the highest probability), the median whichis the value midway in the frequency distribu-tion which divides the area below the distribu-tion curve into two equal parts, and the meanwhich is the arithmetic average of all values.In asymmetrical distributions the median liesbetween the mode and the mean (Fig. 10.3); innormal curves these three measures coincide.

FIG. 10.3 Asymmetrical distribution, positivelyskewed (to the right).

In asymmetrical distributions there isno comparable relationship between standarddeviation (or variance) and the area under thedistribution curve, as in a normal counterpart.Consequently, mathematical transformationshave been used whereby skewed data are trans-formed to normal. Perhaps the commonestis the log normal transformation where thenatural log of all values is used as the distri-bution and many geological data approxim-ate to this type of distribution; however, thismethod should be used with caution. Geovari-ances (2001) recommend the use of gaussiantransformation particularly for variable trans-formation in the geostatistical conditionalsimulation process (see section 10.4.3) and foruse in nonlinear geostatisical techniques suchas disjunctive kriging and uniform condition-ing (Deutsch 2002). An alternative approach isto use nonparametric statistics.

Parametric and nonparametric statisticsParametric statistics, discussed above, specifyconditions regarding the nature of the popula-tion being sampled. The major concern is thatthe values must have a normal distribution butwe know that many geological data do not havethis characteristic. Nonparametric statistics,however, are independent of this requirementand thus relevant to the type of statistical test-ing necessary in mineral exploration. Routinenonparametric statistical tests are available ascomputer programs, e.g. Henley (1981).

Point and interval estimates of a sample

Point estimates are the arithmetic mean (X),variance (S2), and sample size (n). The pointestimate X provides a best estimate of the popu-lation mean (Y) but by itself it is usually wrongand contains no information as to the size ofthis error. This is contained in the variance (S2).A quantitative measure of this, however, isobtained from a two-sided interval estimatebased on the square root of the variance, thestandard deviation (S).

Two-sided interval estimatesIf X is the mean of n random samples taken froma normal distribution with population mean Yand variance S2, then at a 95% probability X isnot more than 2 standard deviations (S) largeror smaller than X. In other words, there is a

Rel

ativ

e F

requ

ency

Variable

Low High

Med

ian

Mod

e

Mea

n

Page 219: Introduction to mineral exploration

202 M.K.G. WHATELEY & B. SCOTT

TABLE 10.2 Two-sided and one-sided confidenceintervals at 95% probability.

Two-sided One-sided

X 10.55% Zn 10.55% ZnS 3.60% 3.60%n 121 121 n 11 11(S/ n) 0.33 0.33t2.5% 1.96 NAt5% NA 1.65t(S/ n) 0.65 0.54Upper limit 11.20% NALower limit 9.90% 10.01%

1. The total number of samples needed to halve thestandard error of the mean (S/ n ), and thus the range of theconfidence interval, is 484 [= (2 × 11)2].2. With a one-sided confidence interval all the risk isplaced into one side of the interval. Consequently thecalculated limit is closer to the x statistic than to thecorresponding limit for a two-sided interval. In the aboveexample the upper limit could have been calculated insteadof the lower.3. See Table 10.1 for t values and Fig. 10.4 for a graphicalrepresentation.NA, not appropriate.

Fre

quen

cyF

requ

ency

X

2.5% 2.5%

11.20%9.90%10.55%

% Zinc

95% of values

Two-sided

X

5%

10.01%10.55%

% Zinc

95% of values

One-sided

FIG. 10.4 Graphical representation of a two-sidedand one-sided confidence interval at 95% confidencelevel. See Table 10.2 for calculations.

95% chance that the population mean is in thiscalculated interval and five chances in a hun-dred that it is outside. The relationship is:

X Y Y

.

.+ > > −

1 96 1 96σ σ

n n

This method is correct but useless because wedo not know σ and the expression becomes:

X Y X . % . %+ > > −t

S

nt

S

n2 5 2 5

where σ is replaced by its best estimate S, andthe normal distribution is replaced by a dis-tribution which has a wider spread than thenormal case. In this t distribution (Davis 2003)the shape of its distribution curve varies ac-cording to the number of samples (n) used to es-timate the population parameters (Table 10.1).As n increases, however, the curve resemblesthat of the normal case and for all practical pur-poses for more than 50 samples the curves areidentical. Examples of the use of this formulaare given in Table 10.2 and Fig. 10.4.

The width (i.e. spread) of a two-sided intervalestimate about the sample arithmetic meandepends on the term:

t

S

n =

TABLE 10.1 Critical values of t for different samplenumbers and confidence levels.

Number of Confidence levelssamples (n)

80% 90% 95% 99%

1 3.08 6.3 12.71 31.825 1.48 2.02 2.57 3.3710 1.37 1.81 2.23 2.7625 1.32 1.71 2.06 2.4940 1.30 1.68 2.02 2.4260 1.30 1.67 2.00 2.39120 1.29 1.66 1.98 2.36>120 1.28 1.65 1.96 2.33

The number of samples (n) is referred to as “the degrees offreedom.”

Page 220: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 203

TABLE 10.3 Effect of changing theconfidence level on the range ofa two-sided confidence interval.The same zinc mineralisationas in Table 10.2 is used. Grade(x) = 10.55% zinc, standarddeviation (S) = 3.60%, number ofsamples (n) = 121.

Confidence level

50% 80% 90% 95%

Percentile of t t25% t10% t5% t2.5%

Value of t 0.68 1.28 1.65 1.96t(S/ n) 0.22 0.42 0.54 0.65Upper limit (U1) 10.77% 10.98% 11.10% 11.20%Lower limit (U2) 10.33% 10.12% 10.00% 9.90%Range (U1–U2) 0.44% 0.76% 1.10% 1.30%

1. The range becomes narrower as the confidence level decreases, and the riskof the population mean being outside this interval correspondingly decreases.At the 50% level this risk is 1 in 2, but at the 95% level it is 1 in 20.2. See Table 10.1 for values of t.

The magnitude, and width, of this term isreduced by:1 Choosing a lower confidence interval (alower t value); an example of this approach isgiven in Table 10.3. The disadvantage of this isthat as the t value decreases the probability (i.e.chance) of the population mean being outsidethe calculated confidence interval increasesand usually a probability of 95% is taken as auseful compromise.2 Increasing the number of samples (n), but therelationship is an inverse square root factor.Thus the sample size, n, must be increased to4n to reduce the factor 1/n by half, and to 16n toreduce the factor to one quarter, clearly thequestion of cost effectiveness arises.3 Decreasing the size of the total variance, ortotal error [S2(TE)], of the sampling scheme.In this sense sampling error refers to variance(S2), or standard deviation (S). A small standarddeviation indicates a narrower interval aboutthe arithmetic mean and the smaller thisinterval the closer the sampling mean is to thepopulation mean. In other words, the accuracyof the sampling mean as a best estimate of thepopulation mean is improved by reducing thevariance. This is achieved by correct samplingprocedures, as discussed below.

One-sided interval estimatesIn this approach only one side of the estimate iscalculated, either an upper or lower limit, thenat 95% probability:

U t

S

nL = − < %X Y5 (UL = lower limit)

U t

S

nu %= − >X Y5 (UU = upper limit)

It is usually more important to set UL, so thatmined rock is above a certain cut-off grade.With a one-sided estimate all the risk hasbeen placed into one side of the interval. Con-sequently, the calculated limit for a one-sidedestimate is closer to the Y statistic than itis to the corresponding limit for a two-sidedinterval. Examples are given in Table 10.2and Fig. 10.4. Through interval estimates, thevariability (S), sample size (n), and the mean(Y) are incorporated into a single quantitativestatement.

10.1.2 Sampling error

Sampling consists of three main steps: (i) theactual extraction of the sample(s) from the insitu material comprising the population; (ii)the preparation of the assay portion which in-volves a mass reduction from a few kilograms(or tonnes) to a few grams for chemical ana-lysis; and (iii) the analysis of the assay portion.Gy (1992) and Pitard (1993) explain in somedetail the formula, usually known as Gy’s for-mula, to control the variances (errors) inducedduring sampling.

Page 221: Introduction to mineral exploration

204 M.K.G. WHATELEY & B. SCOTT

Global estimation error [S2(GE)]

Global error (GE) includes all errors in sampleextraction and preparation (total samplingerror or TE), and analysis errors (AE).

S2(GE) = S2(TE) + S2(AE)

The TE includes selecting the assay portionfor analysis while the analytical error properexcludes this step. Gy (1992) maintains thatanalysts usually complete their work in thelaboratory with great accuracy and precision.In this strict sense, associated analytical erroris negligible when compared with the varianceof taking the assay portion. Consequently, forpractical purposes S2(AE) ≈ 0 and from thisS2(GE) ≈ S2(TE).

Total sampling error [S2(TE)]

The total sampling error (TE) includes allextraction and preparation errors. It includeserrors of selection (SE) and preparation (PE):the former are inherent in both extraction andpreparation while the latter are restricted tomechanical processes used in sample reductionsuch as crushing, etc.:

S2(TE) = S2(SE) + S2(PE)

Selection errors (SE) are minimized by:1 A correct definition of the number of pointincrements, or samples. Sample spacing is bestdefined by geostatistics, described in section10.4.2 A correct definition of the area sampled.Sampling of mineral deposits presents a three-dimensional problem but for practical pur-poses they are sampled as two-dimensionalobjects which equate to the surface into whichsamples are cut. The long axis of the samplepreferably should be perpendicular to the dipof the mineralisation, or at least at an angle toit, but not parallel. The cut in plan is ideally acircle, sometimes a square or rectangle, and inlength should penetrate the full length of thesampled area.3 A correct extraction and collection of thematerial delimited by the above cut. Usuallythis presents a problem with the planes ofweakness in rocks, and their variation in hard-

ness. Following from (2) and (3) an optimumsample is provided by diamond drilling which,with 100% core recovery, cuts almost perfectcylindrical lengths of rock the full length of thesampled area.4 A correct preparation of the assay portionfrom the original sample, as described later.

Selection errors are of two main categories,which arise from:4.1 the inherent heterogeneity of the sampleand4.2 the selective nature of the sample extrac-tion and its subsequent reduction in mass andsize to the assay portion.

The first category comprises the funda-mental error [S2(FE)] and is the irreducibleminimum of the total sampling error andcan be estimated from the sampling model ofGy (1992) (section 10.1.4). Gy (1992) separatesthose of the second category into seven differ-ent types, which are estimated from measure-ments (i.e. mass and grain size) from each stageof a sample reduction system.

Preparation errors [S2(PE)] are related toprocesses such as weighing, drying, crushing,grinding, whose purpose is to bring successivesamples into the form required by the nextselection stage, and ultimate analysis. Sourcesof possible error are:1 Alteration of the sample’s chemical com-position by overheating during drying as, forexample, with coal and sulfides of mercury,arsenic, antimony, and bismuth.2 Alteration of the sample’s physical condi-tion. If the grain size of the sample is importantthis can be changed by drying and carelesshandling.3 Once a sample is collected, precautionsmust be taken to avoid losses during prepara-tion of the assay portion. Losing materialalways introduces error (i.e. increases the vari-ance) because the various sized fractions differin grade and it is usually the finer grainedmaterial which is lost. Losses can be checkedby weighing samples and rejects at each reduc-tion stage.4 Sample contamination must be avoided:sample containers and preparation circuitsmust be clean and free from foreign material.Dust produced by the size reduction ofother samples should be excluded. Accidentalsources of added material include steel chips

Page 222: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 205

from hammers, and that derived from thecrushing and grinding equipment.5 Unintentional sampling errors must beavoided such as mislabeling and the mixing offractions of different samples.

In essence preparation errors result fromthe careless treatment of samples and can becaused by the use of inadequate crushing andgrinding equipment, insufficient drying cap-acity, and inappropriate sample splitting.

Obviously, PE values can be estimatedonly from the results of an operating samplereduction system and, like SE values, cannot beestimated directly for either preliminary cal-culations or the design of such a system. For-tunately, the totality of both can be estimatedfrom the relationship:

S2(TE) ≤ 2S2(FE)

where the variance of the total sampling error(TE) does not exceed twice that of the funda-mental error (FE) (Gy 1992).

Summary

Sampling consists of three stages:1 Extraction of the original sample from in situmaterial (i.e. the population), including itsdelimitation and collection.2 Preparation which involves a reduction inboth mass and grain size of the original sampleto an assay portion for chemical analysis.3 Chemical analysis of the assay portion.

Each of the above three stages is discussed inturn.

10.1.3 Sample extraction

Random samplingCorrect sampling technique requires a randomselection of each sample from the popula-tion. At an operating mine, it is possible thatsampling is nonrandom since it is completedfrom mine development which is planned ona systematic basis. Provided, however, thatthe mineral particles which form the popula-tion are themselves randomly distributed thisobjection is overcome. Many mineral depositsdisplay trends and variations in the spatial dis-tribution of their grade variables and the effect

of this on the concept of their random distribu-tion is a matter of debate.

Sample volume–variance relationshipPreferably samples should be of equal volume,based on similar cross-sectional area. If vol-umes are unequal there must be no correla-tion between this and their analytical results.Although the mean of observations based onlarge volumes of rock should not be expectedto be different from those based on smallvolumes (e.g. larger or smaller diameter drillcore), the corresponding variance of samplesfrom the larger volumes might be expected tobe smaller. This is because the variability ofthe sample depends both on the variance andthe number of samples taken (n):

variability ∝

Sn

2

and instinctively we feel that there should be asimilar relationship with the volume of thesample taken. Indeed, it has been advancedthat:

variability

volume ∝

S2

It appears in practice, however, that there is nosuch simple, direct, relationship. This can betested by splitting samples in half (e.g. mineral-ized parts of a drill core), analyzing them separ-ately, and then combining the results as ifthe whole core sample had been tested. If thevalues from the two halves are statisticallyindependent values from a single population,then the ratio of the variance of the values forthe whole sample to the average values forthe two halves should be 0.5. Results show acalculated ratio of 0.9 or greater so that littleappears to be gained by analyzing the wholesample rather than half of it.

If the sample volume–variance relationshipwere important it would be desirable to analyzeall portions of a core from a drillhole ratherthan follow the common practice of splittingthe core longitudinally to save a half as a recordand use half for analysis. In sampling, however,this relationship should always be consideredand, if possible, tested. This is particularly

Page 223: Introduction to mineral exploration

206 M.K.G. WHATELEY & B. SCOTT

Length Pb +++++ SG Length Length LengthZn% ××××× Grade ××××× SG ××××× Grade

××××× SG

1.5 8.7 3.03 13.05 4.55 39.542.0 15.3 3.27 30.60 6.54 100.061.5 18.1 3.37 27.15 5.06 91.501.0 7.6 2.99 7.6 2.99 22.721.0 5.1 2.91 5.1 2.91 14.842.0 14.9 3.26 29.80 6.52 97.151.5 8.2 3.01 12.30 4.51 37.021.0 12.3 3.16 12.30 3.16 38.87

11.5 137.90 36.24 441.70

Average grade by volume % = 137.90 / 11.5 = 12.0%.Average grade by weight % = 441.70 / 36.24 = 12.2%.

TABLE 10.4 Sample density–volumerelationship.

relevant if the valuable constituents have a lowgrade and are erratically distributed (e.g. dis-seminated gold and PGE values) and little isknown of the variance. In such cases it is sens-ible either to slice the entire core longitudin-ally saving 10% as a record and sending 90% foranalysis, or collect a larger core.

This topic was discussed by Sutherlandand Dale (1984) in relation to the minimumsample size for sampling placer deposits, suchas alluvial diamonds, and they concluded thatthe major factor affecting variance was samplevolume. Annels (1991) explains that the stand-ard error of the mean, expressed as:

Standard error of the mean =

Sn

2

decreases as the number of samples increase.Taking larger samples will have a similareffect, because they will contain more “point”samples than a smaller volume of samples.When undertaking resource estimation, semi-variograms (section 10.4.3) should be con-structed for each sample set. These willdemonstrate whether the variances are propor-tional or whether there is a more fundamentaldifference between the sample sets.

Sample density–volume relationshipIn making reserve calculations, grades andtonnages are determined by using linear %,area %, and volume % relationships. Analyt-

ical data are expressed in terms of weight andunless the weight/volume ratio is unity, orvaries within a narrow range, average gradescalculated by volume % will be in error. Thisproblem is illustrated in Table 10.4 which is anexample of galena and sphalerite mineralisa-tion in limestone where the average grade cal-culated using volume percentage is lower thanthat calculated from weight percentage.

Sampling procedures

IntroductionVariance can be reduced by using well plannedsampling procedures with as thorough andmeticulous extraction of samples as possible.The sample location is surveyed to the nearestsurvey point and cleaned by either chippingor washing from the rock face all oxidizedand introduced material so as to expose freshrock. A sampling team usually consists oftwo people, the senior of whom maintains thesampling notebook. At each location a visualdescription of the rock is made, noting themain rock types and minerals present, percent-age of potential valuable mineral(s), rock alter-ation, other points of interest, and who istaking the sample. Obviously each sample hasa unique number – a safe practice is to usenumbered aluminum sample tags with oneplaced inside the plastic sample bag and theother attached by thin wire about its neck. Anycollection device, whether a scoop, shovel, or

Page 224: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 207

pipe, must obey the three times the maximumparticle size rule, i.e. the width, length, anddepth must be at least three times the maxi-mum particle size. This is defined as the largestscreen size which retains 5% of the material.There are three hand sampling techniques,namely channel, chip, and grab sampling.

Channel samplingIn mineral exploration, a channel is cut inan outcrop, usually the same diameter as thecore being collected, to maintain the volume–variance relationships. It is cut using a hammerand chisel or a circular saw, across the strike ofthe mineralisation (see Fig. 5.3). As the mater-ial is cut it is allowed to fall to the floor on toa plastic sheet or sample tray from which itis collected and bagged. Samples are normally0.5–5 kg in weight, mostly 1–2 kg, and each israrely taken over 2 m or so in length, butshould match the core sample length. Thereis no point in oversampling and sample spac-ing perhaps can be determined from the range(a) (section 10.4.1) derived from a study of asemi-variogram.

A mineral deposit can often be resolvedinto distinct and separate types of mineralisa-tion. Sampling these different types as separateentities rather than as one large sample canreduce natural variation and thus the variance,and keep sample weights to a minimum. Thisstratified sampling is also of value where theseparate types require different mineral pro-cessing techniques due to variation in eithergrain size or mineralogy. Such sample dataare then more useful for mine planning. Suchsamples can be recombined statistically intoone composite result for the whole of themineralisation.

In well-jointed or well-bedded rocks collec-tion by hammering the rock face presents dif-ficulties in that adjacent nonsample materialwill fall with the material being cut but thismust be rejected. Additionally, it is importantthat a representative collection of rock massesis collected. Frequently, mineral values of in-terest occur in altered rocks which may wellbe more friable than adjacent, harder, unminer-alized ground, and overcollection of this easilycut material will provide a biased sample.

Channel sampling provides the best tech-nique of delimiting and extracting a sampleand, consequently, provides the smallest pos-

sible contribution to the total error. The twotechniques which follow are less rigorous, andmore error prone, but less expensive.

Chip samplingChip samples are obtained by collecting rockparticles chipped from a surface, either along aline or over an area. In an established mine,rock chips from blastholes are sampled usingscoops, channels, or pipes pushed into theheap. With a large database of chip samples,statistical correlation between core and chipsamples may establish a correction factor forchip sampling results and thus reduce error.Chip sampling is used also as an inexpensivereconnaissance tool to see if mineralisationis of sufficient interest to warrant the moreexpensive channel sampling.

Grab samplingIn this case the samples of mineralized rockare not taken in place, as are channel and chipsamples, but consist of already broken mater-ial. Representative handfuls or shovelfuls ofbroken rock are picked at random at some con-venient location and these form the sample. Itis a low cost and rapid method and best usedwhere the mineralized rock has a low varianceand mineralisation and waste break into part-icles of about the same size. It is particularlyuseful as a means of quality control of miner-alisation at strategic sampling points such asstope outlets and in an open pit (Annels 1991).

Comparison of methodsBingham Canyon mine in the United Statesand the Palabora Mine in South Africa con-ducted a series of tests whereby the entireproduct from blasthole drilling was collected in30-degree segments from the full 360 degrees(C.P. Fish, personal communication). Oneblasthole may produce between 2 and 3 t. Eachsegment was sampled by the traditional meansdescribed above, and compared to the remain-ing material in the segment after crushing to3 mm and reducing in a 16-segment rotary-divider. The results showed:1 No segment ever gave the same assay as thewhole mass. The spatial variation was greatand there was no consistency in the way thematerial dispersed.2 Subsample analyses gave higher analyticalresults than the 100% mass. Scoop samples

Page 225: Introduction to mineral exploration

208 M.K.G. WHATELEY & B. SCOTT

tend to collect the finer particles as they areinvariably taken from the top part of any heap.The larger, usually less mineralized particlesroll down to the base of the perimeter and arenot sampled. When creating a channel, mater-ial will always collapse into the void. Thus, thetop of the heap is sampled. Similarly, in anupward scoop, the scoop is usually full beforeit gets anywhere near the top of the channel.When forcing a pipe sampler into a pile ofmaterial, the pipe preferentially collects thetop 6 inches (15 cm) of the material and thenblocks.3 Grab sample analyses also gave higher resultsthan the 100% sample.

10.1.4 Sample preparation

The ultimate purpose of sampling is to estim-ate the content of valuable constituents in thesamples (the sample mean and variance) andfrom this to infer their content in the popula-tion. Chemical analysis is completed on a few

grams of material (the assay portion) selectedfrom several kilograms or tonnes of sample(s).Consequently, after collection, sampling in-volves a systematic reduction of mass and grainsize as an inevitable prerequisite to analysis.This reduction is of the order of 1000 timeswith a kilogram sample and 1,000,000 with atonne sample.

Samples are reduced by crushing and grind-ing and the resultant finer grained material issplit, or separated, into discrete mass compon-ents for further reduction (Geelhoed & Glass2001). A sample reduction system is a sequenceof stages that progressively substitutes a seriesof smaller and smaller samples from the ori-ginal until the assay portion is obtained forchemical analysis. Such a system is essentiallyan alternation of size and mass reduction stageswith each stage generating a new sample and asample reject, and its own sampling error. Forthis reduction Gy (1992) established a relation-ship between the sample particle size, massand sampling error (Box 10.1):

BOX 10.1 Gy sampling reduction formula.

Gy (1992) is believed to be the first to devise a relationship between the sample mass (M), its particle size(d) and the variance of the sampling error (S2). This variance is that of the fundamental error (FE): a bestestimate of the total sampling error (TE) is obtained by doubling the FE.

M

CdS

≥3

2

where M is the minimum sample mass in grams and d is the particle size of the coarsest top 5% of thesample in centimeters. Cumulative size analyses are rarely available and from a practical viewpoint d isthe size of fragments that can be visually separated out as the coarsest of the batch. S2 is the fundamentalvariance and C is a heterogeneity constant characteristic of the material being sampled and = cβfg, where:

c = a mineralogical constitution factor =

− ( )1 a

aL

L

[(1 − aL)δA + δ G.aL]

aL is the amount of mineral of interest as a fraction,δA is the specific gravity of the mineral of interest in g cm−3,δG is the specific gravity of the gangue,β is a factor which represents the degree of liberation of the mineral of interest. It varies from 0.0 (noliberation) to 1.0 (perfect liberation) but practically it is seldom less than 0.1. If the liberation size (dlib) isnot known it is safe to use β equal to 1.0. If the liberation size is known d ≥ dlib then β = (dlib/d)0.5 which isless than 1. In practice there is no unique liberation size but rather a size range.f is a fragment shape factor and it is assumed in the formula that the general shape is spherical in whichcase f equals 0.5,g is a size dispersion factor and cannot be disassociated from d. Practically g extends from 0.20 to 0.75with the narrower the range of particle sizes the higher the value of g but, with the definition of d asabove, g equals 0.25.

Page 226: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 209

M

CdS

=3

2

M is the minimum sample mass in grams, dis the size of the coarsest top 5% of the samplein centimeters, C is a heterogeneity constantcharacteristic of the material being sampled,and S2 is the variance. Consequently, in reduc-ing the mass of a sample there is a cube rela-tionship between this reduction and its particlesize if the variance is to remain constant.

In practice the observed variance of asampling system is larger than that calculatedfrom the above model. This is because theobserved variance comprises the total samplingerror (TE) while the above model calculates aparticular component of this totality, the fun-damental error (FE). This important point hasalready been referred to, where S2(TE) = 2S2(FE).

Design of a sample reduction system

The choice of suitable crushing, grinding, dry-ing, and splitting equipment is critical and sys-tems are best designed and installed by thosewith appropriate experience. The design of areduction system is often neglected but it is ofprime importance particularly in dealing withlow grade and/or fine-grained mineralisation.In particular, equipment may have to handlewet, sticky material with a high clay contentwhich will choke normal comminution equip-ment. Also the creation of dust when pro-cessing dry material is an environmental andhealth hazard and may also contaminate othersamples. Dust has to be controlled by adequateventilation and appropriate isolation.

In any reduction system the most sensit-ive pieces of equipment are the crushers andgrinders. Each such item works efficientlywithin a limited range of weight throughputand size reduction and there has to be a real-istic assessment in the planning stage of theexpected number and weight of samples tobe processed each hour or shift. The use ofinadequate comminution equipment at bestmay introduce excessive preparation errors andat the worst can close the entire system.

The calculation of TE from twice the FEprovides a safe estimate (see previously) fromwhich a flexible reduction system can bedesigned. When it is in operation its actual

total error can be calculated and compared withthe initial assumption (see later).

Graphical solution of sample reduction

IntroductionSample preparation is essentially an alterna-tion of size and mass reduction within theparameters of the Gy formula given abovewhich provides for a control of the variance.Such reduction is best shown graphically withlog d (top particle size) contrasted with log M(sample mass), where a horizontal line repres-ents a size reduction and a vertical line a massreduction (Fig. 10.5).

These reductions have to be completedwithin an acceptable variance which can betaken as a relative standard deviation of lessthan 5%. Ideally what is required is a safetyline which divides the above graph into twoparts, so that on one side of the line all reduc-tion would be safe (i.e. with an acceptable vari-ance) while on the other side unacceptableerrors occur. Gy (1992) provides a mathemat-ical expression for such a line as:

Mo = Kd3

Empirically Gy (1992) suggests a value of125,000 for K but emphasizes that for low con-centrations of valuable constituents (<0.01%) ahigher K value should be used of up to 250,000.Consequently:

Mo = 125,000 d3 or log Mo = log K + 3 log d

This relationship on log–log graph paper(Fig. 10.5) is a straight line, with a slope of3.0, which divides the graph into two areas.1 To the left of this safety line Mo is greaterthan M and, consequently, from the Gy for-mula the fundamental sampling variance So2 isless than S2, and is acceptable.2 On the line, Mo = M and So2 = S2, which isacceptable. All points on the safety line cor-respond to a constant fundamental variance.3 To the right of the line Mo is less than M andthe sampling variance is too high (So2 > S2) andunacceptable and either Mo has to be increasedor d decreased to place the reduction point tothe left of the safety line.

Page 227: Introduction to mineral exploration

210 M.K.G. WHATELEY & B. SCOTT

SA

FE

TY

L I

NE

I

H

GF

ED

CB

A

S A F E AREA

U N S A F E AREA

( a )

Top particle size d

100µm mm

108

Mas

s (g

)

107

106

105

104

103

100

10

1

200 500 1 2 5 1 2 5 10 20cm

(b)

Mas

s (g

)

10,000

5000

2000

1000

500

200

SA

FE

TY

LIN

E100

10050201052

Top particle size d

1500200µm mm

100

50

20

10

5

2

I1

HG

FE

DC

B A

FIG. 10.5 Graphical representation of a sample reduction scheme. Fig. 10.5a is an example of a safe schemewhile Fig. 10.5b is unsafe. For explanation see text.

Equal variance linesThe above allocates equal variance to eachsampling stage but this may not always be anoptimum practice. Usually the larger the mass(M) to be sampled the larger the top particle size(d) and the higher its reduction cost. From theGy sampling formula, however, the necessityfor particle size reduction at any stage can bereduced by allocating a higher proportion ofthe total variance to that particular stage. Withthe mass (M) remaining constant this allowsa coarser top particle size. Consequently it ispreferable to allocate a maximum portion ofthe variance to the samples with the coarsestparticle sizes, and less to later successive sam-ples of finer grain size. Gy (1992) suggests thatone half of the total variance is allocated to thefirst sampling stage, one fourth to the next, etc.

As previously, a safe estimate of the totalsampling variance is calculated by:

S FE

CdM

23

( ) =

Using

M = Kd3

then

S FE

CdM

CK

23

( ) = =

and

So TE

CK

( ) = 2

Page 228: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 211

where C is the heterogeneity constant (Box10.1) and K equals 125,000.

From this a series of equal variance lines canbe drawn with decreasing values of K, such asK/2, K/4, etc., which correspond to increasingvariances. For each stage, knowing M and theincreased value of So2(FE) the related coarsergrain size (d) can be calculated from the Gyformula.

Variation of the safety lineThe heterogeneity constant (C) contains theliberation factor β (Box 10.1). The safety linesabove are calculated on the basis that thisfactor is at its maximum value of 1.0 whichis a safe assumption but only correct when thetop particle size is smaller than or equal to theliberation size (dlib) of the valuable compon-ents. If d is larger than dlib then β becomes[dlib/d]0.5, which is less than 1.0.

When the liberation size of the mineral ofinterest is not known the value of 1.0 is used asa safe procedure. This provides for the continu-ously straight equal variance lines as describedabove, with a slope of 3.0. When the liberationsize is known, the above formula can be usedwith a β value of less than 1.0 but approachingthis value as the grain size d is reduced. In thiscase each sampling stage has a unique safetyline slope from about 2.5 increasing to 3.0 as dapproaches dlib, with the less steep sections inthe coarser grain sizes. François-Bongarçon andGy (2002) point out that there are still commonerrors being made when applying Gy’s formula.They describe the main variable, the liberationsize, and illustrate the importance of modelingit with a worked example.

Splitting of samples during reduction

In each stage after reduction in grain size thesample mass is correspondingly reduced bysplitting. The sampling described in this text isconcerned with small batches of less than 40 tand usually less than 1 t which are stationaryin the sense that they are not moving streamsof material on, say, a conveyor belt. In explora-tion, stationary samples include chip samplesproduced by percussion or reverse circulationdrilling (section 10.4).

In the field it may be difficult to calculatethe weight of sample required. In such cases,

collect a minimum of 1000 particles. Thiswill seem a trivial amount when collecting, forinstance, a froth flotation product, but plantshave accurate sampling systems built into theprocess system. It becomes important whendealing with larger particles, such as RC orblasthole rock chips, rocks in stockpiles, oron trucks. Invariably the amount required willbe larger than anticipated, but don’t forget ifthe primary sample is unrepresentative, every-thing else is a waste of time.

Hand or mechanical shovel splittingSplitting samples is best completed by the pro-cess of alternative shovelling which minimizesthe splitting and preparation error contributionto the total error. The sample is spread eitheron a sheet of thick plastic or sheet iron whichis on a flat, clean, and smooth floor whichcan easily be swept. The material is piled intoa cone (step 1) which is then flattened intoa circular cake (step 2) and the first step isrepeated three times for maximum mixing ofthe sample. Studies have shown that rolling ona cloth or plastic sheet does not necessarilycause mixing. If too slow the material justslides, and if too fast, there is a loss of the finerparticles thereby causing bias. From the finalcone, shovelfuls of material are taken con-secutively around its circumference and placedalternatively into two separate piles. In thisway the original cone, and sample, is split intotwo separate but equal smaller cones and thechoice as to which is the new sample is decidedby the toss of a coin and not by the sampler.Further weight reduction is continued inexactly the same manner on the new sample.Alternatively, the cone may be quartered witha shovel and each quarter is further reduced inthe same way. This technique is referred to ascone and quartering.

The width of the hand, or small mechan-ical, shovel should be at least three times thewidth of the diameter of the coarsest sampleparticle size which provides for a maximumpossible size of about less than 100 mm withthe former and less than 500 mm with thelatter tool. Shovels should always be over-flowing and working with them partially filledis not a good practice. Obviously, towards theend of a splitting operation with finer grainsizes a smaller sized tool has to be used.

Page 229: Introduction to mineral exploration

212 M.K.G. WHATELEY & B. SCOTT

36.2 tonnes (80,000 lb)bulk sample

18.1 tonnes fortest work

18.1 tonnescrushed to minus 32 mm

32 mm screen

3.6 tonnescrushed to minus 10 mm

10 mm screen

ROTARY SPLITTER (4:1)

14.5 tonnes

SPLIT

3020 kg 580 kg

Rejects446 kg

72 kg

SPLIT (4:1)

Four replicate18 kg samples

23 cm × 30 cm rollcrusher to minus 5 mm

SPLIT

FIG. 10.6 Reduction of 36.2 tonne(80,000 lb) bulk sample bysuccessive crushing and splitting.This involves a concomitantreduction in grain size from +15 cmto 0.5 cm. (After Springett 1983b.)

and the broad range of gold particle size (i.e. dlib

in Gy’s formula) in the micrometer range,devising an adequate subsampling plant wasdifficult. The most problematical stage wasthe reduction of the initial bulk of 36.2-tonnesamples. This was effected by splitting eachsuch sample into two equal portions with onehalf (18.1 t) for sampling purposes and the otherfor heap leaching gold tests (Fig. 10.6). Figure10.7 illustrates a flow sheet for reduction of18 kg portions, which were treated in the samefashion as drilling samples, to four replicatesamples each of 5 g at a grain size of 0.01 mm(Springett 1983a,b).

In order to provide an adequate number ofincrements, shovel capacity should be betweenone thirtieth to one fiftieth of the originalsample mass to provide some 30–50 shovelfuls,or increments, divided equally between thetwo splits. In a final, fine-grained 100 g samplethe required asssay portion can be obtainedby repeated alternative “shovelling” with aspatula.

An example of the reduction of a bulk samplefor chemical analysis is given in Figs 10.6 and10.7. The example is from a large, low grade,disseminated gold deposit with a grade of about1 g t−1. Due to the high specific gravity of gold

Page 230: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 213

Instrument splittersThe Jones Riffle is the most satisfactory of sta-tionary instrument splitters (Fig. 10.8) whichdivide a dry sample into two equal parts. Itcomprises an even number (between 12 and 20)of equally sized chutes of a slope between 45and 60 degrees where adjacent chutes dischargeon opposite sides of the instrument. For cor-rect implementation, the material to be split isevenly spread on a rectangular scoop designedto fit the splitter width, and discharged slowlyand evenly into the chutes. This preventsloss of fines due to overenthusiastic pouring.To prevent obstruction the chutes should beat least twice the diameter of the coarsestparticle. With repeated splitting it is betterpractice to implement an even rather than anodd number of splitting stages, and to altern-ate the choice of right and left split samples.Asymmetrical splitters are now availablewhich split samples on a 70:30 or 80:20 basis(Christison 2002). This speeds up samplereduction and, with fewer separations, reducesany errors. A sample splitter may be used at thedrill site especially when percussion or reversecirculation drilling is in progress and samples

are collected at regular intervals. A 1 m sectionof 102 mm diameter hole produces approxim-ately 21 kg of rock chips.

Rotary splitters capable of separating labor-atory size (a few kilograms) or bulk samples(a few tonnes) are also available. They usecentrifugal force to split a sample into four orsix containers.

The use of several riffle splitters in series,each one directly receiving the split from thepreceding stage, is poor practice which canintroduce sampling errors.

Comparison of calculated and observed totalsampling errors

A best estimate of the total sampling error (TE)is obtained from the fundamental error (FE):

S2(TE) = 2S2(FE)

This estimate is used to design a sample reduc-tion system from which, when in operation, anobserved S2(FE) can be obtained by samplingeach reduction stage to establish the samplemass (M) and the maximum particle size (d),

23 cm × 23 cm rolls crusher(crushed to minus 1.7 mm)

SPLIT to 450 g

Four replicate 5 g samplesfor assay and reference

Core orreverse circulation drill

cuttings

Disc pulverizerto minus 0.4 mm

Disc pulverizerto minus 0.1 mm

SPLIT

Jaw crusher

18 kg samplesfrom bulk plant

(at 5 mm)

FIG. 10.7 Reduction of 18 kg (40 lb)samples at 0.5 cm size from bulksampling plant (Fig. 10.5), anddrill material, to 5 g samples at−0.01 cm size suitable for chemicalanalysis. (After Springett 1983b.)

Page 231: Introduction to mineral exploration

214 M.K.G. WHATELEY & B. SCOTT

using Gy’s formula. This is then comparedwith the previously estimated total error, onthe following basis:1 When the observed total error is smaller thanor of the same order of magnitude as the esti-mated total error nothing should be changedin the reduction system. This provides a safetymargin against the future wear of the equip-ment which will deliver progressively coarsermaterial for which the sampling varianceincreases.2 When the observed total error is largerthan the estimated total error the former hasto be reduced and particularly at the first (orcoarsest) reduction stage where the varianceis likely to be the major component of S2(FE).This can be achieved by retaining the samenumber of sampling stages but increasing theweight (i.e. the mass) treated at each separatestage and increasing the number of stages butwith the same sample weight at each stage.Usually the latter choice is preferable as the

FIG. 10.8 A Jones sample splitter.For explanation see text.

crushing–grinding equipment does not have tobe enlarged, as in the first option.

The total sampling error of any samplereduction system should be checked at re-gular intervals, say every 6 months dependingupon throughput and use, to ensure that it isworking satisfactorily.

10.1.5 Preparation of samples containingnative gold and platinum

With native precious metals a problem arisesfrom the significant difference between thedensity of an associated gangue (about 2.6) andthat of the metals (110.0 to 21.5). Gold maybe diluted with silver and copper to a densityin the order of 15.0 but even at this level itsdensity is twice that of galena, the densest of allcommon sulfide minerals, and six times thatof quartz. This contrast means that whensuch valuable constituents are liberated fromtheir gangue, flakes and nuggets will segregate

Page 232: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 215

rapidly under gravity and homogenization ofsuch material is difficult (Austr. Inst. Geosci-entists 1987, 1988, Royle 1995).

Calculation of the variance of thefundamental error

Under such circumstances, and with thevery low concentrations of precious metals(≈1 ppm), Gy (1992) suggests a modification ofhis formula for calculating the fundamentalerror.

Normally, the factor C = cβfg (Box 10.1) butfor gold and platinum Gy (1992) recommends afactor of:

C

da

SCML

.

= =2 4 3

2and

where aL is the concentration of the valu-able constituent expressed as a fraction (i.e.1 ppm = 1 × 10−6).

An example of the use of this formula isgiven in Box 10.2.

10.1.6 Analytical errors

According to Gy (1992) analytical errors (AE),in the strict sense, are negligible and errors

which are attributed to analytical laboratoriesarise from the selection of the assay portionand not, on the whole, from actual analyticaltechniques (section 10.1.2). Errors occur, how-ever, and vary from laboratory to laboratory.An example is given in Table 10.5 where anartificial sample was made from gold in silicasand, at 2.74 g t−1 and was sent to four differentlaboratories for analysis (Springett 1983b). Thefollowing conclusions can be drawn:

Laboratory 1 Results are both inaccurate andimprecise.Laboratory 2 Results are accurate butimprecise.Laboratory 3 Results are inaccurate butprecise.Laboratory 4 Results are accurate andprecise.Obviously only the results from Laboratory 4

are acceptable.The scope of the control of error in chemical

analysis is beyond this chapter but these errorsmust be considered (see section 8.2). Externalcontrol is exercised by submitting duplicatesamples and reference materials of similarmatrices to the unknowns (see section 8.2.1). Itis essential that the geologist is aware of theinternal control procedures of the laboratory.These usually employ the analysis of replicate

BOX 10.2 Variance of fundamental error [S2(FE)] with liberated gold mineralization.

Gold mineralization contains 1 g t−1 (aL = 10−6) and is liberated from its gangue (i.e. β = 1) with thecoarsest particles about 2 mm in size (d = 0.2 cm.). Under such circumstances Gy (1992) recommends amodification of his standard formula for the calculation of the fundamental variance. Normally theheterogeneity factor C equals cβfg (Box 10.1), but for free gold he recommends C = 2.4d3/AL andS2(FE) = C/M.

With a 1 tonne sample (M = 106 g), C = 2.4(0.2)3/10− 6 = 19,200 and S2(FE) = 1.92 × 10−2, S = 1.4 × 10−1, or14%. If a normal distribution is assumed the 95% confidence level double-sided interval is ±28%, whichis a low precision. If, say, a 10% precision is required at 95% confidence then a larger sample (M) isneeded. At this level S(FE) equals 5 × 10−2, and S2(FE) equals 25 × 10−4; then

M = C/S2(FE) = 19,200/25 × 10−4 = 768 × 104 g or 7.68 tonnes.

The treatment of such large samples presents a problem, but Gy (1992) recommends the use of stratifiedsampling. In this process the sample is reduced to a grain size of a few millimeters and then concentratedon a shaking table, such as the Wilfley table (Wills 1985). This produces a high density concentrate thatcontains all the coarse gold, an intermediate fraction, and low density tailings which is largely gangue.The intermediate fraction is recycled on the table while the tailings are sampled again on a 1% samplingratio. In this way the 8 tonne sample is reduced to about 80 kg which is then processed as in Figs 10.6and 10.7.

Page 233: Introduction to mineral exploration

216 M.K.G. WHATELEY & B. SCOTT

familiar and the following remarks mainlyapply to work at remote prospects.

As much background information as possibleshould be collected from previous publicationsand reports. A preliminary description of theexercise serves as a source of stimulation thatwill help to isolate points on which furtherthought is needed. The tasks to be completedhave to be briefly but adequately described.

Logistical factors are important. How muchtime is available for the exercise? Can the sitebe revisited or is this the only opportunity forsampling? What budget restraints are there? Isassistance necessary and if so of what type –surveyors, drivers, cooks, laborers, etc.? Decideon the objectives and how it is proposed toachieve them; are they cost-effective?

What is the purpose of the sampling? Isit purely qualitative, where it is necessaryto prove the presence or absence of certainminerals or chemical elements, or is a stat-istically valid estimate of the grade of (say) abase metal prospect required? Is the study pre-liminary (i.e. a return visit is possible) or final(i.e. no chance of return)? Certainly, beforecarrying out extensive systematic sampling atvarious locations, an orientation survey shouldbe completed to ascertain the main character-istics of the prospective target population– a preliminary mean, standard deviation, geo-chemical threshold, and probability of error(see also section 8.3). Finally, check that thesampling procedure is sound and adequateand that the projected sample size (weightand number) is large enough for the purposebut not too large. Carefully assess the cost-effectiveness of the proposed exercise.

Safety

Safety aspects are paramount. Before visiting anew site or abandoned mines, samplers shouldundertake a risk assessment (Grayson 2001).This will consist of hazard identification andrisk assessment, followed by developing pro-cedures to manage the risk and control thehazards. Well-designed risk management pro-grams will also bring health aspects intoconsideration. Effective assessment of hazardsrequired good analytical tools and judgment.Ask for experienced help when starting outin the field. A formalized safety code will be

TABLE 10.5 Errors in assays on a prepared sample of2.74 grams per tonne gold. This is an example wherethe population parameters are known since thematerial was artificially prepared. The results ofthe four different sets of five identical samples(A to E), each analyzed separately at four differentlaboratories (1 to 4), are given below. (After Springett1983b.)

Laboratory (g t−−−−−1 gold)

1 2 3 4

SampleA 5.72 2.74 4.39 2.81B 4.18 2.40 4.01 2.78C 5.66 2.06 3.81 2.67D 4.42 3.43 4.01 2.67E 3.26 2.06 3.81 2.71

Mean 4.65 2.54 4.01 2.73S2 0.87 0.26 0.04 0.003S 0.93 0.51 0.21 0.06

1. The range of values at the 95% confidence level (which isthe mean ±2S) for laboratory 1 is 2.79–6.51 g t−1. This rangedoes not contain the population mean (2.74 g t−1) and thevariance is high. For laboratory 2 the range is 1.52 to3.56 g t−1 which does contain the population mean, but thevariance is high. At laboratory 3 the range is 3.59–4.43 g t−1.This does not contain the population mean, although thevariance is acceptable. Laboratory 4 results have a range of2.61–2.85 g t−1, which contains the population mean andhas a low variance.2. Only the results from laboratory 4 are acceptable.

samples (e.g. every, say, tenth sample isrepeated) and by introducing artificially pre-pared reference materials (e.g. every 30th or40th sample). Some laboratories are found toperform relatively poorly and a rigorous pro-gram of interlaboratory checking by the useof blanks, duplicates and prepared materials isimportant (Reimann 1989).

10.1.7 Sample acquisition

Before sampling

Pre-sampling procedures consist of gaining asmuch familiarity as possible with the geologyand possible problems before visiting the loca-tion to be tested. Obviously if the sampling is atan operating mine these problems are already

Page 234: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 217

available in all exploration offices and must beread and understood. Samplers should rarely,if ever, work alone and visits to mine work-ings (surface or underground) should be accom-panied. Visits to abandoned mine workingsneed particular care and the workings shouldbe assessed fully for potential weak rockconditions, weak roof support, water extent,noxious gases, and other hazards before samp-ling commences.

After sampling

As an example, suppose the calculated mean ofa prospect from 25 samples (n = 25) is 5.72%zinc with a standard deviation of 0.35% zinc.At a 95% probability level a two-sided confi-dence interval is:

5 72

0 35

255 72

0 35

252 5 2 5. %

. % . %

. %. % . %+ > > −t tY

= + > . % .

. % 5 72 2 06

0 355

Y

> − . % .. %

5 72 2 060 35

5

= 5.72% ± 0.15% at a 95% confidence level

That is, there are 19 chances in 20 of the popula-tion mean (i.e. the true grade) being within thisrange, and one chance in 20 of it being outside.

If we are dissatisfied with this result theremedy, in theory, is simple. If we wish a nar-rower confidence interval the site can be revis-ited and additional samples taken (i.e. increasen), but remember the square root relationshipin the above calculation. That is, if the intervalis to be halved (to become 5.72% ± 0.075%),then the total number of samples (n) required isnot 50 but 100 – an additional 75 samples haveto be taken. Alternatively, the value of the tstatistic can be reduced (Table 10.3). At an 80%confidence level the interval is:

5 72 1 32

0 355

5 72 1 320 35

5. % .

. % . % .

. %+ > > −Y

= 5.72 ± 0.09%.

In other words there are 16 chances in 20 of thepopulation mean being within this narrowerrange. As it is narrower the probability of it

being outside the range is higher – four chancesin 20 or quadruple the previous calculation.

Commonsense must be used in the inter-pretation of the results of any sampling pro-gram and particularly with the problem ofre-sampling in order to increase the n statistic.As an example, a zinc prospect 200 km fromthe nearest road or railway is sampled andpreliminary estimates are 2.1% zinc with anestimated overall reserve tonnage of about200,000 t. There is no point in sharpening thisgrade estimate by taking more samples as thelow tonnage, approximate grade and locationmean that this deposit is of little economicinterest. There is nothing to be gained by extrawork which may demonstrate that a betterestimate is 2.3% zinc.

10.1.8 Summary

In a perfect world we would all like to followexactly the sampling theories of Pierre Gyand Francis Pitard. In practice a company doesnot always have the equipment in place, or theopportunity or the skill to implement them,particularly when operating in the field. Itis important therefore to pick the relevantimportant sampling theory facts and use themat the basic level in a way that an operator canunderstand. The person with the shovel doesnot wish to see a complicated mathmaticalformula. An explanation of what is importantcan yield real dividends. It is relatively easy toget from poor sampling to good sampling, butvery difficuilt (and the cost may be exponential)to get from good to perfect.

Sample acquisition and preparation methodsmust be clearly defined, easy to implementand monitor, bias free, as precise and accur-ate as possible, and cost effective. This can beachieved by:1 Making a thorough analysis of the materialwith attention to the particle size distributionand the composition of the particles in eachsize class. These details are required to estab-lish an optimal sampling strategy.2 Ensuring that the batch-to-sample size ratioand the sample-to-maximum particle size ratioare sufficiently large.3 Determining the optimum sample size byconsidering the economics of the entire process.In some cases, the sample size is determined by

Page 235: Introduction to mineral exploration

218 M.K.G. WHATELEY & B. SCOTT

minimizing the cost of collecting the sampleswhen compared with re-sampling.

10.2 PITTING AND TRENCHING

In areas where soil cover is thin, the locationand testing of bedrock mineralisation is maderelatively straightforward by the examinationand sampling of outcrops. However in loca-tions of thick cover such testing may involvea deep sampling program by pitting, trenching,or drilling. Pitting to depths of up to 30 m isfeasible and, with trenching, forms the simplestand least expensive method of deep samplingbut is much more costly below the water table.For safety purposes, all pits and trenches arefilled in when evaluation work is completed.Drilling penetrates to greater depth but is moreexpensive and requires specialized equipmentand expertise that may be supplied by a con-tractor. Despite their relatively shallow depth,pits and trenches have some distinct advant-ages over drilling in that detailed geologicallogging can be carried out, and large and, ifnecessary, undisturbed samples collected.

10.2.1 Pitting

In areas where the ground is wet, or labor isexpensive, pits are best dug with a mechanicalexcavator. Pits dug to depths of 3–4 m are com-mon and with large equipment excavation to6 m can be achieved. In wet, soft ground any pitdeeper than 1 m is dangerous and boardingmust be used. Diggers excavate rapidly and pits3–4 m deep can be dug, logged, sampled, andre-filled within an hour. In tropical regions,thick lateritic soil forms ideal conditions forpitting and, provided the soil is dry, verticalpits to 30 m depth can be safely excavated. Twolaborers are used and with a 1 m square pit,using simple local equipment, advances of upto 2 m per day down to 10 m depth are possible,with half that rate for depths from 10 to 20 m,and half again to 30 m depth.

10.2.2 Trenching

Trenching is usually completed at right anglesto the general strike to test and sample overlong lengths, as across a mineralized zone.

Excavation can be either by hand, mechanicaldigger, or by bulldozer on sloping ground.Excavated depths of up to 4 m are common.

10.3 DRILLING

10.3.1 Auger drilling

Augers are hand-held or truck-mounted drills,which have rods with spiral flights to bring softmaterial to the surface. They are used particu-larly to sample placer deposits. Power augersare particularly useful for deep sampling ineasily penetrable material where pitting is notpracticable (Barrett 1987). They vary in sizefrom those used to dig fence post holes to large,truck-mounted rigs capable of reaching depthsof up to 60 m, but depths of less than 30 m aremore common. Hole diameters are from 5 to15 cm in the larger units, although holes 1 min diameter were drilled to evaluate the Argylediamond deposit in Australia. In soft groundaugering is rapid and sampling procedures needto be well organized to cope with the materialcontinuously brought to the surface by thespiralling action of the auger. Considerable careis required to minimize cross-contaminationbetween samples. Augers are light drills and areincapable of penetrating either hard ground orboulders. For this purpose, and holes deeperthan about 60 m, heavier equipment is neces-sary and this is described in the next section.

10.3.2 Other drilling

For anyone interested in understanding thesubsurface, drilling is the most frequently usedtechnology. The various methods of drillingserve different purposes at various stages ofan exploration program (Annels 1991). TheAustralian Drilling Industry Training Com-mittee (1997) gives a comprehensive account ofmethods, applications, and safety issues. Earlyon when budgets are low, inexpensive drilling isrequired. The disadvantage of cheaper methods,such as augering, rotary or percussion drilling,is that the quality of sampling is poor with con-siderable mixing of different levels in the hole.Later, more expensive, but quality samples areusually collected using reverse circulation ordiamond core drilling as shown in the follow-ing table:

Page 236: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 219

depths of up to 200 m are possible, but depthsof 100–150 m are more usual. Drill cuttings areflushed to the surface by compressed air and aregular and efficient source of this is vital forthe successful operation of this technique. Asthe cuttings come to the surface they can berelated to the depth of the hole. However, suchdirect correlation is not always reliable asholes may not be cased and material may fallfrom higher levels. In wet ground the liftingeffect of compressed air is rapidly dissipatedbut special foaming agents are available toalleviate this drawback. Rigs are usually truckor track mounted and very mobile. With thelatter slopes of up to 30 degrees can be negoti-ated and the ability to drill on slopes steeperthan 25 degrees partly eliminates the need towork on access roads.2 Top hammer drills. As the name suggeststhe hammer unit, driven by compressed air,is at the top of the drill stem and the energyto the noncoring drill bit is imparted throughthe drill rods. These are usually lighter unitsthan down-the-hole hammer drills. They areused for holes up to 10 cm diameter and depthsof up to 100 m, but more usually 20 m. Mostonly use light air compressors and this restrictsdrilling depths to, at the most, only a fewmeters below a water table. It becomes impos-sible for the pressurized air to blow the heavywet rock sludge to the surface. Usually they aremounted on either light trucks or tractors.

General remarksPercussion drilling is a rapid and cheap methodbut suffers from the great disadvantage of notproviding the precise location of samples, as isthe case in diamond drilling. However, costsare one third to one half of those for diamonddrilling and this technique has proved par-ticularly useful in evaluating deposits whichpresent more of a sampling problem than ageological one, e.g. a porphyry copper. There is

Purpose Quality Type

Reconnaissance and exploration Low Auger, rotary and percussion (chips)High Reverse circulation (chips)

Resource or reserve evaluation High Reverse circulation coring in soft formationsDiamond or sonic core drilling in hard formations

Fundamentally, drilling is concerned withmaking a small diameter hole (usually lessthan 1 m and in mineral exploration usuallyonly a few centimeters) in a particular geo-logical target, which may be several hundredmeters distant, to recover a representativesample. Among the available methods rotary,percussion, reverse circulation, and diamondcore drilling are the most important.

Rotary drilling

Rotary drilling is a noncoring method andis unequalled for drilling through soft tomedium hard rocks such as limestone, chalk,or mudstone. A typical rotary bit is the triconeor roller rock bit that is tipped with tungstencarbide insets. Rock chips are flushed to thesurface by the drilling fluid for examinationand advances of up to 100 m per hour are pos-sible. This type of drilling is typically used inthe oil industry with large diameter holes(>20 cm) to depths of several thousand meterswith the extensive use of drilling muds to liftthe rock chips to the surface. The large size ofthe equipment presents a mobility problem.

Percussion drilling

In percussion drilling, a hammer unit driven bycompressed air imparts a series of short rapidblows to the drill rods or bit and at the sametime imparts a rotary motion. The drills vary insize from small hand-held units (as used in roadrepair work) to large truck-mounted rigs cap-able of drilling large diameter holes to severalhundred meters depth (see Fig. 11.3). The unitscan be divided broadly into two types:1 Down-the-hole hammer drills. The hammerunit is lowered into the hole attached to thelower end of the drill rods to operate a non-coring, tungsten carbide-tipped, drill bit. Holeswith diameters of up to 20 cm and penetration

Page 237: Introduction to mineral exploration

220 M.K.G. WHATELEY & B. SCOTT

a rapid penetration rate of around a meter aminute and it is possible to drill 150–200 m inan 8-hour shift. With such penetration ratesand several machines, several hundred samplescan be collected each day. As each 1.5 m of a10 cm diameter hole is likely to produce about20–30 kg of rock chips and dust, sample col-lection and examination requires a high degreeof organization. Like all compressed air equip-ment these drills are noisy in operation.

Reverse circulation

The reverse circulation (RC) drilling techniquehas been in use in exploration since the mid-seventies and can be used in unconsolidatedsediments such as alluvial deposits or for drill-ing rock. Both air and water can be used as thedrill flushing medium and both cuttings or corecan be recovered. The technique employs adouble-walled string of drill rods (Fig. 10.9),with either a compressed air driven percussionhammer or a rotating tungsten carbide coringbit at the cutting end of the string. The mediumis supplied to the cutting bit between the twin-walled drill rods and returned to the surface upthe center of the rods. In the case of percussiondrilling the rock chippings are also transportedto the surface up the center of the rods and fromthere, via a flexible pipe, to a cyclone wherethey are deposited in a sample collection con-tainer (Fig. 10.9).

The advantages of using this method to col-lect rock chippings, rather than auger, rotary orpercussion drilling, are that the entire sampleis collected, the method is extremely quick (upto 40 m per hour can be drilled) and there isvery little contamination. The specialized rods,the need for a compressor and additional equip-ment makes this a more expensive drillingtechnique than auger or percussion drilling,but the additional costs are outweighed by thehigher quality of sample collection. The dualnature of some RC rigs (chips and core) meansthat high quality core can be taken through thezone of interest without the need to mobilizea second (core) rig, thus reducing the overalldrilling costs.

Diamond core drilling

The sample is cut from the target by adiamond-armoured or impregnated bit. This

produces a cylinder of rock that is recoveredfrom the inner tube of a core barrel. The bitand core barrel are connected to the surfaceby a continuous length (string) of steel oraluminum alloy rods, which allow the bit pluscore barrel to be lowered into the hole, andpulled to the surface. They also transmit therotary cutting motion to the diamond bitfrom the surface diesel power unit, and appro-priate pressure to its cutting edge (Fig. 10.10& Table 10.6).1 Drill bits. Drill bits are classified as eitherimpregnated or surface-set. The former consistof fine-grained synthetic or industrial gradediamonds within a metallic cement while thelatter have individual diamonds, sized by theirnumber per carat. In general, impregnated bitsare suitable for tough compact rocks such aschert, while surface-set varieties with largeindividual diamonds are used for softer rockssuch as limestone (Fig. 10.11).

Diamond bits will penetrate any rock in timebut because of their high cost and the needto maximize core advance and core recoverywith minimum bit wear, the choice of bitrequires considerable experience and judg-ment. Second-hand (i.e. used) surface-set bitsalso have a diamond salvage value and con-sequently are not used to destruction. Drillbit diameters are classified with either a lettercode (American practice) or in millimeters(European practice) (Table 10.7).2 Core barrels. As the cylinder of rock (thecore) is cut by the circular motion of the drillbit it is forced up into the core barrel by theadvancing drill rods. Core barrels are classifiedby the length of core they contain. They areusually 1.5–3.0 m in length but can be as longas 6 m. They are normally double-tubed in thesense that in order to improve core recovery aninner core barrel is independent of the motionof the drill rods and does not rotate. Triple-tubed barrels can be used in poor ground andfor collecting undisturbed samples for geo-technical analysis.

Previously, to recover core the barrel hadto be removed from the hole by pulling theentire length of drill rods to the surface, a time-consuming process. Wire-line drilling (Q seriescore) is now standard practice; in this methodthe barrel is pulled to the surface inside theconnecting drill rods using a thin steel cable.This has the advantage of saving time but often

Page 238: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 221

Drill string

Sample

Cyclone

Water

Air /waterDrill rods: each 3 m long

Drill truck Support truck

Sample

Drill bit

Flushing withcompressed

air /water

FIG. 10.9 Schematic diagram and photograph showing a reverse circulation drill. The drill was exploring formineral sands in Western Australia (White 1989).

Page 239: Introduction to mineral exploration

222 M.K.G. WHATELEY & B. SCOTT

FIG. 10.10 A diamond drill. From left to right the components are the power, control, drill, and water pumpunits. (Reproduced by permission of Diamond Boart Craelius Ltd.)

Drill unit weight 890 kg Length of holes:Power unit weight 790 kg AQ (48 mm) 425 mMotor rating 40 kW at 2200 rpm−1 BQ (60 mm) 350 mFlush pump flow 76 liters min−1 NQ (76 mm) 250 mFlush pump pressure 5 Mpa

The drill is a fully hydraulic rig. It is suitable for both underground and surfaceunits and is powered by either an air, diesel or electric motor. To a certain extentpenetration rates can be increased by an increase in pressure on the drill bit. Thisis regulated by hydraulic rams in the drill unit at surface and transmitted to thebit by the drill rods. A limit is reached when all the diamond chips in the bit’ssurface have been pressed as far as possible into the rock. Higher penetrationrates can then only be achieved by increasing the bit rotation speed. Excessivelyhigh bit pressure, with low rotation speed, results in abnormally high wear of thebit. Conversely, low pressure with high rotation results in low penetration andpolishing (i.e. blunting) of the diamonds.

In the above unit, drill water would flush through a 60-mm-diameter hole atabout 50 cm s−1. In a 350-m-length hole return water would take about 12minutes to pass from the drill bit to the surface. The ascending velocity of thefluid should be greater than the settling velocity of the largest rock particlesgenerated by the cutting and grinding action of the bit.

TABLE 10.6 Diamond drilling rigs.In mineral exploration most drillholes are less than 450 m in length.A typical drill rig for this purpose isshown in Fig. 10.10 but manymodels are in use. Some of thespecifications for the drill inFig. 10.10 are given in this table.

Page 240: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 223

flush away part of the sample and this cansometimes be alleviated with the addition ofspecialized additives. Water may be used incombination with various clays or chemicals,which, in addition to the above functions,have added uses such as sealing the rock faceof the drillhole. The science and use of suchdrilling fluids is a subject in itself and ismuch developed in the oil industry whichdrills significantly larger diameter and deeperholes than are required in mineral exploration

FIG. 10.11 On the left-hand side are several types of surface-set diamond bits, on the right are bits withtungsten carbide insets and at the back are impregnated bits. (Reproduced by permission of Diamond BoartCraelius Ltd.)

necessitates having a smaller diameter core(Table 10.7).3 Circulating medium. Usually water is cir-culated down the inside of the drill rods, wash-ing over the cutting surface of the drill bit, andreturning to surface through the narrow spacebetween the outside of the rods and the wallof the drillhole. The purpose of this action is tolubricate the bit, cool it, and remove crushedand ground rock fragments from the bit sur-face. In friable and soft rock this action may

Page 241: Introduction to mineral exploration

224 M.K.G. WHATELEY & B. SCOTT

(this can be adjusted using hydraulic rams atthe surface), the right number of revolutionsper minute, and the correct choice of drill bit.Until recently there was no direct method ofknowing what operating conditions existed atthe drill bit, which are the critical factors ifdrilling efficiency is to be optimized. Recentdevelopments in “measurements while drill-ing (MWD)” have overcome this and allow avariety of direct measurements such as tem-perature, pressure at the bit face, and rate ofwater flow, to be made (Dickinson et al. 1986,Prain 1989, Schunnesson & Holme 1997).Another technique being developed is a re-tractable bit that can be raised, replaced, andlowered through the string of drill rods. At pre-sent when a worn bit requires replacementthe complete drill string has to be withdrawn,which is a time-consuming process (Jenner1986, Morris 1986). However, modern rigs withautomated rod handling can pull the string andunscrew the rods at very high speed.

Drilling advance rates of up to 10 m an hourare possible but depend a great deal upon theskill of the driller and rock conditions. Costsvary from $US50 to 80 a meter for holes up to

Size Hole diameter Core diameter Core area(mm) (mm) (% of hole area)

METRIC SIZESConventional drilling36T 36.3 21.7 3656T2 56.3 41.7 5566T2 66.3 51.7 61

Wire-line drilling56ST 56.3 35.3 39

AMERICAN SIZESConventional drillingAX 48 30.1 39BX 59.9 42.0 49NX 75.7 54.7 52

Wire-line drillingAQ 48 27 31BQ 60 36.5 36NQ 75.8 47.6 39HQ 96.1 63.5 44

The hole diameter is the diameter of rock cut by the drill bit. Part of this isremoved by the cutting and grinding action of the bit and water flushed to thesurface, while the central part is retained as core in the core barrel.

TABLE 10.7 A selection of core,drillhole and drill bit sizes.

(Australian Drilling Industry Training Com-mittee (1997).4 Casing. Cylindrical casing is used to seal therock face of the hole. It provides a steel tube inwhich the drill string can operate in safety andprevents loss of drill strings caused by rock col-lapse and either loss or influx of water. Casingand drill bits are sized so that the next lowersize (i.e. smaller diameter) will pass throughthe larger size in which the hole had been previ-ously drilled.5 Speed and cost of drilling. Drilling machinesused in mineral exploration usually have a cap-acity of up to 2000 m, exceptionally to 6000 m(see section 14.4) and may be inclined fromhorizontal to vertical. The rate of advancedepends upon the type of drill rig, the bit, thehole diameter (usually the larger the holediameter the slower the advance), the depthof hole (the deeper the hole the slower theadvance), the rock type being drilled (hard orsoft, friable, well jointed, etc.), and the skill ofthe driller.

The driller must decide on the best volumeof drilling fluid to be flushed over the drill bit,the right pressure to apply at its cutting surface

Page 242: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 225

300 m long and from $US70 to 150 a meterfor lengths up to 1000 m in accessible areas.Standard references for diamond drilling areAustralian Drilling Industry Training Com-mittee (1997), Cummings and Wickland (1985),the Australian Drillers Guide (Eggington 1985)and The Management of Drilling Projects(1981).

Sonic core drilling

Sonic drilling retrieves core but without thecontamination caused by drilling muds. Sonicdrilling applies the principle of harmonics todrill and case a borehole (Potts 2003). It uses avariable-frequency drill head to transmit vibra-tion energy through the drill pipe and corebarrel to allow continuous core sampling totake place. Sonic drilling can penetrate over-burden, fine sand, boulders, and hard rock. Itcan collect samples up to 254 mm in diameterand can drill up to 200 m vertically or in in-clined holes. The big advantage of sonic drillingis that uncontaminated, undisturbed samplescan be collected because no air, water, or otherdrilling medium is used. Sonic drilling canrealize 100% core recovery, even in glacial till,clay, sands and gravels as well as hard rock. Itoffers rapid penetration, reduced on-site costsand minimal environmental impact. In thelater case the clean up and waste disposal costsare significantly lower.

10.3.3 Logging of drillhole samples

Information from drillholes comes from thefollowing main sources: rock, core, or chips;down-the-hole geophysical equipment; instru-ments inside the hole (see MWD above); andperformance of the drilling machinery. In thissection we are only concerned with geolo-gical logging, but the geologist on site at adrill location must be familiar with all sourcesof information. The collection of geotechnicaldata from core is discussed in section 10.6.

Geological logging

Effective core recovery is essential – that is, thelength or volume (weight) of sample recovereddivided by the length or volume (weight) drilledexpressed as a percentage. If recovery is less

than 85–90% the value of the core is doubtfulas mineralized and altered rock zones are fre-quently most friable and the first to be groundaway and lost during drilling. The core is notthen representative of the rock drilled, it is nota true sample, and it is probably misleading(Fig. 10.12).

Core drillingOften, initial, rapid core logging is done at thedrill site. This information is used to decidewhether the hole is to be either continued orabandoned. Wetted core is more easily exam-ined, using either a hand lens or a binocularmicroscope. Most organizations have a stand-ard procedure for core logging and a standardterminology to describe geological features.Field data loggers are now used to gather com-pany standardized digital data, which are down-loaded to the central database upon return tothe field base or office (see section 9.1). Onionsand Tweedie (1992) discuss the time and costssaved using this integrated approach to datagathering, storage, and processing.

Once the initial logging at a drill site is com-plete, the core is moved to a field base, wherea more detailed examination of the core takesplace at a later date. Nevertheless, the mainstructural features should be recorded (frac-ture spacing and orientation) and a lithologicaldescription (colour, texture, mineralogy, rockalteration, and rock name) with other detailssuch as core recovery and the location of exces-sive core loss (when say >5%). The descrip-tion should be systematic and as quantitativeas possible; qualitative descriptions should beavoided. These data are plotted on graphicalcore logs (see Fig. 5.13) and used as an aid ininterpreting the geology of the current and nextholes to be drilled.

Core is stored in slotted wooden, plastic ormetal boxes short enough to allow two personsto lift and stack them easily (Fig. 10.13; seeFig. 13.10). Core is collected for a variety ofpurposes other than geological description, e.g.metallurgical testing and assaying. For theselatter purposes the core is measured into appro-priate lengths (remembering the principle ofstratified sampling) and divided or split intotwo equal halves either by a diamond saw or amechanical splitter. Half the core is sent forassay or other investigations whilst the other

Page 243: Introduction to mineral exploration

226 M.K.G. WHATELEY & B. SCOTT

affect the measurement of relative density ofthe core (Preston & Sander 1993) or alter thenature of the material.

Rock chips and dust (“sludge”) can be col-lected during core drilling; they represent therock cut away by the diamond drill bit. Drill-ing with air circulation in relatively shallowholes (as in most percussion drilling) deliverscuttings to the surface within a minute orso. However, with core drilling, water circula-tion, and longer holes, there is an appreciable

40° Fragment missing

Chlorite andore minerals

Chlorite and quartz

Chlorite and feldspar

116.96 cm%

44 cm (true thickness)

at a weighted average core recovery of 91%

Sample1

2

3

4

5

5 cm co

re lo

ss

15 cm

15 cm

31 cm

15 cm

23 cm

254.65 m

253.66 m

Slate footwall

Lode zone

Slate

hanging–wall

Samplenumber

Depth Recovery True width(cm)

%Sn cm%

1

2

3

4

5

from(m) to(m) (%)(m)

253.66 253.81

253.81 253.96

253.96 254.27

254.27 254.42

254.42 254.65

0.15 100

100

81

100

100

0.15

0.25

0.15

0.23

12

12

20

12

18

0.03

0.56

4.75

1.27

NIL

0.36

6.72

95.00

15.24

Weighted average lode value = = 2.66% Sn

Core intersection angle average 40°

FIG. 10.12 Typical intersection of atin-bearing vein showing samplingintervals and the uncertaintyintroduced by incomplete corerecovery. (After Walsham 1967.)

half is returned to the core box for record pur-poses. Obviously structural features have to berecorded before splitting and a good practice isto photograph wet core, box by box, before log-ging it, to produce a permanent photographicrecord (Fig. 13.10).

When core from coal seams is to be sampledor the samples are to be collected for geotech-nical analysis, the core should be sealed as soonas the core leaves the core barrel. This is toprevent loss of moisture, which can adversely

Page 244: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 227

contained information is important particu-larly as during mining some drill locations maybe permanently lost.

Noncore drillingIn noncore drilling the chips and dust are usu-ally collected at 1- to 2-m intervals, dried andseparately bagged at the drill site. After wash-ing they are relatively easy to examine with theuse of a hand lens and binocular microscope.Samples can be panned so as to recover a heavymineral concentrate. It is a good practice to

FIG. 10.13 Diamond drill core being examined and stored in a wooden core box. Note the excellent corerecovery. (Reproduced by permission of Diamond Boart Craelius Ltd.)

time lag before cuttings reach the surface. It isestimated that from a depth of 1000 m cuttingscan take 20–30 minutes to reach the surface,with the inherent danger of differential settle-ment in the column of rising water due to dif-ferences in mineral and rock-specific gravitiesand shape. Consequently rock sludge is rarelyexamined during core drilling.

Obtaining core is expensive so it is sensibleto retain it for future examination. How-ever, adequate and long-term storage involvestime, space, and expense, but the value of the

Page 245: Introduction to mineral exploration

228 M.K.G. WHATELEY & B. SCOTT

1 Mobilization and transport of equipment tothe drilling site. This can vary from movementalong a major road by truck to transport byhelicopter.2 Setting up at each site and movementbetween successive borehole locations. Againcosts can vary greatly depending on distanceand terrain.3 A basic cost per meter of hole drilled.4 Optional items costed individually, e.g.cementing holes, casing holes, surveying.5 Demobilization and return of equipment todriller’s depot.

All items of cost should be detailed in thecontract. There may be a difference of interestbetween the driller and the company repres-entative on site, normally a geologist, who isthere to see that drilling proceeds accordingto the company’s plan and the contract. Thedriller may be paid by distance drilled duringeach shift whilst the geologist in charge is moreconcerned with adequate core recovery andthat the hole is proceeding to its desired target.The company representative usually signsdocuments at the completion of each shiftwhere progress and problems are describedand it is on the basis of these documents that

sprinkle, and glue, a sample of rock chips andpanned concentrates from each sample intervalon to a board so that a continuous visual repres-entation of the hole can be made. Again, de-scriptions must be systematic and quantitative.

10.3.4 Drilling contracts

Drilling can be carried out with either in-house(company) equipment or it can be contractedout to specialist drilling companies. In thelatter case the conditions of drilling, amount ofwork required, and the cost will be specified ina written contract. The purpose of drilling is tosafely obtain a representative sample of the tar-get mineralisation in a cost-effective manner.Once the mining company has been assured, inwriting, of the drilling company’s safe workingpractices, the choice of drilling equipment iscrucial and much depends upon the experi-ence of the project manager. Unless the drillingconditions are well known test work should,if possible, be carried out to compare differentdrilling methods before any large-scale pro-gram begins (Box 10.3).

The main items of cost in a contract are asfollows:

BOX 10.3 Comparative sampling results from two different drilling methods. (after Springett 1983b.)

If drilling conditions are not known a program of test work to compare and contrast appropriate drillingprocedures may be desirable before any large scale sampling program is commenced, as in the followingexample.

Silver mineralization 12.2 m thick consisting of fine-grained dolomite and quartz with minorchalcopyrite, galena, and sphalerite occurs in limestone. The silver is principally acanthite but withsome native silver. It has been tested by diamond and percussion drilling, and also by mining from a testshaft.

Diamond drilling: BQ core, diameter 36.5 mm with samples every 1 m over the total length of 12.2 m;core recovery was 93%. The core was split for assay and the total weight of the overall sample was 10 kg.The weighted assay average was 62 g t−1 silver. The main concern in diamond drilling was the possibilityof high grade, friable silver mineralization being ground by the diamond bit and washed away by thecirculating water.

Percussion drilling: With a 1.6-cm-diameter hole samples were collected every 1 m; material recoverywas 87%. All the recovered material was taken for the analysis and totalled 255 kg. The weighted assayaverage was 86 g t−1. Whilst this method was cheaper than core drilling, there was the possibility that thehigh specific gravity silver minerals might sink in the column of water in the bore hole and not be fullyrecovered.

Bulk sample: From a 1.1-m square test shaft all the material was taken for assay, totalling 51.8 t. Theweighted assay average was 78 g t−1.

Accepting the bulk sample result as the best estimate of silver grade the cheaper percussion drillingmethod can be used to sample the deposit rather than the more expensive core drilling.

Page 246: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 229

the final cash payment is made. Consequently,it is important for this representative to bethoroughly familiar with the contract and theproblems that can arise at a drilling site, par-ticularly if in a remote location.

The client (i.e. the company) is at liberty tofix specifications such as, say, a plus 90% corerecovery, a vertical hole with less than 5degrees deviation. It is then at the discretionof the drilling company as to whether or notthese requirements are accepted, but if they areand the conditions are not met the failure isremedied at the expense of the contractor.

10.3.5 Borehole surveying

Drills are usually unpredictable wanderers.In section holes drilled at a low angle to rockstructures tend to follow that structure whileholes at a higher angle tend to become perpen-dicular to the feature. A similar behavior canbe expected from alternations of hard and softrock (e.g. layers of chert in chalk). Steep holesare likely to flatten and all holes tend to spiralcounter clockwise with the rotation of the drillrods. However, wandering also depends uponthe pressure on the drilling bit and its rate ofrevolution, so that two drillholes in similargeological situations may take different paths.Holes with a length of 1000 m have been foundto be several 100 m off course – while indi-vidual lengths of drill rod are essentially inflex-ible, a drill string several hundred meters inlength will bend and curve. Since drillholes areused for sampling at depth, if the orientation ofthe hole is not known the location of the sam-ple is similarly unknown, and geological pro-jections based on these samples will be in error.

Borehole surveys are conducted as a matterof routine in all holes and there are a varietyof instruments for this purpose. They mostcommonly indicate the direction (azimuth) andinclination of the hole at selected intervals,commonly every 100 m, using a small mag-netic or gyroscopic compass. The calculationof corrected positions at successive depths isa straightforward mathematical procedure, ifone knows the location of the top of the hole(X, Y, and Z co-ordinates) and the initial holeinclination.

Deliberate deflection can be achieved eitherby wedging holes (say, 1 degree per wedge)

or using a hydraulically driven downholemotor, which acts as a directional controldevice (Cooper & Sternberg 1988, Dawson &Tokle 1999). This relatively new directional-drilling tool was originally designed for the oilindustry. The downhole motor is powered bythe circulating fluids passed down the centerof the rod string, which drives the bit with-out drill-rod rotation. A spring deflection shoelocated just above the bit acts as a directionalcontrol device and delivers a constant side pres-sure to the drillhole wall forcing the bit tomove (and deflect) in the opposite direction.The amount of spring loading in the shoe is pre-set before the unit is lowered into the hole and,obviously, progress is monitored by closelyspaced surveys. The deflection unit is a specialtool and is not used in normal drilling.

10.3.6 Drillhole patterns

The pattern of drilling in an exploration pro-gram depends primarily upon the intendeduse of the data. In reconnaissance work wherethe geology is poorly known the first holesmay be isolated from each other and drilled forgeological orientation purposes. In explorationfor sedimentary uranium deposits, coal andborates, holes may be drilled up to 10 km apartto locate sedimentary formations of interestand obtain structural data. Later holes may be afew kilometers apart whilst drilling a specifictarget calls for “fences” or lines of holes, at aclose spacing of 100 –200 m. However, the loca-tion and spacing of holes in the last eventual-ity are guided by detailed geological mapping,geochemical, geophysical, and geostatisticalresults (see sections 5.2 and 10.4.1).

Generally, a systematic grid of drillholes(and samples) taken normal to the expectedmineralized zone is a preferred pattern. Thisprovides a good statistical coverage and geo-logical cross sections can be made with aminimum of projection. Such fences maywell be 200–400 m apart with individualholes spaced at 100–200 m intervals, whichallows room for systematic infill drilling ifthe results justify extra work. The inclina-tion of individual holes is obviously import-ant and generally they should be drilled atright angles to the expected average dip of themineralisation.

Page 247: Introduction to mineral exploration

230 M.K.G. WHATELEY & B. SCOTT

Before drilling a hole it is recommended thata section be drawn along its projected length,allowing for hole deviation if this is possible.As the hole progresses the section is modified.Drilling is an expensive and a time-consumingpart of mineral exploration. The objective isto drill a precise number of holes, within bud-get, safely, and provide the exact number ofintersections needed to demonstrate the gradeand tonnage (dimensions) of the mineralisa-tion at an appropriate level of accuracy andprecision. Unfortunately, such optimizationis rarely, if ever, possible. At second best aniterative procedure is followed whereby aseries of successively closer spaced drilling pro-grams are completed with each followed by are-assessment of all existing data – a process ofsuccessive approximations. The main problemis that in the calculation of mineral resourcesthe zone of influence of each sample is notknown until a minimum amount of work iscompleted. If two adjacent samples (taking adrillhole as a sample) cannot be correlated at anacceptable confidence level, then neither hasan acceptable zone of influence in the inter-vening space and further sampling is necessary(see section 10.4.1). Conversely where adjacentsamples show appropriate correlation furthersampling is not required.

There are several methods of assigningzones of influence to either successive samplesor individual boreholes: the mean-square (dn)successive difference test, the use of correla-tion coefficients (R2) or geostatistics using therange (a) of a semi-variogram. However, a relat-ively large nugget effect (section 10.4.3) in thesemi-variogram indicates an element of un-certainty that would call for a reduction inspacing. Such an effect is apparent in the evalu-ation of low-grade disseminated gold depositswhere drilling may be required on a ±50 m gridbasis.

10.4 MINERAL RESOURCE AND ORE RESERVEESTIMATION, GRADE CALCULATIONS

During exploration and initial evaluation of abase metal, industrial mineral, or coal deposit,the principal emphasis is placed upon its geo-logy and the estimation of the quality andquantity of the resources present. Data are

collected from several sampling programs asdescribed above, including trenching, pitting,percussion drilling, reverse circulation drilling,and diamond drilling, as well as during trialmining. The object of this sampling is to pro-vide a mineral inventory. Once the miningengineers and financial analysts have estab-lished that the mineral can be mined at a profit,or in the case of industrial minerals can bemarketed at a profit, it can then be referred toas an ore reserve.

The stock market controversies that sur-rounded the discovery of the Poseidon nickeldeposits in Western Australia in the 1960sled the Minerals Council of Australia to es-tablish a committee to resolve these issues.When the Australasian Institute of Mining andMetallurgy joined it, the resulting committeewas called the Australasian Joint Ore ReservesCommittee (JORC). Other controversies, in-cluding that surrounding the Bre-X deposit inIndonesia in the 1990s (see section 5.4), pro-mpted Australian and north American stockexchanges to adopt rigorous procedures forcompanies to report their resources. The leaderin determining these procedures has beenJORC, which has defined the JORC Code (Aus-tralasian Joint Ore Reserves Committee 2003)that has been taken up by a number of compan-ies and stock exchanges, and forms the basis formost international reporting systems.

As early as 1976 the United States Geolo-gical Survey (USGS 1976) published defini-tions of reserves and resources for mineraland coal deposits. These publications take intoaccount the increasing degree of confidencein the resources and the financial feasibility ofmining them.

Between 1972 and 1989, a number of reportswere issued by JORC which made recom-mendations on public reporting, and resourceand reserve classification. These graduallydeveloped the principles now incorporated inthe JORC Code. In 2004, JORC published itslatest version of the code.

The JORC Code is currently used as a modelfor reporting codes of other countries withappropriate modifications to reflect local con-ditions and regulatory systems. Examplesinclude the USA (SME 2004), South Africa(SAIMM 2004), and UK (IOM3 2004). Agree-ment has also been reached with the United

Page 248: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 231

Nations (UNCE 2001) on a common set ofterms to describe resources and reserves, and aninternational set of definitions and guidelines.

10.4.1 The JORC Code

The JORC code sets out a system for class-ifying tonnage and grade estimates as eitherore reserves or mineral resources, and of sub-dividing these into categories that reflect dif-ferent levels of confidence (Fig. 10.14). Thepurpose of this is purely for public reporting.JORC does not regulate how estimates shouldbe done, nor does it attempt to quantify theamount of data needed for each mineralresource category.

A Competent Person(s) makes these deci-sions based on direct knowledge of the depositin question. The criteria used may be entirelysubjective or they may involve some quantitat-ive measures, or mixtures of both. There areno prescribed methods that can be applied “offthe shelf.” It is very important that the selectedmethods of distinguishing resource classesshould be geologically sensible in the contextof the deposit being studied.

Stephenson (2000) explained that the JORCCode has been operating successfully for over10 years and, together with complementarydevelopments in stock exchange listing rules,has brought about substantially improvedstandards of public reporting by Australasian

mining and exploration companies. The suc-cess of the Code is based upon its early adop-tion in full by the Australian and New ZealandStock Exchanges, and the ability and willing-ness of the mining industry to bring Com-petent Persons to account when necessary.

The JORC Code is designed for reportingof mineral resources, which it defines as aconcentration or occurrence of material ofintrinsic economic interest in or on the Earth’scrust in such form and quantity that there arereasonable prospects for eventual economic ex-traction. The location, quantity, grade, geolog-ical characteristics, and continuity of a MineralResource are known, estimated, or interpretedfrom specific geological evidence and know-ledge, by a Competent Person appointed by theCompany. Mineral Resources are subdivided,in order of increasing geological confidence,into Inferred, Indicated, and Measured cate-gories. Portions of a deposit that do not havereasonable prospects for eventual economicextraction must not be included in a MineralResource.

A mineral resource is not an inventory ofall mineralisation drilled or sampled, regard-less of cut-off grade, likely mining dimensions,location, or continuity. It is a realistic inven-tory of mineralisation, which, under assumedand justifiable technical and economic con-ditions, might, in whole or in part, becomeeconomically extractable.

Exploration Results

Mineral Resources

Inferred

Indicated Probable

Measured Proved

Increasing levelof geologicalknowledge andconfidence

Consideration of mining, metallurgical, economic, marketing,legal, environmental, social and governmental factors

(the modifying factors)

Ore Reserves

FIG. 10.14 A JORC Box used as anearly basis for classifying mineralresources and reserves. It takes intoaccount the increasing degree ofconfidence in the resources and thefinancial feasibility of miningthem. (From Australasian Joint OreReserves Committee 2003.)

Page 249: Introduction to mineral exploration

232 M.K.G. WHATELEY & B. SCOTT

JORC defines ore and reserves as theeconomically mineable part of a measured orindicated mineral resource. Reserves includediluting materials and allowances for mininglosses. Reserves are usually assessed duringfeasibility studies and include considerationof and modification by realistically assumedmining, metallurgical, economic, marketing,legal, environmental, social and governmentalfactors. These assessments demonstrate at thetime of reporting that extraction could reason-ably be justified. Ore reserves are subdivided inorder of increasing confidence into probableand proved ore reserves. A probable ore reserveis the economically mineable part of an indi-cated, and in some circumstances measured,mineral resource. A probable ore reserve isthe economically mineable part of a measuredmineral resource.

Geologists are interested in the amount andquality of in situ mineralisation within ageologically defined envelope. In situ mineralresources tend to be slightly greater than theore reserves. Discussions of the subject aregiven by the Australasian Joint Ore ReservesCommittee (JORC 2003), Whateley and Harvey(1994), and Taylor (1989).

Before resource calculations can proceed,a study of the mineralisation envelope isrequired. This envelope can often be definedby readily identifiable geological boundaries. Insome cases inferences and projections must bemade and borne in mind when assessing con-fidence in the resources. Some deposits canonly be delineated by selecting an assay cut-offto define geographically and quantitatively thepotential mineralized limits (see Chapter 16,where an assay cut-off is used to define theoutline of the Trinity Silver Mine and wherewaste blocks are identified within the depositfor separate disposal). Initially this is highlysubjective. In later stages when confidence ishigher, the ore limits will be carefully calcu-lated from conceptual mining, metallurgical,cost, and marketing data (see section 11.2).

Geologists must provide the basis for invest-ment decisions in mining using the data pro-vided from the above sampling. It is theirresponsibility to ensure that the database thatthey are using is valid (see section 9.1). Oncethe geological and assay (grade) cut-offs (Lane1988, King 2000) have been established, usu-

ally in discussions with the mining engineers,it is possible to start the evaluation procedure.

It is therefore important to understand thecut-off grade theory and grade tonnage curves.

10.4.2 Cut-off grade theory

Geologists will be required to understandthe concept of cut-off grade theory (Lane 1988,King 2000), because they will be required toidentify only the material above the lowestgrade that will be included in the potentiallyeconomic part of the deposit. This is definedas the cut-off grade, which separates the miner-alized rock from barren or low grade rock.Deposits with only one valuable metal, such asgold, would normally use grade (g t−1) to rankthe resource. In a simple process, the highestgrade material is processed to recover the goldand the low grade material sent to a wastedump.

The resource potential will therefore bedetermined by the cut-off grade. It is commonpractice to calculate the resource tonnage at aseries of cut-off grades. This information isplotted on a grade–tonnage graph (Fig. 10.15).Eventually the material above cut-off will beused to develop a mine plan and the blockswithin the mine will be scheduled for extrac-tion. The schedule will be affected by locationand distribution of ore in respect to topographyand elevation, mineral types, physical charac-teristics, and grade–tonnage distribution, anddirect operating expenses associated with min-ing, processing, and converting the commodityinto a saleable form. The ultimate aim is tocalculate the net present value (NPV; see sec-tion 11.5.3).

All this is based upon the tonnage and gradecalculations made by the geologist. Althoughcut-off grades are used to classify material intostreams for processing, stockpiling, or beingsent to waste dumps, Fig. 10.16 shows howerrors in predicting properties of the resourcemay result in misclassification of this material(Wellmer 1998). The ellipse represents a bandof confidence between the estimated and ac-tual values (typically 95% confidence band).The same cut-off has been applied to both axes.It is apparent that there are two areas wherematerial is classified correctly and two areasof misclassification. Misclassified waste has a

Page 250: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 233

0

20

40

60

80

100

120

140

160

Ton

nage

Grade (%)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Cut-off grade (%)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

0

200

400

600

800

1000

1200

(b)

(a)

Cum

ulat

ive

tonn

age Cut-off grade

Cut-off grade

FIG. 10.15 Grade–tonnage graphs. (a) Resourcetonnage frequency distribution at different grades.(b) Resource tonnage cumulative frequencydistribution at different grades.

grade below the cut-off grade but is estimatedto have a grade above cut-off. This misclas-sified waste will be mined and processed asore. Likewise, misclassified ore is classifiedas waste and dumped without extracting anymetal, with the subsequent loss of revenue.The importance of making as good an estimateas possible cannot be overemphasized. Some ofthe methods available to make these estimatesare described below.

10.4.3 Tonnage calculation methods

Classical statistical methods

The use of classical statistical methods inresource calculations is generally restricted to

Est

imat

ed g

rade

Wastemisclassifiedas ore

Wastecorrectlyclassified

Ore misclassifiedas waste

Ore correctlyclassified

Cut-off grade

Cut

-off

grad

e

Actual grade

FIG. 10.16 A confidence band ellipse showing howerrors in predicting properties of the resource mayresult in misclassification of this material (Wellmer1998). The ellipse represents a band of confidencebetween the estimated and actual values (typically95% confidence band).

a global estimation of volume or grade withinthe mineralisation envelope. The simpleststatistical estimate involves calculating themean of a series of values, which gives an aver-age value within the geologically defined area.Calculation of the variance will give a measureof the error of estimation of the mean (Davis2003). When using these nonspatial statist-ical methods it is important that the samplesare independent of one another, i.e. random.If sample locations were chosen because ofsome assumption or geological knowledge ofthe deposit then the results may contain bias.Separate calculations must also be made ofdifferent geological areas to ensure that theresults are meaningful.

The data are usually plotted on frequencydistribution graphs (histograms) (see Figs 8.4,10.1 & 16.10) and scatter diagrams (correla-tion graphs) (see Fig. 8.6). The distributions areusually found to approximate to a gaussian orto a log–normal distribution. Some geostatist-ical techniques only operate in gaussian space.Techniques exist that enable geologists totransform the variable data into gaussian space

Page 251: Introduction to mineral exploration

234 M.K.G. WHATELEY & B. SCOTT

(Geovariances 2001). Cumulative frequencydistributions can also be plotted on normalprobability paper and a straight line on such aplot represents a normal distribution. Signific-ant departures from a straight line may indicatethe presence of more than one grade zone andthus more than one population within the data.Each population should be treated separately(c.f. section 8.3.1).

Conventional methods

In calculating the resource or reserve potentialof a deposit, one formula, or a variation of it, isused throughout, namely:

T = A × Th × BD

where T = tonnage (in tonnes), A = area ofinfluence on a plan or section in km2, Th =thickness of the deposit within the area ofinfluence in meters, BD = bulk density. This

includes the volume of the pore spaces. It isobtained by laboratory measurement of fieldsamples.

The area of influence is derived from a planor section of the geologically defined deposit.The conventional methods commonly usedfor obtaining these areas are: thickness con-tours (constructed manually), polygons, trian-gles, cross-sections, or a random stratified grid(Fig. 10.17). Popoff (1966) outlined the prin-ciples and conventional methods of resourcecalculation in some detail. The choice ofmethod depends upon the shape, dimensionsand complexity of the mineral deposit, and thetype, dimensions and pattern of spacing of thesampling information. These methods havevarious drawbacks that relate to the assump-tions on which they are based, especially thearea of influence of the sampling data, and gen-erally do not take into account any correlationof mineralisation between sample points norquantify any error of estimation.

(c)

a

bcg

(b)(a)

(d)

1

2

3D

(e)

1

32

20

25

10

30

1520

25

1510

152025

2010

15

FIG. 10.17 Conventional methods of estimating the area of influence of a sample using (a) isopachs,(b) polygons, (c) triangles, (d) cross-sections, or (e) a random stratified grid. (From Whateley 1992.)

Page 252: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 235

Large errors in estimation of thickness (orgrade) can therefore be made when assumingthat the thickness or grade of a block is equalto the thickness or grade of a single samplepoint about which the block has been drawn.Care should be taken when using the cross-sectional method that the appropriate formulais used for computing volume between sec-tions, as significant errors in tonnage estima-tion can be introduced where the area andshapes on adjacent sections vary considerably.Wire frame modeling, in technical softwarepackages such as Datamine, offers the user theability to construct a series of sections betweenwhich the user can link the areas of interestwith a wire frame. The package calculatesthe volume, limiting some of the errors intro-duced by manual methods. Manual contouringmethods are less prone to this estimationerror by virtue of the fact that contours areconstructed by linear interpolation, therebysmoothing the data irregularities.

The inverse power of the distance (1/Dn)method can also be used to calculate resources,usually using a geological modeling softwarepackage. The deposit is divided into a series ofregular blocks within the geologically definedboundary, and the available data are used tocalculate the thickness value for the centerof each block. Near sample points are givengreater weighting than points further away.The weighted average value for each blockis calculated using the following generalformula:

Th

V ff

( )

=∑ ×

where Th = the thickness at any block center, V= known thickness value at a sample point,f = 1/Dn weighting function, n = the power towhich the distance (D) is raised.

The geologist has to decide how many ofthe available data are to be used. This is norm-ally decided by choosing a distance factor forthe search area (usually the distance, a, derivedfrom the range of a semi-variogram), the powerfactor which should be employed (normallyD2), and how many sample points should beused to calculate the center point for eachblock. The time needed to calculate many

hundreds of block values on a complex depositrequires the aid of a computer. This methodbegins to take the spatial distribution of datapoints into account in the calculations.

Most of these methods have found applica-tion in the mining industry, some with con-siderable success. The results from one methodcan be cross checked using one of the othermethods.

Geostatistical methods

Semi-variogramsThe variables in a deposit (grade, thickness,etc.) are a function of the geological envir-onment. Changing geological and structuralconditions results in variations in grade orquality and thickness between deposits andeven within one deposit. However, samplesthat are taken close together tend to reflect thesame geological conditions, and have similarthickness and quality. As the sample distanceincreases, the similarity, or degree of correla-tion, decreases, until at some distance therewill be no correlation. Usually the thicknessvariable correlates over a greater distance thancoal quality variables (Whateley 1991, 1992)or grade.

Geostatistical methods quantify this con-cept of spatial variability within a depositand display it in the form of a semi-variogram(Box 10.4). Once a semi-variogram has beencalculated, it must be interpreted by fittinga “model” to it. This model will help to iden-tify the characteristics of the deposit. An ex-ample of a model semi-variogram is shownin Fig. 10.22. The model fitted to the experi-mental data has the mathematical form shownin the two equations below. It is known as aspherical scheme model and is the most com-mon type used, although other types do exist(Journel & Huijbregts 1978, David 1988, Isaaks& Srivastava 1989, Annels 1991).

There is often a discontinuity near theorigin, and this is called the nugget varianceor nugget effect (C0). This is generally attribut-able to differences in sample values over verysmall distances, e.g. two halves of core, andcan include inaccuracies in sampling andassaying, and the associated random errors (seesection 10.1.2). C0 is added where a nuggeteffect exists:

Page 253: Introduction to mineral exploration

236 M.K.G. WHATELEY & B. SCOTT

BOX 10.4 Calculating the semi-variogram.

It is assumed that the variability between two samples depends upon the distance between them andtheir relative orientation. By definition this variability (semi-variance or γ (h)) is represented as half theaverage squared difference between samples that are a given separation or lag distance (h) apart (see theequation below).

In the example (Fig. 10.18), the sample grid shows thickness values of a stratiform mineral deposit orcoal seam at 100 m centers. Suppose we call this distance between the samples (in the first instance100 m) and their orientation, h. We can calculate the semi-variance from the difference between samplesat different distances (lags) and in different orientations, e.g. at 200 m (Fig. 10.22) and 300 m (Fig. 10.23),etc. These examples are all shown with an E–W orientation, but the principle equally applies with N–S,NE–SW, etc. directions. We calculate the semi-variance of the distances for as many different values of has possible using the following formula:

γ ( )

( )

( )

( ) ( )

( )

hz z

n h

i z hi

n h

=− −

=∑ 2

1

2

where γ (h) = semi-variance, n(h) = number of pairs used in the calculation, z = grade, thickness, orwhatever, (i) = the position of one sample in the pair, (i+h) = the position of the other sample in the pair,h meters away from z(i).

Using the above equation above, the γ (h) values calculated for lag 1 (h = 100 m) are as follows:

(1.11 − 1.16)2 + (1.16 − 1.08)2 + (1.08 − 1.03)2 +(1.03 − 1.00)2 + (1.19 − 1.12)2 + (1.12 − 1.03)2 +(1.08 − 1.05)2 + (1.05 − 1.03)2 + (1.03 − 0.97)2 +(0.97 − 1.00)2 + (1.00 − 0.92)2 + (0.92 − 0.94)2 +(1.05 − 1.03)2 + (0.97 − 1.03)2 + (1.03 − 1.00)2 +(1.00 − 0.95)2 = 0.0481(100) = 0.0481/[2 × 16] = 0.0015 (m)2

Using the same formula the γ (h) values for lag 1 to lag 6 (h = 600 m) are as follows:

(100) = 0.0015, (200) = 0.0037, (300) = 0.0036,(400) = 0.0055, (500) = 0.0108, (600) = 0.0104.

The values thus calculated are presented in a graphical form as a semi-variogram. The horizontal axisshows the distance between the pairs, while the vertical axis displays the values of γ (h) (Fig. 10.21).

decreases and the gradient is 0. Beyond thispoint sample values are independent and havevariability equal to the theoretical variance ofsample values (Fluor Mining and Metals Inc.1978, Journel & Huijbregts 1978, Isaaks &Srivastava 1989). This variability is termed thesill (C + C0 on Fig. 10.22) of the semi-variogram,and the point at which this sill value is reachedis termed the range (a) of the semi-variogram.

The range, a, on Fig. 10.22 is 500 m. One cansee that if the drill spacing is greater than500 m then the data from each sample pointwould not show any correlation. They would in

γ (h) = C0 + C

32

12

3

3

ha

ha

( )( )

−⎛⎝⎜

⎞⎠⎟

(for h < a)

γ (h) = C0 + C (for h > a)

The semi-variogram starts at 0 on both axes(Fig. 10.21). At zero separation (h = 0) thereshould be no variance. Even at relatively closespacings there are small differences. Variabilityincreases with separation distance (h). This isseen on the semi-variogram where a rapid rateof change in variability is marked by a steepgradient until a point where the rate of change

Page 254: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 237

1.22 1.11 1.16 1.08 1.03 1.00

1.011.031.121.08

1.08 1.05 1.03

1.03 1.00 0.951.05

1.19

1.03

0.921.000.97

0.97

0.94

100 m0

FIG. 10.18 A portion of a borehole grid with arrows used to identify all the pairs 100 m apart in an E–Wdirection. (From Whateley 1992.)

1.011.031.121.08 1.19

1.03 1.00 0.951.05 1.03 0.97

1.08 1.05 1.03 0.921.000.97 0.94

1.22 1.11 1.16 1.08 1.03 1.00

100 m0

FIG. 10.19 A portion of a borehole grid identifying all the pairs 200 m apart in an E–W direction.(From Whateley 1992.)

effect be random sample points. If, however,one were to drill at a spacing less than 500 m,then the semi-variogram shows us that the datawill, to a greater or lesser extent (depending ontheir distance apart), show correlation. Holes100 m apart will show less variance than holes

400 m apart. Industry standard is to accept asample spacing at two thirds of the range (e.g.350 m of 500 m).

Semi-variograms with a reasonably highnugget variance and where semi-variance alsoincreases rapidly over short distances suggest

Page 255: Introduction to mineral exploration

238 M.K.G. WHATELEY & B. SCOTT

FIG. 10.20 A portion of a borehole grid illustrating some of the pairs 300 m apart in an E–W direction. (FromWhateley 1992.)

that there is a large degree of variability be-tween the samples. When h is greater than orequal to a then the data show no correlationand this gives a random semi-variogram. Thiscan be interpreted to mean that sample spacingis too great.

γ (h)γ

Distance (m)

Sill (C + Co)

C

Co h

Range (a)

1000 200 300 400 500 600

FIG. 10.22 The idealized shape of a semi-variogram(spherical model). (From Whateley 1992.)

It is important to note that the rate ofincrease in variability with distance maydepend on the direction of a vector separatingthe samples. In a coal deposit through which achannel passes, the variability in the coal qual-ity and thickness in a direction normal to the

1.22 1.11 1.16 1.08 1.03 1.00

1.08 1.19 1.12 1.03 1.01

1.08 0.971.05 1.03 1.00 0.92 0.94

1.05 0.97 1.00 0.951.031.03100 m0

FIG. 10.21 A plot of the experimental semi-variogramconstructed from the data derived from the portionof the borehole grid shown in Figs 10.18–10.20. Thecalculations are shown in the text.

Distance (m)

h

0.012

0.010

0.008

0.006

0.004

0.002

00 100 200 300 400 500 600

γ (h)γ

Page 256: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 239

capable of producing a measurement of theexpected error of the estimate (see kriging vari-ance below). The principles of kriging and typesof kriging are discussed by Clark (1979), Journeland Huijbregts (1978), David (1988), Isaaks andSrivastava (1989), Annels (1991), and Deutsch(2002) amongst others. The kriging weights arecalculated such that they minimize estimationvariance by making extensive use of the semi-variogram. The kriging estimator can be con-sidered as unbiased under the constraint thatthe sum of the weights is one. These weightsare then used to estimate the values of thick-ness, elevation, grade, etc. for a series of blocks.

Kriging gives more weight to closer samplesthan more distant ones, it addresses clustering,continuity and anisotropy are considered, thegeometry of the data points and the characterof the variable being estimated are taken intoaccount, and the size of the blocks being esti-mated is also assessed.

Having estimated a reliable set of regulardata values, these values can be contoured ascan the corresponding estimation variances(see below). Areas with relatively high estima-tion variances can be studied to see whetherthere are any data errors, or whether additionaldrilling is required to reduce the estimationvariance.

Cross-validationCross-validation is a geostatistical techniqueused to verify the kriging procedure, to ensurethat the weights being used do minimize theestimation variance. Each of the known pointsis estimated by kriging using between four andtwelve of the closest data points. In this caseboth the known value (Z) and the estimatedvalue (Z*) are known, and the experimentalerror, Z – Z* can be calculated, as well as thetheoretical estimation variance (σ e

2). By plot-ting Z against Z ,* it is possible to test thesemi-variogram model used to undertake thecross-validation procedure. If the values showa high correlation coefficient (R2 near 1), wecan assume that the variables selected from thesemi-variogram provide the best parameters forthe solution of the system of linear equationsused in the kriging matrices to provide theweighting factors for the block estimate andestimation variance calculations (Journel &Huijbregts 1978).

γ (h)γ

Distance (m)

h

Range (a) N–S

Range (a) E–W

1000 200 300 400 500 600

FIG. 10.23 The idealized shape of a semi-variogram(spherical model) illustrating the concept ofanisotropy within a deposit. (From Whateley 1992.)

channel (i.e. across any “zoning”) will increasefaster than the variability parallel to the chan-nel (Fig. 10.23). The different semi-variogramsin the two directions will have different rangesof influence and as such represent anisotropy or“hidden” structure within the deposit.

The squared differences between samplesare prone to fluctuate widely, but theoreticalreasoning and practical experience have shownthat a minimum of 30 evenly distributedsample points are necessary to obtain a reliablesemi-variogram (Whitchurch et al. 1987). Thismeans that in the early stages of an explorationprogram a semi-variogram may be difficult tocalculate. In this case typical semi-variogrammodels from similar deposits can be used as aguide to the sample spacing, until sufficientsamples have been collected from the depositunder investigation. This is analogous to usingthe genetic model to guide exploration.

KrigingKriging is a geostatistical estimation tech-nique used to estimate the value of a point ora block as a linear combination of the avail-able, normally distributed samples in or nearthat block. It considers sample variability inassigning sample weights. This is done withreference to the semi-variogram. Kriging is also

Page 257: Introduction to mineral exploration

240 M.K.G. WHATELEY & B. SCOTT

Estimation variance (σ 2e )

No matter which method is used to estimatethe properties of a deposit from sample data,whether it is manual contouring or computermethods, some errors will always be intro-duced. The advantage of geostatistical methodsover conventional contouring methods is theirability to quantify the potential size of sucherrors. The size of the error depends on factorssuch as the size of the block being estimated(Whitchurch et al. 1987), the continuity of thedata, and the distance of the point (block) beingestimated from the sample. The value to beestimated (Z*) generally differs from its esti-mator (Z) because there is an implicit errorof estimation in Z – Z* (Journel & Huijbregts1978). The expected squared difference be-tween Z and Z* is known as the estimationvariance (σ e

2):

σ e2 = E([Z − Z*]2)

with E = expected value.The estimation variance can be derived from

the semi-variogram:

2γ (h) = E(z(i) − z(i+h))2

By rewriting this equation in terms of thesemi-variogram, as described by Journel andHuijbregts (1978), the estimation variance canthen be calculated.

Knudsen (1988) explained that this calcula-tion takes into account the main factors affect-ing the reliability of an estimate. It measuresthe closeness of the samples to the grid nodebeing estimated. As the samples get fartherfrom the grid node, this term increases, andthe magnitude of the estimation variance willincrease. It also takes into account the size ofthe block being estimated. As the size of theblock increases the estimation variance willdecrease. In addition it measures the spatialrelationship of the samples to each other. Ifthe samples are clustered then the reliabilityof the estimate will decrease. The variabilityof the data also affects the reliability of theestimate, but since the continuity (C) is meas-ured by the semi-variogram this factor isalready taken into account. The estimation

variance is a linear function of the weights(see “Kriging” above) established by kriging(Journel & Huijbregts 1978). Kriging deter-mines the optimal set of weights that minim-ize the variance. These weighting factors arealso used to estimate the value of a point orblock. As it is rare to know the actual value (Z)for a point or block it is not possible to calcu-late Z – Z*(c.f. cross-validation above).

In order to minimize the estimation vari-ance, it is essential that only the nearestsamples to the block being estimated are beused in the estimation process.

Kriging variance (σk2)

Kriging variance is a function of the distanceof the samples used to estimate the blockvalue. A block, which is estimated from a nearset of samples, will have a lower σ k

2 than onethat is estimated using samples some distanceaway. Once the kriging weights for each blockhave been calculated (see “Kriging” above) thevariances can be calculated. Once σ k

2 is deter-mined it is possible to calculate the precisionwith which we know the various properties ofthe deposit we are investigating, by obtainingconfidence limits (σe) for critical parameters.Knudsen and Kim (1978) have shown that theerrors have a normal distribution. This allowsthe 95% confidence limit, ±2(σe), to be calcu-lated (sections 8.3.1 & 10.1):

σ e2 = σe

95% confidence limit = Z* ± 2(σe).

Estimation variance and confidence limit areextremely useful for estimating the reliabilityof a series of block estimates or a set of con-tours (Whateley 1991). An acceptable confid-ence limit is set and areas that fall outside thisparameter are considered unreliable and shouldhave additional sample data collected. Thusthe optimal location of additional holes isdetermined. The confidence limits for criticalparameters such as Au grade or ash contentof coal can be obtained. Areas which haveunacceptable variations may require additionaldrilling, or in a mining situation it may leadto a change in the mine plan or a change in thestorage and blending strategy.

Page 258: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 241

10.4.4 Grade calculation methods

Conventional methods

Often borehole chips or cores, or channels,are sampled on a significantly smaller intervalthan a bench height in an open pit or a stopewidth in an underground mine. It is the geo-logist’s responsibility to calculate a compositegrade or quality value over a predeterminedinterval from the samples taken from the coresor channels. In an open pit the predeterminedinterval is often a bench height. In some in-stances where selective mining is to take place,such as in a coal deposit, the sample compos-ites are calculated for coal and waste separately(see Chapter 13 for a more detailed descriptionof sampling and weighted average quality cal-culations). In an underground vein mine, thecomposite usually represents the minimummining width. Sometimes a thickness × gradeaccumulation is used to calculate a minimummining width (e.g. section 14.6).

The general formula given above for cal-culating the weighted average thickness valuefor the 1/Dn block model, also applies whencalculating weighted average grade or qual-ity values. The weighting function changesthough, and three variables, length, specificgravity (SG), and area are used. For examplewhen compositing several core samples for aparticular bench in a proposed open pit, lengthof sample is used. This is particularly import-ant when the geologist sampling the core hasdone so using geological criteria to choosesample lengths and samples are of differentlengths (e.g. Fig. 10.12). Thus:

G

=

∑ ×∑L G

L

where G = weighted average grade of eachborehole on that bench, G = grade of each coresample, L = length of sample.

When samples are to be composited from,say, core, and the rocks are of significantlydifferent densities, then the weighting factorshould have bulk density (BD) included.Thus:

G

=∑ × ×

∑ ×BD L G

BD L

Area of influence can also be used as part ofthe weighting function. For example, to calcu-late the weighted average grade of a stratiformdeposit which is to be assessed by polygons ofdifferent sizes, the area of each polygon is usedas the weighting function. Thus:

G

=

∑ ×∑A G

A

where A = area of influence of each polygon.

Computational methods

In large ore bodies, or deposits which have alarge amount of data, such as Trinity SilverMine (see Chapter 16), computers are usedto manipulate the data. Commercial soft-ware packages are available (e.g. MineSight,Datamine, see section 9.3) which can per-form all the conventional methods describedabove. Computers offer speed and repeatabil-ity, as well as the chance to vary parametersto undertake sensitivity analyses. In additioncomputers can use exactly the same proceduresas described for geostatistical mineral resourceestimation methods. As well as using thick-ness to calculate volume and (with SG) tonn-age, grade or quality can be estimated alongwith the estimation error.

10.4.5 Software simulation

Geostatistical tools may be used for spatialdata analysis, to model the heterogeneity ofmineral deposits, to assess the uncertainty(risk) in the estimation process, and to use thisinformation in the decision process (Deutsch2002). The fundamental problem explorationgeologists have is to create a model fromlimited sample data. Methods to mitigatethis, such as manual contouring or kriging, aredescribed above. Once a project appears tohave economic potential, engineers becomeinvolved in the evaluation process. They maywish to predict grade changes as the deposit is“mined” in a manner that mimics the proposedmining method. This can be achieved usingconditional simulation.

Normally, sample points will support threeestimated (kriged) points. Thus, on a 250 m

Page 259: Introduction to mineral exploration

242 M.K.G. WHATELEY & B. SCOTT

drilling grid, block size would be 80 m.Mineralized rock with a density of 2.6 t m−3 inan 80 × 80 m block with a 15 m bench willprovide 250 kt of material. This is usually toocoarse to provide an estimate of local variab-ility. This is a situation where conditionalsimulation can be used because it allows theestimator to generate values on a very fine grid.All calculations are performed using gaussiantransformed (normalized) data (Deutsch 2002).The properties of the resultant simulationsare such that they replicate the statistical andgeostatistical properties of the input data, aswell as honoring the input data. The simulatedvalues are controlled, or conditioned, by drill-hole samples, which gives rise to the nameof this technique (Deutsch & Journel 1998,Deutsch 2002).

In coal deposits, some coal quality variables,such as ash and calorific value, have largenumbers of data points. There are often fewer

sulfur or trace element data points (Whateley2002). This creates estimation problems anddifficulties in estimating life of mine (LOM)coal quality variability. This lack of data forsome of the layers within a model presentsproblems to the estimator for the applicationof the method. This is particularly true forsulfur values within some plies when lessthan 30 samples of hard data are availablefor each layer. Adapting the conditional simu-lation technique, by using the more abund-ant, but less reliable whole seam data (softdata) to augment the ply data, can mitigatethis problem. Conditional simulation modelsthus generated still need to be carefullyexamined and verified. Once they are foundto satisfactorily replicate the geological, stat-istical (Fig. 10.24), and geostatistical charac-teristics of raw data, the simulated qualityvariability can be used for predicting LOM coalfeed (Fig. 10.25).

Rel

ativ

e fr

eque

ncy

(%)

20

18

16

14

12

10

8

6

4

2

0

Ply data

Conditionalsimulation

0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2

Total sulfur (db %)

FIG. 10.24 Compound frequency distribution of raw data and simulated sulfur data. The simulated data arefound to satisfactorily replicate the geological, statistical, and geostatistical characteristics of raw data.

Page 260: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 243

FIG. 10.25 Plot of the simulated sulfur variable, used for predicting life-of-mine (LOM) coal feed.

From 100 simulations the probability dis-tribution of a grade variable can be derived.Prediction of grade variability is then possible.

10.5 GEOTECHNICAL DATA

During the sampling and evaluation of a min-eral deposit, the exploration geologist is inan ideal position to initiate the geotechnicalinvestigation of that deposit. Usually, geotech-nical and mining engineers are invited on tothe investigating team after the initial drillingis complete. They will need to know some ofthe properties of the rocks, such as strength,degree of weathering, and the nature of thediscontinuities within the rock, in order tobegin designing the rock-related componentsof the engineering scheme (haulages, under-ground chambers, pit slopes). It is important toinvestigate some of the geotechnical propertiesof the overburden and of the rocks in which themineralisation is found during the exploration

phase of a project. The engineers must decidewhether it is feasible to mine the potential ore,as this information is required for the financialanalyses which may follow (see Chapter 11).

Geotechnical investigations essentiallycover three aspects: soil, rock, and subsurfacewater. A good field guide for the description ofthese is given by West (1991). Soil refers to theuncemented material at or near the surface,and can include weathered or broken rock. Soilis mainly investigated for civil engineering pur-poses such as foundations of buildings. Rock isthe solid material which is excavated duringmining to reach and extract the ore. Subsurfacewater is part of the soil and the rock, and isfound in the pore spaces and in joints, fracturesand fissures.

There are two forms in which the data can becollected: (i) mainly subjective data obtainedby logging core and the sides of excavations(Table 10.8), and (ii) quantitative data obtainedby measuring rock and joint strengths and alsovirgin rock stress, etc., either in the laboratory

Page 261: Introduction to mineral exploration

244 M.K.G. WHATELEY & B. SCOTT

TABLE 10.8 A summary of the description and classification of rock masses for engineering purposes. (Adaptedfrom Engineering Group Working Party 1977, Brown 1981.)

Descriptive indices of rock materialConsideration should be given to rock type, color, grain size, texture and fabric, weathering (condition),alteration, and strength

Rock condition (weathering)Fresh No visible signs of rock material weatheringSlightly weathered Discoloration indicates weathering of rock material and discontinuity surfacesWeathered Half the rock material is decomposed or disaggregated. All rock material is

discolored or stainedHighly weathered More than half the rock material is decomposed or disaggregated. Discolored rock

material is present as a discontinuous framework. Shows severe loss of strengthand can be excavated with a hammer

Completely weathered Most rock material is decomposed or disaggregrated to a soil with only fragmentsof rock remaining

Rock material strengthExtremely strong Rock can only be chipped with geological hammer. Approximate unconfinedcompressive strength >250 MPaVery strong Very hard rock which breaks with difficulty. Rock rings under hammer.

Approximate unconfined compressive strenth 100–250 MPaStrong Hard rock. Hand specimen breaks with firm blows of a hammer, 50–100 MPaMedium Specimen cannot be scraped or peeled with a strong pocket knife, but can be

fractured with a single firm blow of a hammer, 25–50 MPaWeak Soft rock which can be peeled by a pocket knife with difficulty, and indentations

made up to 5 mm with sharp end of hammer, 5–25 MPaVery weak Very soft rock. Rock is friable, can be peeled with a knife and can be broken by

finger pressure, 1–5.0 Mpa

Descriptive indices of discontinuitiesConsideration should be given to type, number of discontinuities, location and orientation, frequency ofspacing between discontinuities, separation or aperture of discontinuity surfaces, persistence and extent,infilling, and the nature of the surfaces

Rock quality designationSound RQD >100 Rock material has no joints or cracks. Size range >100 cmFissured RQD 90–100 Rock material has random joints. Cores break along these joints. Lightly broken.

Size range 30–100 cmJointed RQD 50–90 Rock material consists of intact rock fragments separated from each other by

joints. Broken. Size range 10–30 cmFractured RQD 1–50 Rock fragments separated by very close joints. Core lengths less than twice NX

(Table 10.7) core diameter. Very broken. Size range 2.5–10 cmShattered RQD <1 Rock material is of gravel size or smaller. Extremely broken. Size range less than

2.5 cm

RQD in per cent Length of core in pieces 10 cm and longer

Length of core run= × 100

RQD diagnostic description>90% Excellent75–90 Good50–75 Fair25–50 Poor<25% Very poor

Page 262: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 245

or in situ. Subjective data helps with thedecision making fairly early in a project, butthese data will always need quantification. Itis therefore necessary to study and record thedirection and properties of joints, cleavage,cleats, bedding fissures, and faults, and to studyand record the mechanical properties (mechan-ics of deformation and fracture under load),petrology, and fabric of the rock between thediscontinuities.

It is clear from the above that structuralinformation is extremely important in mineplanning. In the initial phases of a programmost of the structural data come from core.The geologist can log fracture spacing, attitude,and fracture infill at the rig site. A subjectivedescription of the quality of the core can begiven using Rock Quality Designation (RQD)(Table 10.8). Pieces of core greater than 10 cmare measured and their length summed. This isdivided by the total length of the core run andexpressed as a percentage. A low percentagemeans a poor rock while a high percentagemeans a good quality rock.

Bieniawski (1976) developed a geomechan-ical classification scheme using five criteria:strength of the intact rock, RQD, joint spacing,conditions of the joints, and ground water con-ditions. He assigned a rating value to each ofthese variables, and by summing the valuesof the ratings determined for the individualproperties he obtained an overall rock massrating (RMR). Determination of the RMR of anunsupported excavation, for example, can beused to determine the stand up time of thatexcavation.

In order to provide information that willassist the geotechnical and mining engineers itis necessary to understand the stress–straincharacteristics of the rocks. This is especiallyimportant in relation to the mining methodthat is proposed for the rocks, the way in whichthe rocks will respond in the long term to sup-port (in underground mining), or in deep openpits how deformation characteristics affect thestability of a slope. As these rock propertiesbecome better understood, it is possible for thegeotechnical and mining engineers to improvethe design of underground mines, the way inwhich excavations are supported, the effici-ency of mining machinery, and the safety of themine environment.

10.5.1 Quantitative assessment of rock

There are numerous good textbooks whichcover the properties of rocks from an engineer-ing point of view (Krynine & Judd 1957,Goodman 1976, 1989, Farmer 1983, Brady &Brown 1985), the latter reference being particu-larly relevant to underground mining. Themain properties which are used in the engineer-ing classification of rocks for their quantitativeassessment are porosity, permeability, specificgravity (bulk density), durability and slakabil-ity, sonic velocity, and rock strength.

Porosity and permeability are important inthe assessment of subsurface water. Specificgravity is required to determine the mass ofthe rock. Durability and slakability tests revealwhat affect alternate wetting and drying willhave on surface or near-surface exposures.Under conditions of seasonal humidity somerocks have been known to disaggregate com-pletely (Obert & Duval 1967). Sonic velocitytests give an indication of the fracture intensityof the rock. Probably the most important prop-erty of rock is its strength.

Rock strength

The fundamental parameters that define rockstrength are stress and strain, and the rela-tionship between stress and strain. Stress (σ)is a force acting on a unit area. It may be hydro-static when the force is equal in all direc-tions, compressional when the force is directedtowards a plane, tensional when directed awayfrom a plane, torsional (twisting), or shearstress (τ) when the forces are directed towardseach other but not necessarily in line. Strain(ε) is the response of a material to stress by pro-ducing a deformation, i.e. a change in shape,length, or volume. The linear relationshipbetween stress and strain is known as Hooke’slaw and the constant (E) connecting them isYoung’s modulus. It is also called the modulusof elasticity, and gives a measure of stiffness.

The strength of a rock sample can be tested(usually in the laboratory) in several ways.Unconfined (uniaxial) compressive strengthis measured when a cylinder (usually core) ofrock is loaded to a point of failure. The load im-mediately prior to failure indicates the rock’sstrength. Tensile strength is measured by

Page 263: Introduction to mineral exploration

246 M.K.G. WHATELEY & B. SCOTT

“pulling” a cylinder of rock apart. The tensilestrength of rock is generally less than one tenthof its compressive strength. Shear strengthis measured as a component of failure in atriaxial compressive shear test. A cylinder ofrock is placed inside a jacketed cylinder in afluid-filled chamber. Lateral as well as verticalpressure is applied and the rock is tested todestruction (Fig. 10.26). The strength is thatmeasured immediately prior to failure.

Failure

Observations on rocks in the field, in mines,and in open pits indicate that rocks which onsome occasions appear strong and brittle may,under other circumstances, display plasticflow. Rocks which become highly loaded mayyield to the point of fracture and collapse. Arock is said to be permanently strained when ithas been deformed beyond its elastic limit.

With increasing load in a uniaxial (uncon-fined) compressive test, the slope of a stress–strain curve displays a linear elastic response(Fig. 10.27). Microfracturing within the rockoccurs before the rock breaks, at which point(the yield point) the stress–strain curve starts

σ2=σ3

Confiningpressure

Steelcylinder

Fluid filledchamber

Seal

Borehole core

Flexible membrane

Pressure from pump

Axial stress

σ2=σ3

σn

τ

θ

Shear fracture

Confiningpressure

σ2=σ3

Loadσ1(a) (b)σ

σ σ σ σ

σ

σσ

σ

σ1

2

FIG. 10.26 (a) Schematic diagram of the elements of triaxial test equipment. (b) Diagram illustrating shearfailure on a plane, an example of the results of the triaxial test on a piece of core. σ1 is the major principalcompressive stress, σ2 = σ3 = intermediate and minor principal stresses respectively, σn is the normal stressperpendicular to the fracture, and τ is the shear stress parallel to the fracture.

to show inelastic behavior followed by a rapidloss in load with increasing strain. If the loadis released before the yield point is reached, therock will return to its original form. Brittlerocks will shear and ductile rocks will flow,both types of behavior resulting in irreversiblechanges.

The results of the uniaxial (unconfined) com-pressive tests described above give a measureof rock strength and are convenient to make.The conditions that the rock was in prior tobeing removed by drilling (or mining) weresomewhat different, because the rock had beensurrounded (confined) by the rest of the rockmass. When a rock is confined its compressivestrength increases. The triaxial compressivestrength test gives a better measure of the effectsof confining pressure at depth (Fig. 10.26). Corethat has been subjected to triaxial testing,fractures in a characteristic way (Fig. 10.26b).Several failure criteria have been developedfor rocks, the simplest being represented byCoulomb’s shear strength criterion that can bedeveloped on a fracture plane such as the oneshown in Fig. 10.26b and is represented by:

τ = σn tan θ + c

Page 264: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 247

FIG. 10.27 A simplified stress–strain diagramillustrating the idealized deformational behaviorof some rock types. That of granite represents anidealized perfectly elastic material. Clay is a typicalmaterial that behaves as a plastic material whichwill not deform until a certain stress is achieved.Marble is an example of an elastoplastic materialand shale represents a ductile material where stressis not proportional to strain. (Modified after Peters1987.)

Str

ess

σ

Yield point

Grani

te

Marble

Shale

Clay

Strain ε

σ

ε

τ

θ

3''

Failure criterion

σσσσ3' 1' 1''

FIG. 10.28 An example of a Mohr diagram.

where τ = total shearing resistance, σn = normalstress acting at right angles to the failure plane,c = cohesive strength, tan θ = coefficient of in-ternal friction where θ = angle between stressfracture induced and σ1. The theory behindthis equation is explained by Farmer (1983) andBrady and Brown (1985).

Triaxial test results are used to constructa Mohr diagram (Fig. 10.28). The symbol σ3 rep-resents lateral confining pressures in succes-sive tests, while σ1

represents the axial loadrequired to break the rock. Circles are drawnusing the distances between the differentvalues of σ3 and σ1 as the diameter. The com-mon tangent to these circles is known as theCoulomb strength envelope. The angle θ canbe determined from the diagram, rather thanhaving to measure the fracture angle on the

tested rock specimen. A value for c can also bedetermined from the diagram.

The expected confining pressures and axialloads that might be encountered in a mine canbe calculated. Any sample whose test resultsplot as a Mohr diagram that falls below the tan-gent should not fail. A sample which failed in atest with lower confining stresses but similaraxial stresses, would give rise to a circle whichwould cut the tangent suggesting that the crit-ical stresses have been exceeded and failuremight occur (Peters 1987). This does not takeinto account failure that might take place alongdiscontinuities, nor does it consider the effectsof anisotropic stress systems in the rock massor changes that will occur with time.

The effect of discontinuities is probably moreimportant than being able to measure the rockstrength. The discontinuities may be weak andcould result in failure at strengths below that ofthe rock itself.

In situ stress determination

Most in situ stress determinations are made be-fore mining takes place usually from a boreholedrilled from the surface or from an adit. Twocommon procedures are employed in the deter-mination of in situ stress. The first is based onthe measurement of deformation on the bore-hole wall induced by overcoring (Fig. 10.29) andthe second by measuring the component ofpressure in a borehole or slot needed to balancethe in situ stress. In the overcoring system,a strain gauge is fixed to the borehole wall

Page 265: Introduction to mineral exploration

248 M.K.G. WHATELEY & B. SCOTT

using a suitable epoxy resin. Overcoring with acore barrel of larger size causes stress relief inthe vicinity of the strain cell which induceschanges on the strains registered in the gauges(Brady & Brown 1985). These measurementsare used to help in mine design.

Additional tests are carried out during min-ing using various types of strain gauges. Somemines, e.g. Kidd Creek, have their own rockmechanics, backfill, and mine research depart-ments whose rock mechanics instruments areregarded as an indispensable tool for detectingand predicting ground instability (Thiann 1983,Hannington & Barrie 1999). Hoek and Brown(1980) and Hoek and Bray (1977) give excellentexamples of the use of rock mechanics in bothunderground and surface mining respectively.

10.5.2 Hydrogeology

During an exploration program it is importantto note where the water table is encountered ina borehole because if one were to mine thisdeposit later, then it is essential to know wherethe water is, how much water there is, and thewater pressure throughout the proposed mine.Useful tests that can be performed on boreholesinclude measuring the draw down of the waterlevel by pumping, and upon ceasing pumpingmeasuring the rate of recovery. Packer testsin boreholes are used to obtain a quantitativeestimate of the contribution a particular bedor joint may make to the water inflow at a site(Price 1985). A good description of the basicfieldwork necessary to understand the hydro-geology of an area is given by Brassington(1988). Two properties of the rocks that can bemeasured are the porosity and the permeabil-ity. Porosity describes how much water a rockcan hold in the voids between grains and injoints, etc. Permeability is a measure of howeasily that water can flow through and out ofthe rock. The flow of water is described empir-ically by Darcy’s law, and the relative ease ofthe flow is called the hydraulic conductivity.Darcy’s law is a simple formula:

v = Ki

where v = the specific discharge or the velo-city of laminar flow of water through a porousmedium (the critical Reynold’s number at

End ofborehole

Strain cell

Drilling rig

Borehole

Potentialmineabledeposit

Overcoring

Strain cell

FIG. 10.29 A diagrammatic representation of theprinciple of overcoring.

Page 266: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 249

Original water table

Cone of depression

Shaft

Ore bodyGround surface

Dewatering borehole

Intersecting conesof depression Flow line

Point ofdischarge

Equipotential line

Piezometer installedto measure the

water table level

Dra

wdo

wn

Piezometric surface

FIG. 10.30 Diagram to illustrate dewatering cones and a piezometric surface. The shaft and dewateringborehole are sited close enough to each other for their cones of depression to intersect and produce a loweringof the water table which effectively results in the dewatering of part of the mine. The flow net illustrates theflow of underground water through rocks of uniform permeability. Flow takes place towards the points ofdischarge, which in this case are pumps at the bottom of the shaft and dewatering boreholes. (Modified afterMcLean & Gribble 1985.)

which turbulence development is seldomreached), K = hydraulic conductivity, i =hydraulic gradient (the rate at which the headof water changes laterally with distance).

The hydraulic conductivity can be measuredin boreholes by undertaking pumping tests.The head of water at a point (e.g. a borehole)within a rock mass is the level to which thewater rises when the borehole is drilled. Thetube inserted in a borehole to measure thiswater level is called a piezometer. Severalboreholes may intersect points on a piezo-metric surface (Fig. 10.30), and the maximumtilt (i) of this surface from the horizontaldefines the hydraulic gradient (McLean &Gribble 1985).

Groundwater moves slowly under the influ-ence of gravity from areas of high water tableto low areas (Fig. 10.30). The pumps lower thepiezometric surface in the proximity of theborehole creating a difference in pressure, withthe pressure much less at the point of discharge(the pumps) than in the areas between thepumps. This results in the water flowing incurved paths towards the pumps.

Rocks which are permeable and allow waterto flow freely through the joints, fractures, andpores are known as aquifers. Rocks which havea low permeability and are impervious and donot allow water to flow through them areknown as aquicludes.

Having found an orebody, finding waterfor the processing plant is often the next con-cern. A plant requires a considerable volumeof water and it can happen that the amount ofwater available may make or break a project.

10.6 SUMMARY

Mineralisation is composed of dissimilar con-stituents, is heterogenous, and much mineral-isation data may have an asymmetrical ratherthan a normal distribution. Simple statist-ical sampling methods used for investigatinghomogenous populations may therefore givea highly misleading result if used to interpretmineralisation data where in many cases asample can represent only a minuscule propor-tion of the whole. The pitfalls and problems

Page 267: Introduction to mineral exploration

250 M.K.G. WHATELEY & B. SCOTT

that this may cause have been discussed andemphasis placed on the need to pay carefulattention to sampling error.

Sample extraction must be random andpreferably of equal volume, but as the samplesmay amount to several kilograms or eventonnes in weight there must be a systematicreduction of mass and grain size before ana-lysis can take place. Sample reducing systemsrequire thoughtful design and must be care-fully cleaned between each operation. Gold-and platinum-bearing samples present specialproblems because their high density may leadto their partial segregation from lighter ganguematerial during sample preparation.

Samples may be acquired by pitting andtrenching and by various types of drilling. Ifcoring methods are used then effective corerecovery is essential and the core logging mustbe most methodically performed. Drilling con-tracts should be carefully thought out andbased on sound knowledge of the drilling con-ditions in the area concerned. Drillholes devi-ate from a straight line and must be surveyed.Drilling patterns must be carefully planned andthe geologist in charge ready to modify them asresults come to hand.

The analytical data derived from the aboveinvestigations are used in mineral resourceand ore reserve estimations, and for grade cal-culations. Classical statistical methods haveonly a limited application and geostatisticalmethods are normally used. The explorationsample collection work provides a great oppor-tunity to initiate geotechnical investigations.The collection of geotechnical data will pro-vide mining engineers with essential informa-tion for the design of open pit and undergroundworkings. It will also provide a starting pointfrom which the hydrogeology of the target areacan be elucidated.

10.7 FURTHER READING

Two textbooks which cover statistical analysisof geological data are the eminently readableDavis (2003) and Swan and Sandilands (1995),while Isaaks and Srivastava (1989) give a veryreadable account of geostatistics. The mathem-atical details of geostatistical estimations areelaborated by Journel and Huijbregts (1978).

Annels (1991) covers this subject as well asdrilling, sampling, and resource estimation.Discussions of the problems of resource defini-tion are covered by the Australasian Joint OreReserves Committee (the JORC code: JORC2003), Annels (1991), and Whateley and Harvey(1994). The JORC website hosts a number oftechnical papers that explain the JORC code,its history and implementation. A very goodsource for case histories is the volume ofEdwards (2001).

An introductory text for geotechnical inves-tigations is written by West (1991), whileHoek and Brown (1980) and Hoek and Bray(1977) are still the most sought after books thatdescribe the application of rock mechanics tounderground and surface mines respectively.Brassington (1988) gives a good description ofthe basic fieldwork necessary to understandthe hydrogeology of an area.

10.8 APPENDIX: USE OF GY’S SAMPLINGFORMULA

1 IntroductionThese are worked examples of the calcula-tion of the total variance (TE) of sphaleriteand chalcopyrite mineralisation and the two-sided confidence of that mineralisation. Theseare followed by a worked example of how tocalculate the mass of a sample to be takenknowing the confidence limits.

Using the short form of the Gy formula forthe fundamental variance (section 10.1.4):

So2 (FE) = C/K

Taking K as 125,000:

So2 (FE) = C/125,000

and the variance of the total error (TE) is esti-mated as twice that of the fundamental error(FE). It is assumed that the liberation factor βis constant and equals unity, which is a safeassumption.

2 Calculation of total variance (TE) anddouble-sided confidence interval: sphalerite.A prospect contains sphalerite mineralisationof mean grade 7.0% zinc in a gangue of density2.7 kg m−3.

Page 268: Introduction to mineral exploration

10: EVALUATION TECHNIQUES 251

Sphalerite contains 67% zinc with a specificgravity of 4.0 kg m−3.(a) Fundamental variance of each samplingstage (see also Box 10.1). Critical content (aL) = 7% zinc = 10.5% sphalerite as a fraction= 0.105.

Constitution factor (c):

=

−− +

( . ).

[( . ) . ( . ) . ]1 0 105

0 1051 0 105 4 0 0 105 2 7

= 8.52(3.58 + 0.28) = 32.89, say 33.

Heterogeneity constant (C)

= cβfg = 33 × 1 × 0.5 × 0.5 = 4.13.

Fundamental variance:

=

CK

= 24.38/125,000 = 33 × 10−6

(b) Total variance (TE) of the sampling scheme.At each reduction stage the total variance isequal to double the fundamental variance. Thecomplete sampling scheme consists of fourreduction stages (Fig. 10.5a) each ending on thesafety line (see text). Consequently, the totalvariance equals eight times (2 × 4) the funda-mental variance as calculated above.

Total variance (TE) = 8 × (33 × 10−6) = 264 × 10−6.

Relative standard deviation = ( )264 10 6× −

= 16.2 × 10−6

Absolute standard deviation as % of mineral =10.5% sphalerite × 1.6% = 0.17% sphalerite.Absolute standard deviation as % of metal =0.17% sphalerite × 67% = 0.11% zinc.The confidence level at 95% probability is ±2SD = ±2(0.11%) = 0.22% zinc.The two-sided confidence level is then 7%±±±±± 0.22 zinc at 95% probability, which is anacceptable result. Generally, a relative stand-ard deviation of <5% is acceptable.

3 Calculation of total variance (TE) anddouble-sided confidence interval: chalcopyritemineralisation.A prospect contains chalcopyrite mineralisa-tion of grade 0.7% copper (X) in gangue ofdensity 2.7 kg m−3.

Chalcopyrite contains 33% copper with a den-sity of 4.2 kg m−3.(a) Fundamental variance of each samplingstage.Critical content (aL) = 0.7% = 2.1% chalcopyriteas a fraction = 0.021.

Constitution factor (c)

=

..

1 0 0210 021

[(1 − 0.105)4.0 + (0.105)2.7]

= 46.62(4.11 + 0.06) = 194.4, say 195.

Heterogeneity constant (C) = cβfg = 195 × 1 ×0.5 × 0.25 = 24.38.

Fundamental variance:

=

CK

= 24.38/125,000 = 195 × 10−3.

(b) Total variance (TE) of the sampling scheme.As above but, suppose, the system consists offive reduction stages each ending on the safetyline. The total error variance now equals 10times (2 × 5) the fundamental variance.

Total variance (TE) = 10 × (195 × 10−6)= 1950 × 10−6

Relative standard deviation = (1950 × 10−6)−2 =44 × 10−3 = 4.4%.Absolute standard deviation as % ofmineral = 2.1% chalcopyrite × 4.4% = 0.09%chalcopyrite.Absolute standard deviation as % of metal =0.09% × 33% = 0.03% copper.The confidence at 95% probability is ±2 SD =±2(0.03%) = 0.06% copper.The two-sided confidence level is then 0.7% ±±±±±0.06% copper at 95% probability, which is anacceptable result. Generally a relative standarddeviation of <5% is acceptable.

4 Calculation of the weight of a sample (M) tobe taken knowing the confidence interval.(a) Introduction (see section 10.1.4)

M = Cd3/S2(FE)

(b) Sphalerite mineralisation.Sphalerite mineralisation assays 7% zinc andthe confidence level required is ± 0.2% zinc at

Page 269: Introduction to mineral exploration

252 M.K.G. WHATELEY & B. SCOTT

95% probability. The sphalerite is liberatedfrom its gangue at a particle size of 150 µm: thegangue density is 2.7 kg m−3. Sampling is under-taken during crushing when the top particlesize (d) is 25 mm.Critical content (at) = 7% zinc = 10.5% sphal-erite = 0.105 as a fraction.Constitution factor (c) as in previous example= 33.Liberation factor (β) = (0.015/2.5)0.5 =(0.006)0.5 = 0.077.Heterogeneity constant (C) = cβfg = 33 × 0.077 ×0.5 × 0.25 = 0.32.Top size (d) = (2.5)3 =15.62.Cd3 = 0.32 × 15.62 = 5.0.Fundamental variance (FE).

Relative standard deviation 2S =

0 27

. %. There-

fore S = 0.014.

Mass of sample =

CdS FE

3

2( )=

5 00 014 2

.( . )

= 25.5 kg.

This is the sample weight to constrain thefundamental variance. For the total variancea sample weight of twice this quantity isrequired [i.e. S2(TE) ≤ 2S2(FE)], or 51 kg.

To illustrate the benefits of samplingmaterial when it is in its most finely dividedstate, assume that the same limit (±0.2% zinc)is required when sampling aftergrinding toits liberation size when d is 0.015 cm and βequals 1.0:

C = 33 × 1 × 0.5 × 0.25 = 4.13

d3 = (0.015)3 = 3.3 × 10−6

Cd3 = 13.63 × 10−6 and

M ===== 6.9 grams, say 7 g

Page 270: Introduction to mineral exploration

11Project Evaluation

Barry C. Scott and Michael K.G. Whateley

11.1 INTRODUCTION

It is said that mineralisation is found, orebodiesare defined, and mines are made. Previouschapters have outlined how mineralized rockis located by a correct application of geology,geophysics, and geochemistry. Having locatedsuch mineralisation its definition follows andthis leads to a calculation of mineral resourceswhere a grade and tonnage is demonstrated atan appropriate level of accuracy and precision(see Chapters 3 and 10). The question now be-comes how does a volume of mineralized rockbecome designated as a resource and by whatprocess is this resource selected to become anore reserve (see section 10.4) for a producingmine? What steps need to be taken to ensurethat there is a logical progression that movesthe project with the initial unconnected ex-ploration data to a final reserve estimate thatmeets the requirements of potential investorsand bankers? A mine will come into existenceif it produces and sells something of value.What value does a certain tonnage of mineral-ized rock have? What is meant by value in thiscontext? How is this value calculated andevaluated? These questions are considered inthis chapter.

11.2 VALUE OF MINERALISATION

Legislation in a number of countries now re-quires directors of listed companies to complywith a corporate governance code. In the UK,the Turnbull report (Turnbull 1999) recom-mends a systematic approach to the identifica-

tion, evaluation, and management of signific-ant risks to a company. In a mining companythis requirement includes technical issues suchas determination of the ore reserves, selectionof appropriate mining and processing methods,as well as financial, social, environmental, andreclamation aspects of the project. Similarlythere is a range of business issues to be add-ressed, such as the project’s ownership struc-ture, permitting, marketing, and governmentrelations and their interactions. The valueof mineralisation is a function of these aspectsand several other factors, which are listedbelow.

11.2.1 Mine life and production rate

Mine life

Other things being equal the greater the ton-nage of mineable mineralisation the longer themine life and the greater its value. It followsthat:

Mine life in years

total ore reserveaverage annual production

=

Factors relating to mine life include thefollowing:

Legal limitationsExploitation of the mineralisation may beunder the constraints of a lease granted by itsformer owner such as the government of thestate concerned. If the lease has an expiry datethen the probability of obtaining an extensionhas to be seriously considered.

Page 271: Introduction to mineral exploration

254 B. SCOTT & M.K.G. WHATELEY

Market forecastWhat will be the future demand and value forthe minerals produced? This important factoris considered separately below.

Political forecastIf a lease agreement has been negotiated withthe local government how long is this govern-ment likely to remain in power? Will a newgovernment wish to either cancel or modifythis agreement? (See also section 1.5.4.)

Financial returnThe development of a mine is a form of capitalinvestment. The main concern is for this in-vestment to produce a suitable level of return.The optimization of this financial return willprobably decide the number of years of produc-tion. As a very general rule the minimum life ofa mine should be 7–10 years.

Production rate

A commonly used guide for the AverageAnnual Production Rate (AAPR) is:

AAPR

(ore reserve)

.

.

=0 75

6 5

This is a broad generalization but for ore re-serves of 20 Mt, the AAPR is 1.5 Mt a−1, and for10 Mt it is 0.9 Mt a−1. Factors relevant to theannual production rate include the following:

Type of miningThe two basic types of mining are either under-ground (Fig. 11.1) or from the surface (open pit),(Fig. 11.2). As production from the former canbe more than three times the cost of the latter,underground extraction may require at leastthree times the content of valuable componentsin the mineralisation to meet this higher cost.This in turn implies a lower tonnage, but ahigher grade of ore reserve and a lower annualproduction rate.

A basic characteristic of open pit mining isthe surface removal of large tonnages of wasterock (overburden) in order to expose the ore forextraction (Fig. 11.2). This removal of wasterock and ore usually proceeds simultaneouslyand the resultant stripping ratio of tonnageof waste removed to tonnage of ore is a funda-

mental economic specification. With a strip-ping ratio of 4:1, 4 t of waste has to be removedfor every 1 t of ore recovered; if the miningcosts were $US 2 t−1 then the extraction costper tonne of ore is $US10 t−1. Conversely, if,say, the stripping ratio were 8:1 at the samemining cost then the extraction cost per tonneof ore would be $US 18 t−1.1 Open pit mining costs vary from about $US1to $US3 t−1 of rock extracted. An economy ofscale is apparent in that the higher charge is forproduction levels of lower than 10,000 tonnesper day (tpd), while the lower cost refers to out-puts exceeding 40,000 tpd of ore plus wasterock.2 Underground mining is generally on asmaller scale than open pit mining. An outputof 10,000 tpd would be modest for an open pitbut large for an underground mine. This gener-ally lower level of production is in part relatedto the more selective nature of undergroundextraction in that large tonnages of waste rockdo not have to be removed in order to exposeand extract ore. This is exemplified in under-ground block caving. Northparkes Cu/Au minein Australia is one of the most productiveunderground mines in the world, producing5 Mt of ore per year (Chadwick 2002). Blockcaving is designed as a low maintenance work-ing environment with minimal operating costsand high productivities.3 Underground mining costs vary greatly de-pending upon the mining method employed.In large orebodies where production is mech-anized, costs per tonne of ore extracted can becomparable with those of an open pit. At theother end of the scale, in narrow veins usinglabor-intensive methods, costs can be as highas $US30–40 t−1.4 Generally underground mining costs arehigher than those at the surface. However, asthe depth of an open pit increases, the ratioof waste to ore mined becomes greater andmore waste rock per tonne of ore is extracted(Fig. 11.2). This has the effect of increasing theoverall extraction cost until eventually it canequal underground mining costs.

Scale of operationsThe economies of scale mean that usually thelarger the annual output the lower the produc-tion cost per tonne of mineralized rock mined.

Page 272: Introduction to mineral exploration

11: PROJECT EVALUATION 255

FIG. 11.1 Mining terminology. Ore was first extracted at the mineralized outcrop using an open pit. Later anadit was driven into the hillside to intersect the ore at a lower level. An inclined shaft was sunk to mine fromstill deeper levels and, eventually, a vertical shaft was developed. Ore is developed underground by drivingmain haulage and access drifts at various levels and connecting them by raises from which sublevels aredeveloped. It is mined from the lower sublevel by extracting the mineralisation upwards to form a stope.Broken ore can be left in a stope to form a working platform for the miners and to support the walls againstcollapse (shrinkage stoping). Alternatively, it can be withdrawn and the stope left as a void (open stoping) orthe resultant space can be filled with fine-grained waste rock material brought from surface (cut-and-fillstoping). Ore between the main haulage and sublevel is left as a remnant to support the former access untilthe level is abandoned. (From Barnes, 1988, Ores and Minerals, Open University Press, with permission.)

FIG. 11.2 Development of an open pit mine. Duringthe early stages of development (a–a′) more ore isremoved than waste rock and the waste to ore ratiois 0.7:1. That is, for every tonne of ore removed 0.7tonnes of waste has also to be taken. The waste rockis extracted in order to maintain a stable slope onthe sides of the open pit, normally between 30 and45 degrees. Obviously the greater the slope the lessthe waste rock to be taken, and the lower the wasteto ore ratio. As the pit becomes deeper the ratio willincrease; at stage b–b′ it is 1.6:1. (From Barnes 1988,Ores and Minerals, Open University Press, withpermission.)

b a a' b'

Waste dump

Aditportal Adit portal

Abandonedopen pit

Inclineshaft

Verticalshaft

Shaft pillar

ShaftstationwithsumpbelowRaise

Loadingstationfrom orebin

Developmentraise

Main haulage level

Winze

Broken ore

Stope

Unbrokenore

Developmentraise

Haulage drift or level

Sub-levelOrebin

Unminedore

Orestoped

out andspace

filledwith

waste

Page 273: Introduction to mineral exploration

256 B. SCOTT & M.K.G. WHATELEY

Obviously there is an upper limit to this gen-eral principle otherwise most mines would beworked out in their first year of production!

Accessibility of the mineralisationHow accessible is the mineralisation? Thereare limitations on open pit and shaft depths,and shaft size, which restrict output. What typeof mining extraction is to be followed? Under-ground methods vary considerably in theirnature and scope depending upon the geologyof the mineralisation and the host rocks. Somemethods are suitable for large-scale productionin wide zones of mineralisation whilst othersare designed for small workings in narrow veins.

Capital expenditure limitationsThere may well be restrictions on the availabil-ity of capital for developing and equipping themine. Consequently, it may not be possible toachieve the maximum desirable rate of annualproduction despite the presence of adequateand accessible ore reserves.

11.2.2 Quality of mineralisation

Grade is but one facet of quality, albeit animportant one. Other factors such as mineral-ogy and grain size enter into the definition (seesection 2.2). Commodities like coal, iron ore,bulk materials, etc., can often be sold as a run-of-mine (ROM) product that does not requireprocessing. However, most base and preciousmetal ores usually require some method ofprocessing to separate waste rock and ganguefrom the valuable constituents (Wills 1997).Concentration plants vary from low cost (sep-arating sand from gravel) to high cost, complexplants as in the separation of lead and zincsulfides (Fig. 11.3).

11.2.3 Location of the mineralisation

The valuable products have to be taken tomarket for sale, and supplies for the miningoperation follow the same route in reverse.Local availability of power, water supply, andskilled labor must be considered, and localhousing, educational and recreational facilitiesfor the workforce. Clearly mineralisation adja-cent to existing facilities has greater value thanthat in a remote, inhospitable location.

11.2.4 Market conditions

The sale of minerals extracted from a mine isusually its only income. From this revenue themining company has to pay back the capitalborrowed to develop the mine and pay the re-lated interest, production and processing costs,transport of the product to market, productmarketing, dividends to shareholders, and localand national taxes (Fig. 11.4).

Can the product be sold and at what price?It cannot be assumed that demand is growingsufficiently fast to absorb all feasible new pro-duction. There has to be a reasonable assess-ment of future demand and price over the lifeof the mine, or at least its first 10 years, which-ever is the shorter. The product price (seesection 1.2.3) is the most important singlevariable in a feasibility study and yet the mostdifficult to predict. Historic price trends (Kellyet al. 2001, LME 2003) over the last 20 yearsshow a declining market price for most basemetals and energy materials (Fig. 11.5; seeFigs 1.4–1.6, 1.8).

Many mineral products are sold on commod-ity exchanges, in London and New York, wherea price is set each working day and considerablefluctuations occur (Figs 1.4–1.6). Most mineowners have little or no control over theseprices, which are vital for the well being of pro-ducing mines. Market principles suggest thatduring a period of low prices production costsshould be minimized and revenue maximized.One way of achieving this would be to adjustthe cut-off grade to extract the maximum pos-sible tonnage of easily accessible mineralisa-tion at the highest grade. Alternatively duringperiods of higher price, lower grade materialcould be extracted.

Present day mineral and metal prices areavailable in publications such as the MiningJournal, Engineering and Mining Journal, andIndustrial Minerals, and websites such as theLondon Metal Exchange (LME 2003) whichsummarize average weekly and monthly com-modity prices, as well as providing annualreviews.

11.2.5 Economic climate

The prediction of general business sentiment,the demand for products, and inflation and

Page 274: Introduction to mineral exploration

11: PROJECT EVALUATION 257

Coarse-ore bins

Conveyor belts

Zincflotation cells

Thickeners

Filters

Leadconcentrate

Zincconcentrate

Tailings(Gangue)

Leadflotation cells

Hydro-cyclone

Crushersand

screens

Fine-orebins

Grindingcircuit

D

C

BA

FIG. 11.3 Separation of lead and zinc sulfides. The ore has an average grade of 9.3% zinc and 2.0% lead, assphalerite and galena respectively. At the main site the broken ore at 15 cm size is crushed via a primary (A) and a secondary crusher (B) to less than 40 mm size. The former crusher is in open circuit in that nomaterial is returned for re-crushing. The secondary, finer, crusher is in a closed circuit and the crushed rockpasses over screens (C) on which only material less than 40 mm size passes through to the fine-ore bins.Oversize is returned to the crusher B. The crushed ore is then ground, in water, to less than 0.075 mm, whichis the effective liberation size of the contained sulfides, in the grinding circuit D. This is in a closed circuitwith hydrocyclones, which separate oversize material (>0.075 mm) and return it for further grinding and sizereduction, in D. The undersize is now a finely ground powder suspended in water and forms a slurry (about40% solids) which can be pumped for treatment in the flotation units, or cells. In flotation the surfaceproperties of the liberated sulfides are used whereby air bubbles become attached to individual particles ofsphalerite which float to the surface of the cells and are skimmed off as a froth, whilst all other mineralsand rock particles sink. Each sulfide has different surface properties depending upon the pH of the slurryand added chemicals. In this way sulfides are separated from their gangue, and from each other. A sphaleriteconcentrate is produced with, say, 54% zinc, and galena concentrate with 65% lead. The separation processis reasonably efficient and recovers 91% of the contained zinc values but only 77% of the lead. The wasterock, or tailings in the form of a fine slurry, contains about 0.2% zinc and 0.9% lead. (From Evans 1993 andWills 1988.)

Page 275: Introduction to mineral exploration

258 B. SCOTT & M.K.G. WHATELEY

exchange rates over the life of the mine arecrucial. Apart from rising equipment and laborcosts, inflation and exchange rates may havedisastrous effects if they give rise to higherinterest rates on capital borrowed to developthe mine or on the value of local currency.

As far as mining companies are concernedeconomic cycles are doubly important becausein times of recession, and inflation, they sufferfalling mineral prices, rising interest rates onloans, but operating costs continue to rise. Inthe late 1980s the world economy entered aperiod of recession. This led to a dramaticfall in the copper price (Fig. 11.5) to a low of$US0.74 lb−1 in November 1993. Many pro-ducers incurred heavy financial losses on theiroperations. The reverse, however, happened inthe mid 1990s when booming economies tookthe copper prices to $US1.47 lb−1 in July 1995.Conversely, in the late 1990s and early 2000s,a new recession commenced with depressedmineral prices and low profitability – inNovember 2001 the copper price fell to$US0.60 lb−1, although it has risen sharplysince then with sharp increase in demand fromChina. The overall long-term trend is down-ward (Fig. 11.5), which means that producersmust always look for ways to reduce theircosts.

11.2.6 Political stability of host country

Fundamentally a company developing a min-eral property wishes to have its invested capital

1

2

34

DISTRIBUTION OF INCOME

1 Wages, salaries and benefits to employees2 Taxes and imposts to governments3 Interest to providers of loan capital4 Dividends to shareholders5 Reinvested in R.G.C. Group a) depreciation b) exploration and evaluation c) retained earnings

5c

5b

5a

FIG. 11.4 Distribution of cash flow in 1984 byRenison Goldfields Consolidated Ltd. (AfterConsolidated Gold Fields PLC Annual Report1985.)

0

500

1000

1500

2000

Cu

pric

e (U

S$

t–1)

2500

3000

3500

4000

1989 1991 1993 1995 1997 1999 2001 2003

FIG. 11.5 Copper price trends on the London Metal Exchange, 1989–2003. Compare with long term trendscorrected for inflation in Fig. 1.5. (After LME 2003.)

Page 276: Introduction to mineral exploration

11: PROJECT EVALUATION 259

returned plus a healthy profit otherwise theinvestment will be taken elsewhere. Othernecessary and associated objectives includesafety and health of employees, efficient min-eral extraction, care for the environment, andpayment of local and national taxes. Thegovernment of the host country may have adifferent order of priorities, such as continuityof employment, social objectives, and taxes tocentral government for national social policies.It gives a low order of priority to dividends toshareholders and repayment of loans that wereused to develop the mine. In addition an inflowof foreign investment is sometimes dislikedwith the thought of foreign shareholders con-trolling a national resource of the country.Thus it is essential to operate within a frame-work of acceptable law and, if necessary, amutually agreed contract between developerand host. This may require delicate negotiationand it has to produce a true partnership other-wise the agreement will not last, particularly ifthere is a change of government or leadership.There are known deposits with measuredresources that have no value because of theimpossibility of achieving such a satisfactoryagreement (see also section 1.4.4). Investmentin mineral development almost ceased inZimbabwe in the late 1990s and early 2000sbecause of Government policies relating toownership, inflation, and fiscal policy.

11.2.7 Sustainable development, health andsafety factors

The circumstances under which mining devel-opments are viewed and judged have changedover the last few years. Government endorse-ment is no longer the only external factor inthe approval process. Assessment of the social,environmental, health, and safety implicationsof project development now form an integralpart of all applications in the mining leaseapproval process. Potential concerns of localcommunities as well as other interested par-ties such as non-governmental organizations(NGOs), shareholders, and investors can havesignificant influence on a project’s viability aswell as its design and scope (see section 4.3).

The long-term implications for sustainabledevelopment must also be considered. Sustain-able development is development that meets

the needs of the present without compromisingthe ability of future generations to meet theirown needs (Brundtland 1987). It involves inte-grating economic activity with environmentalintegrity and social concerns (MMSD 2002). Itis here that the mining industry has the oppor-tunity to enhance the contribution that miningand metals can make to social and economicdevelopment (Walker & Howard 2002).

Frequently, Environmental Impact Assess-ments (EIA) are a requirement for permitting orregulatory approval with public consultationan underlying condition (Environment Agency2002). It is at this time that concern to theNGOs can be addressed. The range of biophy-sical impacts to be considered may be quiteexhaustive and include effects on air and waterquality, flora and fauna, noise levels, climate,and hydrological systems. The reporting for-mat can be quite prescriptive, demandingcareful attention to content, procedures, andprotocol (see section 17.2).

Typically, an EIA would explain how thecompany will mitigate issues such as blastingvibrations, dust, noise, atmospheric and waterpollution, and increased traffic density. Withmodern equipment and design most, if not all,of these can be overcome but at a cost. Thiscost is, in effect, an added operating chargewhich may be one of the factors that has aninfluence upon the value of a potential mine(see also section 1.4.2).

In comparison with environmental concerns,legal requirements for consideration of socialimpacts may be limited to general demographicand economic concerns. The understandingof social risk is relatively new. Appropriateconsideration needs to be given to the socialelements. This could include changes that aproposed development would create in socialrelationships, community, people’s quality andway of life, ritual, political/economic pro-cesses, attitudes, and values. Normally thisis combined with the EIA in the form of aSocial and Environmental Impact Assessment(S&EIA) report. It would include the sustain-able development, health and safety aspectsof a project. They also have to be managed toensure that new projects and expansions cantake place effectively. They provide an essen-tial contribution throughout the mining cyclefrom the conceptual study to well beyond mine

Page 277: Introduction to mineral exploration

260 B. SCOTT & M.K.G. WHATELEY

closure. They encompass both human and bio-physical considerations and they must be fullyintegrated with engineering, financial and otheraspects of a project. These reports also have toconsider the closure costs. Closure planning isnow considered early in a mine life and is oneof the factors that has a bearing on the valueof the mineralisation. Decommissioning, reha-bilitation and sustainable development allcontribute to closure costs.

11.2.8 Government controls

Any potential mine will have to operate withina national mining law and a fiscal policyenforced by an inspectorate. Fiscal policy isusually of prime importance for it decidesitems such as taxation of profits, import dutieson equipment, and transfer of dividends andcapital abroad if foreign loans and investors areinvolved. If this transfer cannot be guaranteedthen the value of the proposed mine, and themineralisation, will be considerably less (seealso section 1.5.3).

11.2.9 Trade union policy

It is a foolish investor who does not consult thelocal trade unions before starting a new miningproject. Union disputes can stop a mine aseffectively as machinery breakdowns and con-tinual strikes and disputes over working prac-tices can make a mine unprofitable. In otherwords industrial disputes can change ore tomineralized rock of no value. In some countriestrade unions are, in effect, a branch of the gov-ernment and if a contract is agreed betweenhost and developer this will include a tradeunion agreement.

11.2.10 Value of mineralisation – a summary

Mineralisation has value if a saleable productcan be produced from it. “Value” is a financialconcept and is related to the several factorsdiscussed above rather than to just grade andtonnage. The main factors which enhance thevalue of mineralisation are an increase intonnage, grade, mineral recovery, product salesprice, and political stability, together withdecreases in costs of mining, ore beneficiation,transport to market, capital, and taxes.

Of course not all these factors are under thedirect control of a mining company wishingto develop mineralisation into ore reserves.Changes in the way in which global businessoperates, and the impact of ever more sophist-icated and immediate means of communica-tions, have brought large mining companiesunder increasing scrutiny by, and pressurefrom, a wide range of interests. The demand isfor these companies to demonstrate, throughaction, that they understand their responsibil-ity towards the communities in which theirbusinesses are or will be located. As a result,issues that integrate economic activity withenvironmental integrity and social concernshave become central to the planning and man-agement of sustainable development goodpractice (MMSD 2002).

11.3 CASH FLOW

The factors in section 11.2 can be expressed interms of revenue and cost (income and expend-iture) of a proposed mining operation. All rel-evant items are brought together in an annualsummary which is referred to as a cash flow. Inany particular mineral project:

Cash flow

= cash into the project (revenue) minus cashleaving the project (cost)

= (revenue) − (mining cost + ore beneficiationcosts + transport cost + sales cost + capitalcosts + interest payments + taxes)

Cash flow (Fig. 11.6 & Box 11.1) is calculatedon a yearly basis over either a 10-year period orthe expected life of the mine, whichever is theshorter. It is a financial model of all relevantfactors considered in section 11.2. It is the pro-cess by which the value of a tonne of mineral-ized rock is calculated to determine whetherthis tonnage is ore. Such mineralisation onlyhas value by virtue of its ability to produce aseries of annual positive cash flows (revenueexceeding expenditure) over a term of years.Another way of expressing this is to say thatthe value of a tonnage of mineralized rock isthe value in today’s terms of all future annual

Page 278: Introduction to mineral exploration

11: PROJECT EVALUATION 261

BOX 11.1 Cash flow calculation.

From the results of an order of magnitude feasibility study on an occurrence of sphalerite and galenamineralisation, the first 3 years of a 10-year cash flow are given below. The plan is to mine the miner-alisation underground and produce galena and sphalerite concentrates at the main site. These would besold to a lead and zinc smelter for metal production.

Calculations are for year end; interest payments on borrowed capital to build the project are excluded.Units are $US1000 and 1000 tonnes.(a) Total initial capital (CAPEX) to establish the mine is $140,600. Construction will take 2 years: years−1 and −2.(b) Ore production is 2400 t yr−1 but only 1600 t yr−1 in the first year of production (year 1). Ore grade is9.3% zinc and 2.0% lead.(c) Ore treatment is as in Fig. 11.3.(d) Revenue: the selling price of zinc is taken as $0.43 lb−1 and for lead as $0.21 lb−1. Translating thisvalue to ore gives a revenue of $46.6 per tonne.(e) Operating costs per tonne:

Mining $7.25Mineral processing 8.10Overheads 4.35

$19.70 per tonne ore

The cost includes transporting the sphalerite and galena concentrates to a local port, and loadingcharges. The concentrates are sold to the smelter at the port and it is then the smelter company’sresponsibility to ship the concentrates to their plant.(f ) There is a royalty payment to the previous mineral owners of 4.5% of gross revenue less productioncost.(g) Calculation f is made before the payment of tax.

Cas

h flo

w

Pro

babi

lity

of f

ailu

reS

tage

+veProduction

Year

Production

0

Exploration

evaluation

Develop-

ment

−ve

−6 −4 −2 1 3 5 7 9

FIG. 11.6 The cash flow of a mineral project over 15years. During the first 4 years the mineralisation isevaluated and over the next 2 years the mine is builtat a total capital cost of $US250M. Over these 6years the cash flow is negative in that there is norevenue generated. In the first year of production(Year 1) revenue commences and continues duringthe life of the mine. From this revenue is paid all themine operating costs and repayment of the capitalloans, plus interest, which were used to develop themining operation. The lenders of capital requiretheir money back within the first 3–5 years ofproduction. Also shown is the probability of failureof the project, and its main stages. Obviously thesize of the database increases with time and theprobability of failure decreases correspondingly.

Page 279: Introduction to mineral exploration

262 B. SCOTT & M.K.G. WHATELEY

cash flows arising from mine production. Howthese values are determined is discussed in sec-tion 11.5.

The basic data for the financial model, orcash flow, is collected in a series of studies,which are considered next. Some geologistsmay reject this involvement with finance.However, it has to be remembered that whileall acknowledge the expertise of mine making(i.e. transposing mineralized rock into ore)which lies with geologists and other engineers,the ultimate success or failure of a project ismeasured in financial terms. Decisions aremade with whatever geological information isavailable and it is the geologist’s responsibilityto see that as much critical information aspossible is taken into account. Many geologicalassumptions are included in financial decisionsregarding the development of a mineral pro-perty and geologists should participate fully inthese financial deliberations.

11.4 STUDY DEFINITIONS

A mining company may be exploring for newdeposits (“greenfields”) or evaluating a mineextension (“brownfields”) possibly as a resultof the discovery of additional ore, an increasein commodity price, a change in the miningmethod, or the introduction of different tech-nology. For each project technical and eco-nomic studies are required to determine if theproject is valuable to the company concerned.Initially, especially during exploration, theamount of information available is limited andthis constrains the range and detail possible instudies. As more data are collected across arange of disciplines, including geology, mining,mineral processing, marketing, etc., the studiesbecome more extensive and eventually “homein” on the single project with the most value.The end result is a series of study stages that inmost cases involve increasing levels of detailand expenditure.

The definition of each study stage is import-ant and this is part of the job of the competentperson (see section 10.4.1). In most cases studystages are defined by a set of objectives at thestart, a series of work programs designed toachieve these objectives, and a decision pointat which the project may progress to the next

stage, pause while more data are collected orperhaps be discarded in favor of a more attrac-tive investment opportunity. Naturally, thedetail of the content of the program will varyaccording to the mineral type, geological andmetallurgical complexity, location, environ-mental considerations, etc., so that there is noready-made recipe that applies in all situations.

Reports at the end of each stage describetechnical issues such as geological, miningengineering, mineral processing and extractivemetallurgy, infrastructure, environmental,social, financial, commercial and legal aspectsof the project (Goode et al. 1991, Smith 1991,Barnes 1997, Northcote 1998, White 2001).Similarly there is a range of business issues tobe addressed, such as the project’s ownershipstructure, permitting, marketing and govern-ment relations. Business and technical issuesalso interact, e.g. when a particular productspecification has to be achieved to enable mar-ket penetration.

From these inputs capital and operating costsare derived to prepare estimated cash flowsfrom which the value of the mineralisation inquestion is determined. The value provides abase on which decisions can be made, includingstrategic or political, technical, financial, andsocial aspects. The project moves progressivelyfrom an exploration target, where “back of theenvelope” estimates are made, to a feasibilitystudy from which there is sufficient confidencein the information for a financial institute tolend the capital that will be required to bringthe mine on stream.

Estimates may be ranked according to theinformation available and the extent and depthof study completed. A ranking proposed here(N. Weatherstone, personal communication)ranges from conceptual through order-of-magnitude and pre-feasibility to feasibilitystudies. An example for geological studies isshown in Fig. 11.7. A conceptual study reportmay only comprise a single sheet compiled ina few hours or days, whilst a feasibility studymay involve a team of geologists, engineers,and other specialists for several months oryears. A decision point is reached after eachstage, which marks a significant point in theproject development. The project only con-tinues if there are sufficient indications ofsignificant tonnage and grade, mining and

Page 280: Introduction to mineral exploration

11: PROJECT EVALUATION 263

FIG. 11.7 Flow chart showing the range of feasibility studies and associated decision points.(After N. Weatherstone personal communication.)

Conceptual studies

Resource notdefineable

Decision point

Order of magnitudestudies – data collection

and validation, interpretation,modeling and resource

estimationInsufficient data

Inferred

Discardproject

Decision pointresources

Pre-feasibility studiesDrilling and other data

collection and validation,interpretation, modeling and

resource estimation

Indicative

resources

Feasibility studiesInfill resource and

geotechnical drilling,modeling and resource

estimation

Measured

resources

Detailed engineeringOre reserves

+ve NPV

–ve NPV

–ve NPV

–ve NPV

+ve NPV

+ve NPV

+ve NPV

Insufficient data

Decision point

Insufficient data

Decision point

processing are feasible, and indications are thatit is economically viable.

It is recognized in the minerals industry thatthe chances of the successful establishment ofa new mine are low. In other words, the con-version of mineralized rock into an ore reserveis a high-risk venture. A large and successfulCanadian mining company investigated 1000mineral projects over a 40-year period. Ofthese only 78 warranted major work, only 18

were brought into production but 11 did notreturn the company’s capital investment. Onlyseven were viable from the thousand projectsconsidered.

Consideration has to be given to the reader-ship of these studies as there are different levelsof understanding regarding the details of aproject and differences of emphasis are alwayspossible. Broadly is it intended for internal orexternal (outside the company) use? If the

Page 281: Introduction to mineral exploration

264 B. SCOTT & M.K.G. WHATELEY

former, is it for an operational or an administr-ative group? External readers could be a gov-ernment department, prospective investors, orbanks that may loan money for the project.

11.4.1 Conceptual study

Most successful exploration programs start atthis study level. A conceptual study establishesthe likely presence of a resource. This estimatemay be only preliminary, but is the first step todefining an inferred resource as classified bythe JORC code. It is based upon the interpreta-tion of the size, shape and grade of the potentialresource and the possible mining and process-ing methods. The resource estimate may onlybe based upon reference to deposits of similargenesis and size and may only be an office-based evaluation. There might be sufficientindications of tonnage and grade to encouragefurther work. It is important to identify poten-tial issues, such as access, topographical, ormining lease constraints.

During the study similar deposits elsewherewill be used as the basis to establish whethertheir mining or processing methods are tech-nically feasible on your deposit. Infrastructureand engineering requirements will be costedto ±35%, although estimates will be too unreli-able to attach confidence limits. Estimates canbe calculated from empirical factors and com-parison with similar operations, or in a mineextension, current and historical statistics andcosts are immediately relevant and the accu-racy of such estimates may be within ±10%.Such projects also follow the study definitionsprocess.

For evaluation of exploration projects, muchinformation is taken from sources such as thepublications of state and national bureaux ofmines and from reports of projects in the min-ing literature. Such examples and much usefuldata are included in O’Hara (1980), Mular(1982), Bureau of Mines (USA) Cost EstimatingSystem Handbook (1987), Camm (1991), Goodeet al. (1991), Smith (1991), Craig Smith (1992),and references such as the Annual ReviewReference Manual and Buyers Guide of the Ca-nadian Mining Journal. Websites that providecurrent information include that of Minecost(Minecost 2004) and Western Mine Engineer-ing (Westernmine 2004). Major social and/or

environmental sensitivities will need to beidentified. The project should indicate itspotential to produce enhanced value (NPV, seesection 11.4.3) when compared to other poten-tial investments.

11.4.2 Order-of-magnitude study

An order-of-magnitude study is normally afairly low cost assessment. However, the costof order-of magnitude studies is increasing overthe years, as companies have to address risksearlier in the process. The study is based uponlimited drilling and other sample collectionto establish whether there could be a viableproject that would justify the cost of progress-ing to a pre-feasibility study. The results definethe presence of sufficient inferred resources towarrant further work. Mineralogical studies atthis stage will identify undesirable elementsand other possible metallurgical issues. Theseresources should define sufficient tonnageabove a given cut-off grade to enable engineersto determine possible mining options and pro-duction rates and thus the preliminary size ofmining and processing equipment. Social andenvironmental baseline studies will be initi-ated. Preliminary capital and operating costscan be established to somewhere in the regionof ±30%. Financial modeling of options andsensitivities will be undertaken.

11.4.3 Pre-feasibility study

The aim of the pre-feasibility study is to evalu-ate the various options and possible combina-tions of technical and business issues, to assessthe sensitivity of the project to changes in theindividual parameters, and to rank variousscenarios prior to selecting the most likely forfurther, more detailed study.

The pre-feasibility study incorporates majorsampling and test work programs. Upon com-pletion of a pre-feasibility study, geologicalconfidence is such that it should be possible topublicly declare ore reserves (from measuredand indicated resources), and any other mineralresources that may become mineable in thefuture with further study. Similarly, the min-ing method and production rate will have beenselected. Extensive mineral processing andmetallurgical test work, possibly using a pilot

Page 282: Introduction to mineral exploration

11: PROJECT EVALUATION 265

plant, will have demonstrated that a productcan be extracted using a viable process. Themine infrastructure and labor requirements,and the impact of the potential mine on the en-vironment will have been evaluated. Detailedevaluation of capital and operating costs canbe established to ±25%. These figures will beused, along with a preliminary mine schedule,in a financial and commercial evaluation. Anumber of sensitivities will have been assessed,including alternative methods of mining andprocessing the mineralisation, and the effectsof various levels of output. The integration ofsocial and environmental aspects is most im-portant in the consideration of options for majorproject decisions. Findings from the social andenvironmental baseline studies are used as thebasis to predict and evaluate the likely effectsor impacts, whether negative or positive. Theoption that demonstrates the highest valuewith acceptable (lowest) risk will be selectedas demonstrably viable. Such a study can costseveral hundred thousand dollars.

11.4.4 Feasibility study

The aim of the feasibility study is to confirmand maximize the value of the preferred tech-nical and business options identified in thepre-feasibility study stage.

During the feasibility study the environ-mental and other statutory approvals processes,commenced during the pre-feasibility study,will continue but at an accelerated pace. Con-sultations and negotiations with local commu-nity groups, landowners, and other interestedparties will proceed to the point of basic agree-ment. Social and environmental impact assess-ments will be required that meet statutory andcompany requirements. The timing of formalsubmissions for permitting and other statutoryprocesses needs careful consideration. Sub-mission too early may prematurely commit theproject to conditions which may constrainscope and compromise value.

Sufficient sample collection and test workhas taken place during a feasibility study formore of the resource estimate to be reported inthe measured category. On large projects sev-eral million dollars may have been spent tobring the project to feasibility study level. Sen-sitivity analyses will have been run to assess

the major factors that may have an impactupon the reserve estimate. This will help quan-tify the risk associated with the reserves,which at this stage will fall within the compa-ny’s acceptable risk category. If the company isapproaching a financial institute to borrow thecapital that will be required to bring the mineon stream, these risk factors will be assessedin detail. Often, financial institutes use inde-pendent consultants to audit the resource andreserve estimates. The reserves will be basedupon a final mine design. This design willprovide the information to enable the companyto request bids for detailed engineering studies.Similarly, a final process flow sheet will havebeen developed. Costs will have been devel-oped to ±15%.

11.4.5 Detailed engineering

Once the decision is taken to proceed todetailed engineering, costs will be developed to±10%. The project is almost certain to progressto a mine. This is the final stage before majorcapital expenditure is incurred and involvesproducing detailed plans and work schedulesso that orders can be placed for equipmentand the mine development can begin. Theseinvestigations are thorough, penetrating, andusually completed by outside consultant com-panies who specialize in this type of work.Such a study usually takes several monthsor years, depending upon the initial database,and can cost several million dollars for a largeproject.

11.5 MINERAL PROJECT VALUATION ANDSELECTION CRITERIA

Previous sections have considered the value ofmineralisation in absolute terms (i.e. a sum ofmoney expressed as a cash flow) and this is thebasis of mineral project evaluation. Anotheraspect of value is its use in a relative sense,that is the ranking of a group of similar pro-jects, or options within one project (King 2000).This relative value is based upon rankingeach project’s absolute value and is particularlyuseful when a company has several projectscompeting for funds. Any company has finitemanagerial and financial resources and a

Page 283: Introduction to mineral exploration

266 B. SCOTT & M.K.G. WHATELEY

choice has to be made between two or morealternatives based on some tangible measure-ment of economic value or return. There arefour main techniques, discussed below, all ofwhich are based on the annual cash flow for theproject in question.

11.5.1 Payback

This is a simple method which ranks mineralprojects in order of their value by the numberof years of production required to recover the

initial capital investment from the project cashflow. In Box 11.2 the projects are ranked inorder of least time of payback: B, D, C, and A.The method serves as a preliminary screeningprocess but it is inadequate as a selection crite-rion for it fails to consider earnings after pay-back and does not take into account the timevalue of money (section 11.5.3). It is useful inareas of political instability where the recoveryof the initial investment within a short periodof time is particularly important and hereproject B would take precedence.

BOX 11.2 Mineral project evaluation and selection criteria.

Mineralisation at four similar prospects has been investigated and an order of magnitude feasibilitystudy has been completed on each. At each location it is thought that a mine can be constructed in oneyear. Production commences in year 1. Which, if any, are worth retaining? Cash flow values are inarbitrary units.

Year

A B C D

Initial investment (CAPEX) −1 (3100) (2225) (2350) (2100)Operating margin

1 500 2000 150 11002 500 1000 450 9003 1000 500 1000 11504 2000 500 3400 950

Payback (years) 3.6 1.2 3.2 2.1Operating margin / CAPEX 1.3 1.8 2.1 1.9NPV at 15% discount −560 +835 +656 +780DCF ROR 11% 40% 21% 30%

The value of the mineralisation at prospect A does not reach the minimum DCF return of 15% and isdiscarded. The remaining three meet this requirement and can be ranked in value:

Criteria Most favorable Middle Least favorable

Payback B D COperating margin / CAPEX C D BNPV B D CDCF ROR B D C

The mineralisation at prospect B has the greatest value although it is not the lowest cost to develop, butit has the great virtue of generating a large cash flow at the beginning of the project. Prospect A, the leastfavorable, has the largest CAPEX and a low cash flow in years 1 and 2. At a discount rate of 15% themineralisation at prospect B is worth 835 units, C is worth 656 units, and D 780 units. On the basis ofthis study prospect B is the one to be retained.

Page 284: Introduction to mineral exploration

11: PROJECT EVALUATION 267

11.5.2 Operating margin to initialinvestment ratio

There are several variations on this theme butit produces a dimensionless number whichindicates the amount of cash flow generatedper dollar invested. It is an indication as to howfinancially safe the investment is and the largerthe number the higher the value of the miner-alisation. It is easy to calculate and considersthe whole life of the project compared withpayback which considers only the first fewyears. All factors in the cash flow (such asrevenue, operating cost, taxes, etc.) are takeninto consideration and all affect the rationumber. However, as in payback, the timevalue of money is not taken into consideration.As a simple example consider spending $100today to receive $300 in 3 years, or spending$100 today to receive $400 in 10 years. Theincrease in the ratio overfavors the latter yetmost would prefer the former. Using thisapproach the projects in Box 11.2 can be rankedin order of value as C, D, B, and A.

11.5.3 Techniques using the time value ofmoney

These techniques are commonly used in thefinancial evaluation of mineralisation andmineral projects. The opening theme is thatmoney has a time value (Wanless 1982). Dis-regarding inflation, money is worth more today

than it will be at some future date because itcan be put to work over that period. If a dollarwere invested today at an interest rate of15% compounded annually, it would amountto $1.00 × (1 + 0.15) or $1.15 after 1 year and$1.00 × (1.15)5 = $2.01 after 5 years. This is thecompound interest formula:

S = P(1 + i)n

where S is the sum of money after n periods ofinterest payment, i is the interest rate, and Pthe initial investment value. In this examplewe can say that the 5-year future value of the $1invested today at 15% is $2.01. Reversing thisviewpoint from the calculation of future valuesto one of value today, what is the value today of$2.01(S) if this is a single cash flow occurring in5 years’ time at the accepted rate of interest of15%? The answer is obviously $1.00 and thepresent value (P) is a variation of the compoundinterest formula above:

P

Sn n

=+

=+

= =( )

$ .( . )

. ( . ) $ .1

2 011 0 15

2 010 497 1 005

This expression is the present value discountfactor, more usually termed the discount fac-tor, and tables of calculated factors are readilyavailable (Table 11.1). From the formula, andthe table, it is evident that this factor decreaseswith increasing interest rates and number of

TABLE 11.1 A selection of present value discount factors.

Years

1 2 3 4 5 10 15

Discount rate 5% 0.952 0.907 0.864 0.823 0.784 0.614 0.48110% 0.909 0.826 0.751 0.683 0.621 0.386 0.24015% 0.870 0.756 0.658 0.572 0.498 0.247 0.12320% 0.833 0.644 0.579 0.482 0.402 0.162 0.065

For explanation see text.The table demonstrates the decrease in the magnitude of the discount factors with increasing years and increasing discountrate, governed by the basic formula:

discount factor =1

(1 − i)n

where i = interest or discount rate (as a fraction) and n = number of years.

Page 285: Introduction to mineral exploration

268 B. SCOTT & M.K.G. WHATELEY

years. From this two important characteristicsin valuing mineralisation and mineral projectsdevelop. Firstly, as discount factors are highestin the early years, the discounted value of anyproject is enhanced by generating high cashflows at the beginning of the project. That is,the ability to extract higher grade and moreaccessible (lower mining cost) ore at the begin-ning of a project can significantly enhance itsvalue. This is the declining cut-off gradetheory. Not all mineralisation is amenable tothis arrangement but if this can be achievedit will provide a higher value than a comparablelocation in which this is not possible. Sec-ondly, discount factors decrease with time andby convention and convenience cash flowsare not calculated beyond usually a 10-yearinterval as their contribution to the valuebecomes minimal. This has the added advan-tage of limiting future predictions. In cashflow calculations it is difficult to predict withsome degree of accuracy what is to happen nextyear – several years ahead is in the realm ofspeculation although best estimates have tobe made.

Thus the value of money today is not thesame as money received at some future date.This concept is concerned only with the inter-est that money can earn over this future inter-vening period and is not concerned withinflation. In this calculation money received atsome time in the future is assumed to have thesame purchasing power as money receivedtoday; it has constant purchasing power and isreferred to as constant money. The effect ofinflation on project value, however, is import-ant and is considered in section 11.7.

Net present value (NPV)

The second formula above is used to determinethe present value of the expected cash flowfrom a project at an agreed rate of interest. Thecash flow from each year is discounted (i.e.multiplied by the factor from Table 11.1 appro-priate to the year and interest rate). A summa-tion of all these yearly present values over thereview period provides the present value (PV) ofthe evaluation model. From this PV the initialcapital investment is deducted to give the netpresent value (NPV). Some cases are given inBox 11.2.

The NPV method includes three indicatorsregarding the value of a mineral project, pro-vided that the net present value summation is apositive sum:1 the initial capital investment is returned;2 the financial return on this investment is thespecified interest rate;3 the net present value summation provides abonus payment which is sometimes called theacquisition value of the mineralisation con-cerned. Projects can be ranked according to thesize of this bonus provided that the same inter-est rate is used throughout (Box 11.2). Thereason why project B is superior to projects Dand C is that the cash flows in project B arelarger in the earlier years and the initial invest-ment is returned sooner.

The selection of an appropriate interest rateis critical in the application of NPV as a valua-tion and ranking technique, as this discountsthe cash flow and determines the net presentvalue. A minimum rate is equal to the costof initial capital used to develop the project.Additional factors such as market conditions,tax environment, political stability, paybackperiod, etc. have to be considered over theproposed life of the project. This is a matter ofcompany policy and usually the interest ratefor discounting varies from 5% to 15% abovethe interest rate of the required initial capitalinvestment. In times of high interest rates thisdiscount rate is particularly onerous.

Discounted cash flow rate of return(DCF ROR)

This technique is a special case of NPV wherethe interest rate chosen is that which will ex-actly discount the future cash flows of a projectto a present value equal to the initial capitalinvestment (i.e. the NPV is zero). This DCFreturn is employed for screening and rankingalternative projects and is commonly used inindustry. Since this discount rate is not knownat the beginning of a calculation an iterativeprocess has to be used which is ideally suitedfor computer processing. A very approximatefirst estimate can be obtained by dividing thetotal initial capital expenditure by the averageannual cash flow, and dividing the result into0.7, but it is dependent upon the shape of theoverall cash flow (Box 11.2). Also over a narrow

Page 286: Introduction to mineral exploration

11: PROJECT EVALUATION 269

BOX 11.3 Sensitivity analysis.

This example is taken from a feasibility study completed on a coal mine in western Canada. In the studyit was apparent that two of the critical factors were variation in production cost and in tonnage of coalsold as these could both have a serious effect on the cash flow. These factors were varied and expressed asDCF ROR:

VARIATION IN MINE PRODUCTION COST

Base case +5% +10% −5% −10%26.6% 22.3% 18.5% 31.5% 36.9%

VARIATION IN TONNAGE OF COAL PRODUCED SOLD

Base case +5% +10% −5% −10%26.6% 33.9% 42.9% 20.2% 14.5%

Obviously the lower the tonnage produced the lower the revenue and the lower the DCF ROR, etc.None of the variations bring the DCF ROR to less than the hurdle of 15%. Results are best presentedgraphically.

range of discount rates the DCF ROR valuecorrelates as a straight line (Box 11.3). Conse-quently by calculating a series (3–4) of positiveand negative NPVs at selected discount ratesthe DCF rate can be seen graphically at thepoint where NPV equals zero.

Generally, mining companies finance miner-alisation for development which has a valueequal to or exceeding a DCF rate of return of15% after tax or 20% before tax has beendeducted.

Comment

These quantitative economic modeling tech-niques have contributed much to an improve-ment in the process of investment decisionmaking. During the 1980s, however, there wasan increasing concern with the totality of min-eral projects and a greater emphasis on aspectssuch as business risks associated with rapidlyrising capital requirements due to inflation, theunpredictability of future economic condi-tions, sophisticated financing arrangements,and host government attitudes.

Some mining organizations have used com-parative production cost ranking to reduceinvestment risk. They require that the produc-tion cost to market for a new project is in the

lower quartile of all major primary producers ofthat commodity. This is based on the premisethat if the commodity prices fall, and there is acorresponding reduction of project revenue,then their operation will be protected by acushion of other and higher cost producers,who, it is assumed, will be forced to reduceproduction at an early date thus stabilizing thecommodity price. It is worth noting, however,that a project with such a low production costwould have a favorable investment rate of re-turn (either NPV or DCF) and would be rankedhighly by these techniques.

The NPV method of valuation requires man-agement to specify their desired rate of returnand indicates the excess or deficiency in thecash flow above or below this rate while DCFprovides the project’s rate of return and doesnot consider the desired rate. As a comparisonbetween the two it is said that NPV providesa conservative ranking of projects comparedwith DCF.

Long versus short run considerations have tobe assessed. As an example compare two occur-rences of mineralisation, one with a 15-year lifeand a 15% DCF and another with a 5-year lifeand a 20% DCF. If the short life, high DCFproject is selected, management must considerthe business opportunity at the end of the

Page 287: Introduction to mineral exploration

270 B. SCOTT & M.K.G. WHATELEY

project. What is the probability of finding anddeveloping a series of such short life, high re-turn mineral deposits? The long-term future ofthe company may be better served by choosingthe mineral project with the lower DCF. Lastly,there is the condition of selecting a series ofsmall projects with high DCF. Small projectsoften require the same management time andattention as larger ones. With a finite amountof time its effective use will tend towards largeprojects with, perhaps, lower DCF rather thana series of smaller schemes.

In summary, the objective of valuation is tosummarize and convey to management in asingle determinant a quantitative summary ofthe value of a mineral deposit. Clearly there isno single perfect method of assessing this andgood management uses every available rel-evant method and is aware of their inherentweaknesses and strengths. In the final analysisin the evaluation of mineral projects there is nosubstitute for sound managerial judgement.

11.6 RISK

Risk pervades our entire life and the way weact. It can be described in two ways – eitherwith qualitative expressions (it’s a sure thing,we have a fair chance of discovery in the UpperPalaeozoic, etc.) or in a quantitative sense us-ing probability. A probability of 1.0 means thatthe event will occur while 0.0 means that itwill never happen; negative probabilities donot exist. A probability of 0.31 means thatthere are 31 chances in 100 that the event willoccur and 69 in 100 that it will not. What is anacceptable risk? Risk is very much a personalassessment. For instance many people wouldnot work in an underground mine because it isseen as being dangerous, yet every day accepthigher risks such as dying from fire in the homeor in a road accident (Rothschild 1978).

In the previous section four methods ofanalyzing cash flows in the valuation of miner-alisation were presented. The projected cashflows were assumed to occur (i.e. probability =1.0) and they did not include quantitative state-ments of risk: they were treated as decisionsunder certainty. However the valuation of min-eralisation involves the introduction of manyfactors (grade, tonnage, mining cost, transportcost, etc.) which are not constants (i.e. there

is not a single, unique value) but variables.Consequently mineral valuation comprises aseries of uncertain decisions and the risk ofthese decisions being wrong can be assessed.

11.6.1 Risk analysis

Qualitative assessment

Adjustment of discount rates for NPVand DCFA commonly used overall method of allowingfor risk is to use an abnormally high discountrate in the valuation:

Safe long term rate + risk premium rate= risk rate

For instance if the accepted minimum valua-tion rate is 15% DCF and there is a perceivedrisk, an added rate is included to bring thisminimum to, say, 23%. This is the simplestmethod of allowing for risk but also the leastsatisfactory as the added rate is a matter of per-sonal judgement; its determination with anydegree of accuracy is impossible.

Adjustment of costsA danger here is that of overadjusting. Considera mineral valuation where as a safety factor it isdecided to increase the best estimate of operat-ing costs, to reduce ore grade and selling prices,and apply a high discount rate as an additionalrisk allowance. It will have to be a remarkableproject to survive such treatment.

A better approach is to calculate a base casefrom the most accurate information available.Risk factors are then added in a sensitivityanalysis (Box 11.3) by considering the variablecomponents one at a time while all others arekept constant. In any evaluation certain com-ponents have a greater effect upon the size ofthe cash flow (and hence value) than others andthe purpose of a sensitivity analysis is toidentify them so that further investigations canimprove their reliability (i.e. make them lessrisky). Commonly grade is varied increment-ally, as are other components such as extrac-tion cost, mineral recovery, etc.

The most significant component is revenuewhich is dependent upon the future sales priceof the product and this is the most difficult toforecast. Capital and operating costs only move

Page 288: Introduction to mineral exploration

11: PROJECT EVALUATION 271

upwards but are not likely to increase at a rateexceeding 15% a year. However selling prices,and particularly for those commodities sold onexchanges, may decrease or increase and inextreme instances can either double or halve ina year (Figs 1.5 & 1.6). Forecasting commodityprice trends is a specialist operation and thereare groups who provide this service on a con-sultancy basis, such as Brook Hunt, London. Ifno reliable price trend predictions are availablea break-even price can be calculated from acash flow at a required rate of return – this isthe price at which operating and financingcosts can be met and the mine can continuein production. Once this minimum price isknown, the possibility of its being maintainedover a standard 10-year period can be assessed.

These sensitivity results are best appreciatedgraphically with variation in the single com-ponent concerned plotted against the relatedDCF ROR (Box 11.3).

Quantitative assessment

Discussion on components used in the cal-culation of a cash flow has not included anyquantitative statement of risk. Each compon-ent has been thought of as if it were a constantwith a probability of occurring of 1.0. Mostof these components are variables, a conceptreadily apparent from considering the grade ofa deposit which is an average value withinits own distribution and probability of occur-rence. The same principle applies to estimatesof mining cost, recovery, selling price, etc.

Normally a spreadsheet model will onlyreveal a single outcome, generally the mostlikely or average case. The use of simulationenables the user to automatically analyze theeffects of varying inputs on the modeled sys-tem. Probability of occurrence can be com-puted in a spreadsheet, by randomly generatingvalues for uncertain variables many times over.The result is a probability distribution of theuncertain variable, or a simulated model. Onetype of spreadsheet simulation is referred to asMonte Carlo simulation (Decisioneering 2003,Lumina 2003).

It is possible to use these component dis-tributions in the calculation of a cash flow withtheir probability of occurrence, instead of fixedvalues with no risk element. The probabilitydistribution of DCF or NPV can be calculated

and also the probability of a certain rate of re-turn (Wanless 1982, Goode et al. 1991, Hatton1994b, Decisioneering 2003).

11.7 INFLATION

Mining is among the most capital intensive ofindustries, and ranks near the top of the indus-trial sectors in this respect. This is related tothe complex technology of modern productionsystems; high investment in major infrastruc-tures such as town sites, railways and ports,and expenditure on social and environmentalaspects. To offset this trend projects becamelarger (i.e. higher production) to benefit from aneconomy of scale which would either controlor reduce increasing capital (and production)costs per unit of production. Generally, how-ever, the capital cost of mineral projects isrising significantly. This inherent increase isexacerbated by inflation. As an example, aninflation rate of 7% a year for 5 years will in-crease estimates by 40% in year 5. The impactof high inflation rates on capital (and operatingcosts) for projects with a long pre-productionperiod of several years can be considerable.

Related to inflation, monetary exchange ratesare an important item to consider in mineevaluation. When a company operates a mineunder one currency but sells its products in an-other, changes in exchange rates between themcan have serious consequences for cash flows.

When inflation was in low single figures itseffect on the valuation of mineral projects wasnegligible and could be ignored but with higherrates its effect on calculated rates of return maybe considerable. The value of money, measuredby what it will buy, becomes progressively lessas time passes. The overall rate of increase inprice (or fall in the value of money) is measuredby a gross national product deflator and is theweighted average of the various price increaseswithin a particular national economy. Thereare other measures of inflation (Camm 1991)which may be more appropriate in particularcircumstances such as retail and wholesaleprice indices (Economagic 2003), the indices ofcapital goods in various categories, and wageindices for a country, an industry, or a particu-lar skill.

The differential inflation (or escalation)which exists between the various components

Page 289: Introduction to mineral exploration

272 B. SCOTT & M.K.G. WHATELEY

tion. This is because, if the components whichform the cash flow are inflated at the inflationrate and the resultant flow is then deflated forDCF and NPV calculations at the same rate,this will produce a flow identical with that pro-duced by ignoring inflation altogether.

Opinions differ on the treatment of inflationin mineral project evaluation but a safe methodis to use it as a type of sensitivity analysison the base case. Separate specific cost andrevenue items, which correlate well with pub-lished cost indices, can be used over a limitedperiod of, say, 4 years ahead, and thus accountcan be taken of differential inflation over thisperiod. Beyond this, separate predictions areunrealistic and it is advisable to use only asingle, uniform rate for the complete cash flowfor the remainder of the project duration. Thebase case cash flow prediction is in constantmoney whilst the inflation model will be incurrent money.

In most projects the greater part of thefinance to build a mine is obtained from banks.Banks test the ability of a project cash flowto repay these loans, and related interest pay-ments, by applying their own estimates offuture inflation rates to the company base case.Many mineral companies then prefer to pre-pare cash flows for their base case, and sensitiv-ity analysis, in constant money terms and leavethe preparations of inflated flows in currentmoney to the loan-making banks (Gentry 1988).

11.8 MINERAL PROJECT FINANCE

The development of mineralisation into a pro-ducing mine requires a skilful combinationof financial and technical expertise (Box 11.4).For the purpose of this section assume that afeasibility study (section 11.4.4) has been com-pleted and the company concerned intends toproceed with the development of the definedmineralisation.

11.8.1 Financing of mineral projects

Traditional financing

Generally finance to develop mineral projects(Institution of Mining & Metallurgy 1987) isobtained from three sources (Potts 1985).

Infla

tion

(%) Opcost

General

Revenue

Differentialinflation

0 1 2 3

Years

4 5 6

FIG. 11.8 Differential inflation. Revenue isincreasing at a slower rate than general inflation,while the reverse is true with mine operating costs(opcost). There is a positive differential inflationwith the former and negative with the latter.

in the cash flow is an important factor. Whileinflation may be assumed always to be a posi-tive factor, escalation can be either positive ornegative as the particular component moves ata rate either faster or slower than the measureof general inflation (Fig. 11.8).

11.7.1 Constant and current money

Cash flows may be calculated in either con-stant or current money terms. Constant moneyis assumed to have the same purchasing powerthroughout the valuation period and thus oneyear’s money may be compared directly withthat of any other. This is the basis on whichDCF and NPV are calculated, as above and inBoxes 11.1–11.3.

In current money the figures used in calcula-tions are adjusted for the anticipated rate of in-flation each year and are, in essence, the figureswhich are expected to be entered into the booksof account for that year and represent the netavailable profit or loss. Money from differentyears is not directly comparable and DCF andNPV derived from this type of cash flow aremisleading.

If the rate of inflation is the same for capitaland operating costs, and revenue (i.e. there isno escalation), and there are no tax or otherfinancial complications (which is seldom, ifever, the case), the DCF yield may be correctlyestimated by completing the calculation inconstant money terms, without regard to infla-

Page 290: Introduction to mineral exploration

11: PROJECT EVALUATION 273

BOX 11.4 Case history of Ok Tedi Mine.

The Ok Tedi Mine, Papua New Guinea (PNG), provides an example of the valuation of mineralisationand associated project financing. This porphyry copper deposit (Rush & Seegers 1998) was discovered in1968 by Kennecott Copper Corporation (USA). Negotiations between this company and the hostgovernment on the development of the project broke down in 1975 and Kennecott left PNG. The PNGGovernment continued further exploration and development work. In 1976 Broken Hill Proprietary(Australia) (BHP) agreed to evaluate the deposit and established that more than 250 Mt of mineralisationat 0.7% copper (in the form of chalcopyrite) occurred in the porphyry and 115 Mt of material at 1.2 g t−1

gold in the capping gossan. A definitive feasibility study was completed in 1979. A consortium wasformed which included BHP, Amoco Minerals (a subsidiary of the Standard Oil Company of Indiana,USA), and a German group of companies that largely represented copper smelting activities. Thissponsoring consortium, managed by BHP, presented proposals for the development of the deposit to thePNG Government in November 1979 which were accepted in early 1980 (Pintz 1984). The proposals,based on the feasibility study, were that development would take place in four stages for the productionof gold metal, and copper concentrates for export to Germany and other locations. A copper smelter andrefinery was not envisaged in PNG.

Construction: 1981–84

Production To produce 22,500 t d−1 gossanStage 1: at 2.9 g t−1 gold to produce1984–86 about 22,000 kg of gold a year, depending on the grade and recovery

Production Gold production to continue asStage 2: previously, but also to mine copper1986–89 ore (average grade 0.7% copper) at a commencement rate of 7000 t d−1 and expanding to

60,000 t d−1 by the end of 1989. Stripping ratio for the overall open pit is 2.5:1

Production Gold mining is phased out andStage 3: copper production continued atpost 1990 60,000 t d−1 ore. At an average grade of 0.7% copper, 80% recovery and 300 production

days a year this is equivalent to 100,800 t yr−1 copper metal in shipped chalcopyriteconcentrates containing 25–27% copper

In the early years low cost material with high value (the gold ore) was produced in order to repay rapidlythe project finance. In later years after this gold ore had become exhausted copper ore would be minedfrom a large open pit. Copper sulfide concentrates only were to be produced on site and these con-centrates sold to a German copper smelter group.

A separate project company was formed, Ok Tedi Mining Co. Ltd, with the following shareholders andsponsors: BHP 30% (managers), German Group 20%, AMOCO 30%, PNG Government 20%.

In 1981 the estimated total cost of Stage 1 was $US855M with a project debt to equity ratio of 70:30.Consequently the amount of equity required from the sponsors was $US256M and the remaining $599Mwas raised as project finance from a consortium of banks in several countries.

Construction was completed on time and production stage 1 was successfully concluded. However,the commencement of stage 2 coincided with a decrease in the price of copper and the ensuing delays inproject development caused renegotiation of several aspects of the initial agreement between the projectcompany and the host government. The project is continuing, although the lack of a tailings facilitycaused intense controversy in the early years of the twenty first century and BHP Billiton’s 52%shareholding has been transferred into a company promoting sustainable development in PNG.

Page 291: Introduction to mineral exploration

274 B. SCOTT & M.K.G. WHATELEY

EquityThis is finance provided by the owners (i.e.shareholders) of the company developing theproject through their purchase of shares in thecompany when it is floated on a stock ex-change. Throughout the life of the mine theseshares may be bought and sold and if the mineis successful they may be sold at a higher pricethan their original value. Another return toshareholders is dividends on each share whichare paid out of profits after other financial com-mitments have been met. A successful minemay pay dividends from the start of productionto the end of its life but for many mines fluct-uating mineral prices mean that dividendamounts vary dramatically from year to year.Usually a reduction in the dividend leads to adrop in the share price (Kernet 1991).

DebtIn this case money can be supplied by sourcesoutside the company, usually a group of banks.Debt finance places the company in a funda-mentally different position than that of equityfinancing. Lenders may have the power toforce it to cease trading (i.e. close down) ifeither interest charges or loans are not paid ina previously agreed manner. Thus, ultimately,control of the company is in the hands ofthe suppliers of finance and not the miningcompany itself.

Retained profitsA successful company may retain some of itsprofit and not distribute all of it in the form ofdividends to shareholders. In this way a sourceof finance can be accumulated within a com-pany that is preparing to develop a mineralproperty.

The traditional means of financing mineraldevelopment in the first half of this centurywas a combination of the issue of equity, debtfinance, and the use of retained profit. Thismethod was adequate while the capital cost ofdevelopment was millions or tens of millionsof dollars. During the last three to four decadesthe capital cost and size of major mineralprojects has increased rapidly and large projectsnow cost several hundreds of millions of dol-lars, possibly a billion dollars. With these levelsof expenditure very few, if any, mineral com-panies are able to finance new ventures using

traditional methods. Thus companies havesought other methods of financing which madean optimum use of their financial strength andtechnical expertise but preserved their borrow-ing capability as much as possible. One methodof achieving these ends is project finance.

Project finance

Project financing differs fundamentally fromtraditional financing. The organization provid-ing the finance for the project looks eitherwholly or substantially to the cash flow of theproject as the source from which the loans (andinterest) are repaid and to its assets as securityfor the loan. In this way mineral projects arefinanced on their own merit rather than fromthe cash flow of the mining company that ispromoting the scheme (the sponsor).

Lenders to the project like to see securityattached to the revenue of the cash flow in theform of firm sales contracts for the mineralproducts. It is common for them to form a con-sortium to spread any lending risk as widely aspracticable. Project finance is then a type ofnonrecourse borrowing, which is not depend-ent upon the sponsor’s credit. However, for thisthe sponsoring company isolates the newproject from its other operations (Fig. 11.9) andusually has to provide written guarantees thatit will be brought to a specified level of produc-tion and managed effectively.

Banks prefer to have a safety margin as pro-tection against a deterioration in the projectcash flow. They will therefore agree to financeonly a proportion (say 60–80%) of the cost of aproject with the sponsor providing the remain-der. Full debt financing is rare. Two importantconsiderations which decide this proportion offinance are the length of the payback period(see earlier) and the quality of the managementwho are to bring the new mine into production.If the loan is to be repaid over a relatively shortperiod of time (say 3–5 years) a bank will bemore likely to lend a larger proportion of thetotal capital cost. Management is perhaps theparamount factor in the evaluation of a project.Bad management can destroy a project whichgood management could turn it into the nextRio Tinto plc!

As a final comment it should be rememberedthat the lending banks and the sponsor have a

Page 292: Introduction to mineral exploration

11: PROJECT EVALUATION 275

Sponsor Lender

Project

Lender Sponsor

Companyproject

Project

Project finance

Traditional finance

FIG. 11.9 Traditional and projectfinancing of capital requirementsfor mine development. Forexplanation see text.

difference of emphasis in project development.A mining company will wish to achieve a posi-tive return on its investment (i.e. the equity ofthe project company) to satisfy the demandsof its shareholders. A bank, however, will wishto ensure that its money is returned on time,with interest and loan repayments taking pri-ority over dividend payments to shareholders.The payment of dividends, however, can haveconsiderable effect on the price of the com-pany’s shares traded on stock exchanges. Con-sequently, a compromise is usually reachedwhere the major proportion of the cash flow isreserved for project loan repayment, a smallerpart (say 20%) is placed at the disposal of theproject company.

It is a matter of opinion as to whether cashflows used in DCF analysis should includedrawdown and repayment of debt (i.e. projectfinance), and interest payments. If not thenprojects are evaluated as if they are financedby 100% equity, although in practice thiswill not happen (i.e. financing will be a com-bination of loan and equity). Such a proce-dure is said to separate investment decisionsfrom financing decisions (i.e. repayment ofdebt, etc.). It is a fundamental principle thata basically poor project cannot be trans-formed into an attractive one by the mannerin which it is financed. If an all-equity evalua-tion demonstrates a strong cash flow andmeets standard parameters (see earlier), thenafter this it can be determined how it is bestfinanced.

Banks are not inherently concerned with therate of return of the proposed project. Theirreturn is more or less fixed, based upon theinterest rate at which they agree to lend money.Banks are more concerned with protection oftheir loans and they tend to examine cashflow predicted in the feasibility study from thisperspective. A criterion is that the NPV at anagreed rate of interest must be at least twice thetotal initial loan plus interest payments. Thereis the obvious implication that a bank will lendmoney for a mineral project provided that itsloan and interest payments are secure, evenif actual project cash flows fall to half theirestimated level.

11.8.2 Risk in project finance

Project loans are provided on an agreed timeschedule of drawdown and repayment, basedon the technical data and cash flows in the finalfeasibility report. The interest rate on loans isusually a premium rate (say 3%) above anagreed national base rate. As this base rate fluc-tuates the loan interest rate will also change,and in times of inflation with high base rates(as in 1990 and 1991) repayments can be oner-ous for a project company.

Even the best of feasibility studies maybe incorrect in actual practice and events towhich both sponsors and lenders particularlylook during the construction and operation ofa new mineral project are as follows describedbelow.

Page 293: Introduction to mineral exploration

276 B. SCOTT & M.K.G. WHATELEY

During construction

It is probable that the greatest risk in projectfinancing occurs during the construction phase.This phase may legitimately take several yearsand during this time the loan is being usedand no revenue is being generated. Mine con-struction is usually contracted out to majorcivil engineering companies and some formof completion guarantee from the managingcontractor is crucial. This is usually to com-plete and commission the project in a specifiedtime period and to make penalty paymentsto provide funds if construction is not com-pleted on time. Complications which can delayconstruction include an increase in capitalcosts due to inflation or technological diffi-culties, delays in the delivery of plant andequipment, strikes, delays caused by abnormalweather, financial failure of a major contractor,and poor management.

A delay in completion and subsequent opera-tion postpones the generation of the cash flowfrom which loans and interest were to be repaid.It also creates a requirement for additionalmoney over and above that originally negoti-ated in order to finance this delay period.Obviously the key factors are the quality of thefinal project feasibility report and the reliabil-ity of the managing contractor.

During production

Adverse factors during production include:errors in the quantity and quality of ore reserveestimation, technical failure of equipment, fallin the price of the mineral product, foreignexchange fluctuations such as a currency de-valuation, poor labor relations which provokeindustrial strife, environmental difficulties,and government interference in the runningof the project of which the final condition isexpropriation. Any of these events may causedelay in the repayment of the loan. In cases ofgrave default the lenders have the ultimateright to take over the project and replace itsmanagement.

General

Banks, in reviewing a feasibility study, payparticular attention to the degree of confidence

which is attached to estimates of mineablegrade and tonnage, annual production rate,capital, and operating costs. While banks maybe prepared to accept the risk of commodityprices falling below those projected, they arenot prepared to accept technical risks. Ofincreasing importance to banks and projectsponsors is the potential liability in the eventof inadvertent damage to the environment.Consequently, the manner in which a feasibil-ity study examines social and environmentalissues is reviewed with great care.

11.9 SUMMARY

Cash flows in constant money are tested bysensitivity analysis in the valuation and rank-ing of mineralisation, mineral properties, andprojects before mining companies proceed todeveloping a mine. A minimum acceptance rateof return (the “hurdle” rate) is usually stated,below which projects are not considered, e.g. a20% DCF ROR before tax and a 15% DCF RORafter tax. This usually means that the initialcapital investment is returned in the first 4years or so of mineral production. Few com-panies consider inflation and even fewer con-sider a quantitative assessment of risk (section11.6).

11.10 FURTHER READING

O’Hara (1980) and Mular (1982) provide muchuseful basic data on the evaluation of orebodiesand the estimation of mining costs, but arenow rather dated. A more relevant source is theUS Bureau of Mines Cost Estimating SystemHandbook (1987), which provides a rigorousbasis for cost estimating. Websites that providecurrent information include that of Minecost(Minecost 2004), CRU International (CRU2004), Brook Hunt (Brook Hunt 2004), andWestern Mining Engineering (Westernmine2004). A good system for order-of-magnitudeestimates is in Camm (1991) and Craig Smith(1992). Two other articles of note are the de-scription of the Palabora copper open pit inSouth Africa (Crosson 1984), and the NevesCorvo copper–zinc–lead–tin project in Portugal(Bailey & Hodson 1991).

Page 294: Introduction to mineral exploration

11: PROJECT EVALUATION 277

In the authors’ opinion there is a scarcity ofsuitable textbooks and articles on the financialaspects of mineral evaluation, and Wanless(1982) remains a good choice. Aspects of thefinancing of mineral projects were presentedin nine papers at a meeting “Finance for themining industry” under the auspices of theInstitution of Mining and Metallurgy in 1987.MINDEV 97 was an AusIMM event presentingthe A–Z of “how-to” successfully develop amineral resource project (Barnes 1997). Paperswere presented by major mining companies,contractors, and equipment suppliers, coveringthe earliest stages of permitting, preparingfeasibility studies, and assembling the develop-

ment team, through to financing, procurement,construction, and commissioning. Case stud-ies of mineral evaluation and related minedevelopment are comparatively rare but prob-ably the best published is that of the Ok Tedicopper-gold open pit mine (Pintz 1984) inPapua New Guinea.

For up-to-date information on environ-mental impact assessments, the reader isdirected towards the bi-monthly publicationEnvironmental Impact Assessment Reviewby Elsevier Science Ltd. Lord Rothschild’s(1978) short paper on “Risk” is a readableand excellent introduction to this importanttopic.

Page 295: Introduction to mineral exploration

278 B. SCOTT & M.K.G. WHATELEY

Page 296: Introduction to mineral exploration

Part II

Case studies

Page 297: Introduction to mineral exploration
Page 298: Introduction to mineral exploration

12Cliffe Hill Quarry,

Leicestershire –Development of

Aggregate ReservesMichael K.G. Whateley and William L. Barrett

12.1 INTRODUCTION

The annual production and use of constructionaggregates in the UK, including both sandand gravel and hard rock, amounts to approx-imately 220 Mt. The materials are used forroad building, construction, civil engineering,concrete, house building, chemicals, and otherspecialized applications. Within such a widemarket rocks with particular properties are bet-ter suited to certain applications than others.The physical properties of the rocks are definedin terms of their response to different tests,most of which are covered by the relevant Brit-ish Standards. Some international, Americanand local (but not British Standards) tests arealso used.

Tarmac Quarry Products Ltd (Tarmac) is oneof the leading aggregate, ready mixed concrete,and waste disposal companies in the UK; it alsooperates in Europe and Scandinavia, as well asthe Middle and Far East. In 1999 it becamea subsidiary of Anglo American plc. Tarmacoperates about 100 active quarries and sandpitsin the UK, spread from Ullapool in northwestScotland to the south coast of England.

The biggest market for aggregates in the UKis the southeast of England where ironicallythere is no hard rock and demands have histor-ically been met from sand and gravel sources(both land based and marine dredged) and bysupplies railed in from elsewhere in England.Notwithstanding the expected efforts to ex-pand local sand and gravel output from both

land and marine sources, an aggregates supplyshortfall of between 15 and 30 Mt yr−1 p.a. ispredicted for the next decade for southeasternEngland. The deficiency in indigenous aggre-gates will have to be satisfied by increasingimports into the region from elsewhere in theUK and also from overseas. Leicestershire iswell placed as a supplier and has suitable qual-ity rock resources to benefit from this demand.

In the mid 1970s it became apparent that theprofitable microdiorite (markfieldite) quarry ofCliffe Hill in Leicestershire (Fig. 12.1) belong-ing to Tarmac Roadstone Ltd, East Midlandswas running short of recoverable reserves.Significant resources remained at the site, butthese were sterilized beneath the processingplant (Fig. 12.2).

The markfieldites of the Charnwood Forestarea occur as a number of relatively smalligneous bodies intruded 550 Ma ago into lateProterozoic pyroclastics and metasediments.Markfieldite differs from other microdioritesin having a granophyric groundmass and it hasa general uniformity and strength which putsit amongst the best and most consistent gen-eral purpose construction aggregate materialsin the UK. An impressive durability meansthat the markfieldite can also be used for allclasses of railway track ballast, as well as roadstone.

Charnwood Forest contains major faults,e.g. the Thringstone Fault, which brings Car-boniferous rocks against the Proterozoicsequence to the west of Stud Farm (Fig. 12.1).

Page 299: Introduction to mineral exploration

282 M.K.G. WHATELEY & W.L. BARRETT

CharnwoodForest

Bardon Hill

Stud FarmBilla Barra Hill

Cliffe HillQuarry

Stanton underBardon

Bradgate Park

Rai lway

l i ne to Le ices te r

M1

A50Fig 12.2

N

13

11

9

45 47 49 51 53

0 1 km

FIG. 12.1 Distribution of markfieldite (a microdiorite) in the southern part of Charnwood Forest,Leicestershire, UK. Stippled areas are known and projected areas of markfieldite as given in Evans (1968).Hatched areas are other intrusive rocks. The grid coordinates are in kilometers.

This case study gives a clear example ofthe procedure adopted by one company in theexploration for and development of a hardrock resource. Readers wishing to expand theirreading into sand and gravel and limestoneresources are referred to British GeologicalSurvey publications on the procedures for theassessment of conglomerate resources (Piper& Rogers 1980) and limestone resources (Coxet al. 1977). Additional background reading canbe found in the book by Smith and Collis (2001).

12.2 PHASE 1 – ORDER-OF-MAGNITUDE STUDY

Cliffe Hill Quarry is adjacent to the M1 motor-way and a railway giving it excellent accessto major markets in the southeast of England.The initial desk study in 1977 concentrated onextensions to Cliffe Hill Quarry and adjacent

Knowing the location of faults is important inany quarry. Associated minor faulting is alsoimportant to quarrying operations. In CliffeHill Quarry one apparently minor shear zonewas mapped. As a minor feature it could easilyhave been missed during core logging, howeverit offset the Precambrian basement by almost60 m (Bell & Hopkins 1988).

In order to maintain its production from theEnglish Midlands, and to maximize the exploi-tation of a valuable national resource, Tarmacinitiated a search for alternative sources ofsimilar quality markfieldite in the neighboringareas. This search was conducted in a numberof phases with each subsequent and morecostly exercise only being undertaken if clearlyjustified by the results of the preceding one.All costs quoted in connection with this casestudy relate to values current at the time of theexpenditure.

Page 300: Introduction to mineral exploration

12: CLIFFE HILL QUARRY AGGREGATE RESERVES 283

tween the Triassic Mercia Mudstones and themarkfieldite. The mudstones are also referredto as overburden which generally includesglacial deposits as well.

The above exercises could be regarded asseparate phases or subphases, but since theytook place in swift succession they have beenregarded collectively as Phase 1. A cost (exclud-ing overheads and geologist’s time) of around£1500 was incurred for this work.

12.3 PHASE 2 – PRE-FEASIBILITY STUDY

The preliminary desk study had indicated areasof interest, but the Stud Farm holdings wereinsufficient to support a viable quarry (Bell &Hopkins 1988). Negotiations were entered intowith the owner of adjacent land. An agreementwas reached which allowed Tarmac accessto the land for exploratory drilling. A small ex-ploration drilling budget was approved and inmid 1978 nineteen, 150-mm continuous flightauger holes were drilled with a Dando 250 top

Rai

lway

line

to

Bur

ton

West

Lane

Stud Farm

Tythe Farm

EXTN

NQ

Billa Barra Lane

Stanton Lane

Stantonunder

BardonThornton Lane

Mai

nS

t ree

t

Cliffe HillQuarry

ProcessingPlant

M1

RWL

500 m0

4847464544

Fig. 12.4

N

11

10

FIG. 12.2 Cliffe Hill and Stud Farm Quarries, showing the location of some of the borehole sites (dots) and theproposed new railway line (RWL). The pecked and dotted line represents the outline of the area included inthe planning permission. The pecked line represents the outline of the proposed new quarry (NQ) with thepossible extension shown to the east (EXTN).

land and mineral right holdings on Stud Farm.The study included a detailed literature searchwhich indicated that both Billa Barra Hill tothe northwest, and Stud Farm around 1 kmto the west of the existing quarry might be ofpotential interest (Fig. 12.1). The 1:63,360 and1:10,560 Geological Survey maps and the ac-count by Evans (1968) of the Precambrian rocksof Charnwood Forest pointed to additionalareas of diorite near Stud Farm. A well sunkon Tythe Farm (Fig. 12.2) at the end of the lastcentury intersected markfieldite at a relativelyshallow depth.

Subsequent ground surveys eliminated theformer site as being composed of fine-grainedtuff, volcanic breccia, and metasediments, butrevealed scattered float boulders of mark-fieldite in the soils at the latter site. In view ofthe total absence of exposures at Stud Farm, afew pits were dug using a mechanical digger,and although these failed to reach bedrock theyconfirmed the existence of further pieces ofmarkfieldite in the soils. The term bedrockin this chapter refers to the unconformity be-

Page 301: Introduction to mineral exploration

284 M.K.G. WHATELEY & W.L. BARRETT

drive, multipurpose drill rig. These holes,drilled alongside established farm tracks overthe higher parts of the farm, proved thatrock, harder than could be penetrated with theaugers, existed at exploitable depths over anarea large enough to contain around 20 Mt ofmaterial. The results from this drilling enabledthe geologists to establish overburden distribu-tion, overburden types, water-bearing zones,the depth of weathering, and the gradient ofthe bedrock surface (Bell & Hopkins 1988). Atthis stage there was no proof of the natureof the harder material since no samples wererecovered. Nevertheless, it was assumed thatthe impenetrable rock was markfieldite and,although the drill pattern was irregular, itshowed that there was only thin overburden.Geologists calculated the resources and over-burden ratios for a theoretical quarry, andwithin the margins of error in such calcula-tions there appeared to be sufficient resourcesfor a viable quarry. The auger drilling in Phase2 is estimated to have cost between £2000and £2500.

12.4 PHASE 3 – FEASIBILITY STUDY

Phase 3 of the Cliffe Hill Quarry project in-cluded all the exploration drilling, a limitedamount of geophysics, the testing of the coresamples, the design of the landscaping works,the detailed specification of the processingplant, and the submission of the original plan-ning application, and a number of subsequentamendments to it, to the relevant local govern-ment offices.

12.4.1 Drilling

The Estates and Environment Department ofTarmac acts as a contract drilling companyto the operating divisions and the Stud Farmwork, although welcome, caused a number ofproblems. With the continuing land acquisi-tion program rig-time was becoming scarceand although it had been hoped to undertakeany additional work at Stud Farm in-house,the rapid exploration success meant that thefollow-up work could not be fitted into theprogram of the existing, company owned rigs.The company therefore took the decision to

FIG. 12.3 A top-drive, multipurpose, flight augerdrilling rig used to evaluate the resources on theStud Farm property adjacent to Cliffe Hill Quarry.

purchase a brand new six cylinder, 100-hp top-drive, multipurpose, drilling rig (Fig. 12.3) toevaluate fully the resource potential at StudFarm. A rig of the type used, together with aback-up vehicle and the relevant in-hole equip-ment, would have cost around £120,000, butthis figure is included as depreciation in thetotal drilling costs given below.

Drilling, using open-hole methods in theoverburden and coring at various diameters inthe bedrock, commenced in August 1979 andcontinued with some breaks until December1989. During this period about 240 boreholestotalling around 17,000 m were drilled, mainlyon a 50 × 50 m grid (Fig. 12.2). A planningapplication for Stud Farm was prepared inparallel with the exploration work.

The experience gained while drilling theoriginal holes during Phase 2 indicated thatcertain aspects of the drilling would have to beimproved if sufficient data were to be collected.The indications were that there was between

Page 302: Introduction to mineral exploration

12: CLIFFE HILL QUARRY AGGREGATE RESERVES 285

10 and 50 m of overburden. Although the flightaugers could probably cope with this depth,removing the clay from the auger and clearingthe hole proved slow and difficult. The claywas also expansive and, when wetted by thewater flush from a diamond bit, the hole tendedto close and trap the core barrel. These relatedproblems were ultimately solved by using adrag bit and water flush. A drag bit is a bladedbit which is used when the sticky materialsuch as clay and marl would clog up air flushbits. The drag bit drilling technique producedan open hole to bedrock almost as fast as theaugers could when cleaning time was takeninto account. In addition, the clay had to someextent expanded due to the water flush. Bydrilling with sufficient annulus size (the gapbetween the drill rods and the sidewall upwhich the drilling medium (air, water, or foam)carries the drill cuttings to the surface) and cas-ing placement to bedrock, a good core recoverywas usually possible (Bell & Hopkins 1988).

During the early drilling program the drillerremarked on the apparent low penetration rateof the core barrel. At the time this had beenassumed to be the result of operating in anuncased hole. Once the main program began itbecame apparent that, although the high rota-tion speed improved penetration, the produc-tivity was lower than expected. A series ofchecks and experiments narrowed the problemdown to polishing of the impregnated core bits.After a series of trials, conducted in conjunc-tion with one of the leading British bit manu-facturers, a bit matrix which gave the optimumbalance between bit wear and penetration wasdeveloped. Once this problem had been solvedthe drilling proceeded at a rapid rate (Bell &Hopkins 1988). During the summer of 1981a series of boreholes was drilled along a linejoining Stud Farm with Cliffe Hill Quarry(Fig. 12.2). The aim was to investigate the routeof a proposed tunnel to link the two quarries.The purpose of the boreholes was twofold: (i) todelineate the bedrock–overburden interface,in order to ensure that the tunnel which wasplanned to link Cliffe Hill Quarry to Stud Farm(section 12.4.5) remained in solid rock through-out its length, and (ii) to identify the engineer-ing properties of the bedrock with particularreference to jointing and faulting. Some heavyfaulting was identified in two boreholes.

Between September 1981 and August 1982the drilling was concentrated on two areas ofthe site:1 Investigation of the proposed plant site.Some 10 boreholes were drilled where accesspermitted.2 Drilling of a newly acquired area to thenortheast of Stud Farm.

The results confirmed the already familiarpicture of highly variable overburden thick-nesses, with the bedrock often “dipping away”beneath rapidly thickening Triassic MerciaMudstone overburden.

Between April 1983 and March 1985 drillingwas carried out on a 50 m grid across the area ofthe proposed quarry (Fig. 12.2) with the aim of:1 confirming the geophysical interpretation;2 identifying the volume of weatheredmaterial;3 building up detailed knowledge of the faults,rock quality, etc., to enable detailed quarryplans to be drawn up.

The drilling results were used to produce theoverburden isopach map (depth to bedrock)(Fig. 12.4). This map was initially drawn inde-pendently of the geophysical data and clearlyshows the success of the previous electromag-netic and resistivity surveys in revealing thedistribution of the overburden (Figs 12.5 &12.6). Although some inaccuracy is evident inthe deeper areas, with the resistivity methodunderestimating thickness of overburden, thetwo shallow ridges and steep sides were accur-ately delineated. Drilling indicated that thesefeatures can be attributed to faulted blocks ofmarkfieldite alternating with late Proterozoicmetasediments, as seen in Cliffe Hill Quarry.

12.4.2 Geophysics

As indicated previously, the planning applica-tion was being prepared in parallel with theexploration work and the work was coordin-ated by a team of geologists, engineers, estatessurveyors, and landscape architects. The con-tinual acquisition of additional borehole dataduring the planning of the project meant thatoccasionally parts of the scheme had to beredesigned. However, the boreholes revealedthat the areal extent of the markfielditewas somewhat larger than originally anticip-ated. It became apparent that a geophysical

Page 303: Introduction to mineral exploration

286 M.K.G. WHATELEY & W.L. BARRETT

Figs 12.5 & 12.6

New Railway

X X′

5545

3525

155

5

15253545

N

0 500 m

Resistivitysounding

Tythe Farm 0 30 m

BH 35

BH 36

BH 3

BH 18

BH 37 BH 38

A

B

15

5

15

25

35

45

515253545

55

N

FIG. 12.4 Overburden isopach map on the Stud Farm property drawn from data derived from borehole results,conductivity measurements, and resistivity soundings.

BH 35

BH 36

BH 3

BH 18

BH 34

BH 37 BH 38

A

B

35

30

30

515

510 1520

Borehole

Conductivitymeasurement

Tythe Farm 0 30 m

N

FIG. 12.5 Overburden isopach map on the Stud Farmproperty drawn from data derived from conductivitymeasurements.

FIG. 12.6 Overburden isopach map on the Stud Farmproperty drawn from data derived from resistivitysoundings.

investigation of the extent and form of theintrusion was essential. It was hoped that thisknowledge would reduce the number of majorinterpretational changes that might be madein the future and also enable the drilling to bemore effectively planned.

Refraction seismics

Following a trial seismic refraction survey inDecember 1980, a full survey was carried out

across the whole of the site in February 1981. Astandard 12-channel refraction seismic tech-nique was employed, with a Nimbus seismo-graph recording on 12 geophones the firstarrivals of the ground waves produced from theexplosive source. The aim was to produce anoverall picture of the bedrock topography.

The results of the survey indicated that theoverburden was thin near the topographicallyhigh center of the property, and that overbur-den increased in all directions away from this

Page 304: Introduction to mineral exploration

12: CLIFFE HILL QUARRY AGGREGATE RESERVES 287

center. Unfortunately the major geophysicalsurface did not correlate well with the surfaceof the markfieldite as identified by borehole in-formation. It was assumed that this was be-cause the seismic method was delineating thebase of a weathered or fractured layer. In thisparticular deposit the depth of weathering is upto 20 m on the higher ground, but significantlyless in the bedrock on the flanks of thehill where the older rocks were more deeplyburied below the marls of the Triassic MarciaMudstone. Much of the weathered rock hasto be quarried conventionally and can be usedfor lower specification purposes. The seismicrefraction method in this application had theeffect of making the bedrock surface appeargenerally deeper than it is and also less variableat depth.

Electrical resistivity and electromagnetics

Planning permission was granted in August1983 by the Leicestershire County PlanningDepartment. Immediate priority was given tothe detailed investigation of the site for devel-opment. An accurate overburden volume figurewas required and insufficient boreholes hadbeen drilled. Further geophysical surveys werecommissioned to obtain additional detail of theoverburden distribution.

Overburden thicknesses were required on a50 m grid across the area investigated in Phase1, and it was decided to use a combinationof electrical resistivity and electromagnetictechniques to produce the detail required.

The area was traversed using the GeonicsEM34 Ground Conductivity Meter (20-m and40-m coil separations). The Mercia Mudstone,which composes most of the overburden, isconsiderably more conductive than the under-lying markfieldite. Thus an increased thick-ness of the overburden produced an increasedconductivity reading. The readings, reflectingthe degree of conductivity of the different ma-terials, were then used to produce a conductiv-ity contour map. Therefore, by traversing thesite it was possible to produce a qualitativepicture of overburden variation. Although theassumption of a two-layer model of overburdenand bedrock oversimplified the situation some-what, e.g. an additional boulder clay layer, etc.could be present, the marked contrast in con-

ductivity meant that the method proved to bevery successful in delineating the shallower(<20 m deep) bedrock features (Fig. 12.5). Al-though borehole information allowed somecorrelation of conductivity with depth, theresults remained essentially qualitative and aresistivity survey was carried out to quantifythe data (Fig. 12.6).

The quantification of the conductivity con-tours was attempted by taking a series of resis-tivity soundings (electrical depth probes) alongsectors of equal conductivity. This techniqueproduces a more accurate depth reading at apoint, by avoiding significant lateral variationsin overburden thickness. The resistivity sound-ing methods used employed the British Geo-logical Survey (BGS) multicore offset soundingcable (a Werner configuration) with an AbermTerrameter. An electrical current is passedthrough two electrodes and the potential dif-ference is recorded across a further pass ofelectrodes to produce a resistance value in thestandard array. The curves produced are com-puter processed to give depths to bedrock at themeasurement sites.

With the highly variable nature of the over-burden thickness, the electromagnetic resultsproved to be very important for accuratelypositioning the resistivity lines so that theydid not cross any sharp lateral variations inbedrock topography. This greatly increased thequality of the resistivity data.

Combining the electromagnetic and resist-ivity results with borehole data produced anoverburden isopach map (Fig. 12.4). The mostimportant features identified from this plan arethe two shallow “ridges” (marked A and B onFigs 12.5 & 12.6) away from which the overbur-den thickness increases rapidly. Neither ofthese features had been identified by the earli-est drilling and, because of their narrow nature,might have remained undetected for sometime. Their presence resulted in a significantreduction in calculated overburden volume.

Reflection seismics

Recent modification to seismic source, detec-tor, recording equipment, and field techniqueshas improved the seismic reflection method ofexploration to enable it to be used sucessfullyfor exploration at shallow depths. The system

Page 305: Introduction to mineral exploration

288 M.K.G. WHATELEY & W.L. BARRETT

FIG. 12.7 A migrated section of part of line X–X′ (Fig. 12.4), showing an image of an ancient wadi cut inmarkfieldite and filled with Triassic mudstone. (After Ali & Hill 1991.)

has been used recently (Ali & Hill 1991) togenerate, extract and record high-resolutionseismic data which eliminated noise, in par-ticular low frequency, high amplitude, source-generated ground roll. High-frequency energywas needed to identify and distinguish themfrom other interfering waves.

Shallow reflection seismic methods wereused on the Stud Farm property to locate andimage the unconformity between the TriassicMercia Mudstones and the markfieldite. It wasassumed that the homogeneity of the overlyingmudstone would result in no reflections fromwithin their sequence, and therefore make thedetection of reflections from the unconformityeasier (Hill 1990). However, some reflectionswere obtained from within the mudstone andthese are thought to represent sandy and cal-careous horizons (Fig. 12.7).

Data were recorded along two lines (Fig.12.4), usually with sixfold coverage and 1-mgeophone spacing. Despite poor weather andpoorly consolidated near-surface materials,some reflected waves were recognized. A prob-able function was chosen based on experiencegained in a nearby quarry, with similar geology(Ali & Hill 1991) and stacking of the datacarried out. Careful processing and correction

of the data produced a section on which theunconformity is clearly imaged as a steep sided,ancient wadi (Fig. 12.7).

12.4.3 Sample testing

In any deposit, and particularly in a new one, itis essential to determine the precise physicalcharacteristics of the materials, both in order toestablish the potential markets and to deter-mine the type of processing plant that will berequired. In some deposits a detailed chemicalprofile of the material may also be necessary.

The intrusive rocks of this part of Leicester-shire are known to be hard, durable, and mainlyconsistent, but it is nevertheless essential tocarry out a large number of British Stand-ard and other special tests to quantify anyvariations and to establish the precise charac-teristics of the materials. The core samplesobtained were of two sizes (35 mm and 60 mmin diameter) and all were subjected, wheresample quantity permitted, to the followingtests as stipulated in British Standard 812(1975): Relative Density, Water Absorption,Aggregate Impact Value (AIV), and Aggre-gate Crushing Value (ACV). A few sampleswere also tested under the same standard for

Page 306: Introduction to mineral exploration

12: CLIFFE HILL QUARRY AGGREGATE RESERVES 289

Aggregate Abrasion Value (AAV) and PolishedStone Value (PSV).

The Leicestershire quarries have historicallysupplied stone to the profitable British RailHigh Speed Track ballast market. In order toqualify for this market the stone must achievea very low and consistent value in a non-BritishStandard wet attrition test. A larger size mater-ial is required for this test and consequentlyonly the 60-mm cores could be used. As testingprogressed, it proved possible to establish a us-able correlation between the water absorptionvalues determined on all samples and the wetattrition values.

Over 120 samples of core were tested forsome, or all, of the properties mentioned above.The results provided the data necessary to de-termine which markets could be penetratedand what type of basic plant would be requiredto produce the materials to satisfy those outletsin terms of both anticipated quantity and prod-uct specification.

12.4.4 Reserve estimation

The confidence in the resources grew as theamount of data increased and several refine-ments of the estimation method evolved asfollows:1 After the initial phase of flight auger drillingan inferred resource estimate (USGS 1976) ofsome 20 Mt ± 100% was quoted.2 By the time the first planning applicationwas submitted, a generalized bedrock and de-posit shape had been defined by widely spacedboreholes and some typical cross-sections hadbeen drawn. A resource estimate based on thearea of the cross-section, extended to the midpoint between sections likely to be recoveredusing an average quarry configuration, gavean enhanced figure with a confidence limit of±20%.3 By 1989 with much of the relevant drillingcompleted, two sets of cross-sections (N–S andE–W) at 50-m intervals were drawn. With aprovisional quarry development scheme super-imposed, a more precise reserve estimate wasproduced. The extractable area on each cross-section was multiplied by the distance to themidpoint between the adjacent sections oneach side. The N–S sections were consideredbest but the figures were also verified using the

E–W sections. In all cases the volumes wereconverted to tonnages by multiplying the vol-ume by the average relative density of the rockin a saturated, surface dried condition. In thiscase the value 2.7 was used for the fresh rock.The final calculations were considered moreprecise and were accorded a confidence limitof ±10%.

12.4.5 Quarry planning

As the drilling and sample testing continuedwith a progressive clarification of the com-bined overburden and weathered rock thick-nesses, and of the overall structure, shape, andquality of the deposit, three possible operatingschemes were devised for developing the newquarry in conjunction with recovering the size-able resources which would become availablefollowing the dismantling of the plant at theold quarry.

Bell and Hopkins (1988) described thesethree options as follows. A straightforward op-tion to work out Cliffe Hill Quarry completely,while towards the end of its life, develop-ing Stud Farm as its natural successor. Thisscheme had several drawbacks. The dimen-sions of both quarries would impose a ceilingon potential output. The extension of CliffeHill would require a road diversion and reloca-tion of the existing plant and unfortunately theonly feasible site for the latter would positionit closer to an adjacent village. This potentialplant site was also the only area suitable foroverburden disposal. Even if these problemscould be resolved for Cliffe Hill it would benecessary to install a second new plant whenStud Farm opened. This scheme was costedout on the basis of capital investment and thereturn it would yield. The main stumblingblock proved to be the cost of establishingtwo plants, each with a relatively short life.These would have been economically viableabove certain quarry outputs but neitherquarry could supply the required outputs on along-term basis.

The second option was the creation of aquarry at Stud Farm at an early date with a newprocessing plant and quarry. Once this wascommissioned the existing Cliffe Hill plantwould be demolished but crushing and screen-ing would continue using a mobile plant. An

Page 307: Introduction to mineral exploration

290 M.K.G. WHATELEY & W.L. BARRETT

Rai

lway

line

to

Bur

ton

West

Lane

Stud Farm Billa Barra Lane

Stanton Lane

Stantonunder

BardonThornton Lane

Cliffe HillQuarry

M1

500 m0

4847464544

Tunnel

N

11

10

FIG. 12.8 Plan showing the proposed development of the new quarry (on the left), the current workings inCliffe Hill Quarry, and the proposed field pattern and associated landscaping of the new quarry.

analysis of this proposed development indi-cated that a mobile operation of economic sizeat Cliffe Hill could only be located in one placeand this would result in a substantial steriliza-tion of reserves. There was also some doubtas to whether the mobile plant was capable ofproducing the different product types required.The financial analysis indicated that produc-tion costs using a mobile plant, which couldpossibly require replacement several times dur-ing the Cliffe Hill reserve extraction, combinedwith the additional cost of a new plant at StudFarm, made the financial return marginal.

The third option was the most radical. Thisenvisaged a new high capacity processing plantat Stud Farm which would be fed from boththe Stud Farm and Cliffe Hill Quarries (Figs12.2 & 12.8). The latter when extended wouldbe linked directly to the new plant by an under-ground tunnel (to ensure the least disturbanceto the inhabitants of Stanton under Bardon) andact as a satellite producer of primary crushedrock. Working both quarries simultaneouslycould supply considerably more output than asingle quarry. The necessary output levels alsomeant that each quarry would be operating at

an optimum level from the point of view ofproduction costs. This proposed scheme wasoriginally received by the management withsome scepticism, but when costed, it provedcapable of meeting the company’s financialinvestment and return criteria and was agreedto by all the relevant company experts beforebeing submitted to the local planning author-ity in October 1980. Comprehensive but posi-tive and wide-ranging discussions between thecompany and the Leicestershire county plan-ners followed before consent was granted inAugust 1983.

During the discussions it became clear thatthe third option favored by the company wouldnot be supported by the planners on the basisthat simultaneous working on both sides ofStanton under Bardon (Fig. 12.2) would be toodisruptive to the villagers. The planners didhowever insist that all mineable resources atboth sites should be recovered at some stage. Inthe light of these factors, the company optedfor developing a full-scale operation to workout the Stud Farm site completely, followingthe completion of which the remaining re-serves at the old quarry would be recovered via

Page 308: Introduction to mineral exploration

12: CLIFFE HILL QUARRY AGGREGATE RESERVES 291

some suitable connection, for processing atStud Farm.

Once planning permission had been granted,Tarmac undertook a full financial feasibilitystudy, after which further applications weresubmitted in 1986, both to permit an accelera-tion of the development of the new quarry, andfor the construction of a direct rail link fromthe site to the Leicester to Burton railway line(Fig. 12.2). Both proposals received planningconsent in the same year. Numerous controlsand legally binding agreements are associatedwith the planning consents, and were formul-ated in discussions between the planning au-thority and the company prior to the consentsbeing granted.

An approximate estimate of some of the ma-jor costs for Phase 3 (as defined above) under anumber of headings are tabulated in Table 12.1.

12.5 PHASE 4 – DETAILED ENGINEERING

The Phase 3 drilling, testing, and other evalua-tion exercises having proved, to an acceptablerisk level, the existence of an economicallyviable quality and quantity of material, themajor cost operations (collectively designatedas Phase 4) commenced in 1986.

In anticipation of the commencement ofmajor operations, extensive site investigationsurveys were carried out to provide the datafor the compilation of the statutory Mines &Quarries (Tips) Regulation 9 reports, withoutwhich bulk excavations and depositions on the

vast scales envisaged are prohibited in the UK.Such reports must be with the Inspectorateat least 30 days before the commencement ofthe operations to which they relate. The quarrydevelopment as a whole will necessitate theexcavation of several million cubic meters ofoverburden (down to a depth of 40 m) and thedisposal of this material into natural-lookinglandforms on the surrounding lands.

The major operations constituting Phase 4of the development were: further drilling andtesting, overburden removal, landform crea-tion, detailed site investigation for foundationdesign for the plant site (Barrett 1992), plantsite excavation, quarry development, blasting,road and rail access construction, the cultiva-tion and restoration of the completed over-burden disposal areas, and the erection of theplants and ancillary structures. A brief sum-mary of the Phase 4 progress on an annual basisis as follows.

In 1986 earthworks commenced at the newsite, with the clearing of hedges, the lifting andstorage of topsoil from approximately half thesite, and the excavation of around 1.6 × 106 m3

of overburden, mainly from the proposed plantsite at the western extremity. The excavatedmaterial was used to create a new landformon the southwestern perimeter of the site(Fig. 12.8). Two electricity powerlines werealso re-routed during this period.

In 1987 earthworks continued with thestripping of a further 1.6 × 106 m3 of overbur-den. This operation exposed bedrock over aworkable area and completed the excavation of

TABLE 12.1 Some major Phase 3costs. cost center Order of magnitude

of expenditure in1987–88 £ values

1 Exploration drilling to end of 1986 280,0002 Sample testing to the same date 25,0003 Geological and geophysical surveys 80,0004 Landscape design work 75,0005 Engineering design work and research 100,0006 Financial and administrative 75,000

considerations7 Compilation and submission of 20,000

planning application

Approximate total of major costs 655,000

Page 309: Introduction to mineral exploration

292 M.K.G. WHATELEY & W.L. BARRETT

Cost center Order of magnitudeof expenditure in1988–89 £ values

1 Overburden excavation and disposal 15,000,0002 Post Phase 3 drilling and testing 192,0003 Landscaping design, cultivation, 375,000

supervision4 Plant purchase, erection, 30,000,000

commissioning5 Road improvements and rail link 3000,0006 Quarry development and sundry 4000,000

other costs

Approximate total of major costs 52,567,000

TABLE 12.2 Some major Phase 4costs.

the plant area. The overburden was disposed ofalong the southwestern perimeter to completethis new landform. Some 8 ha of new landformwere cultivated and seeded, and rock removalto form the primary crusher slot and platformcommenced. Highway improvements wereundertaken, and erection of the plant and asso-ciated infrastructure commenced towardsthe year’s end.

During 1988 the erection of the processingplant, the asphalt and ready mixed concreteplants, the electrical substations, weigh-bridges, and offices was started and continuedthroughout the year. The development ofquarry benches and further overburden strip-ping was also carried out in 1988. The overbur-den was deposited to create a new landformalong the northern and western perimeterof the site. All completed landform sectionswere cultivated and seeded and tree plantingcommenced around the main entrance to thequarry.

In 1989 the erection of the plants and ancil-lary structures was essentially completed andplant commissioning commenced towards theend of the year. Further quarry developmentwas carried out to create the approximately200 m of 15-m-high faces that will be needed tosustain the anticipated production. The raillink was also completed and officially openedin October (Figs 12.2 & 12.8).

Cultivation and planting works continuedon the completed landforms and the CentralElectricity Generating Board commenced theerection of new pylons to enable the 132 kVpowerline to be re-routed in 1990.

1990 saw most of the planned workscompleted and several million tonnes of rockproduced. In the future further scheduled over-burden removal and bank building phases willbe required, and are incorporated with appro-priate safeguards in the planning consents.

Phase 4 costs are shown in Table 12.2.

12.6 QUARRYING AND ENVIRONMENTALIMPACT

The Stud Farm (now known as New Cliffe Hill)quarry is nearing the end of its life with annualproduction of 4.5 Mt of crushed aggregate, andextraction will probably be complete in 2005.As envisaged in the original plan, productionwill then move back to Old Cliffe Hill withprocessing at the New Cliffe Hill plant. In orderto accomplish this, the two quarries have beenlinked by a 725-m-long, 9 × 6 m tunnel. Thisshould enable production to continue until atleast 2024, subject to a review by the planningauthorities in 2007. The Cliffe Hill quarries arerun by Midland Quarry Products, a joint ven-ture of Tarmac with Hanson Quarry Products,formed in 1996 to enable more efficient work-ing of their reserves in the area.

The Cliffe Hill quarries are a long-term op-eration and have had to adapt to the changingpublic attitudes and social makeup of the localpopulation. During much of the twentieth cen-tury the impact of quarrying was overshad-owed by that of deep level coal mining in thenearby area. Since the cessation of coal miningin the early 1990s the area has become home to

Page 310: Introduction to mineral exploration

12: CLIFFE HILL QUARRY AGGREGATE RESERVES 293

residents commuting to nearby cities andhas benefited from government aid for urbanregeneration as well as the creation of a na-tional forest in the surrounding area. CliffeHill has adapted to these changes by making anumber of environmental initiatives and im-proving communication with the local popula-tion. In particular a quarry liaison committee,including representatives of local residents,local councils, and quarry management, meetsevery 3 months. The company also producesa 6-monthly newspaper to keep residentsinformed of activities at the quarries.

Although the existing operations are, atleast, tolerated by local residents it would bevery difficult to open a new quarry in the area.Any operation is likely to based on existingresources.

12.7 SUMMARY

The ultimate total expenditure to developa site from greenfield into a productive andprofitable quarry is high. The importance,therefore, of carrying out sufficient explorationand testing to guarantee as nearly as possiblethe predicted quality and quantity of mater-ial is essential. The rapidly increasing costof each phase (Table 12.3) emphasizes thenecessity for a logical approach, with eachsubsequent phase only being undertaken ifclearly warranted by the results from the pre-ceding one. In this particular case explorationcosts accounted for approximately 1.25% ofthe total costs of the development of thequarry operation.

Phase Cost in 1987–89£ values

Phase 1 – Order-of-magnitude study 1500Phase 2 – Pre-feasibility study 2500(auger drilling)Phase 3 – Feasibility study 650,000(coring, testing, etc.)Phase 4 – Detailed engineering 52,567,000

Total 53,221,000

TABLE 12.3 Summary of costs byphases.

Page 311: Introduction to mineral exploration

294 M.K.G. WHATELEY

13Soma Lignite Basin, Turkey

Michael K.G. Whateley

13.1 INTRODUCTION

The variables in a coal deposit (quality, thick-ness, etc.) are a function of the geological envir-onment. Excellent descriptions of all aspects ofcoal deposits and the assessment of these de-posits are given by Ward (1984), Scott (1987),and Whateley and Spears (1995). Coal or ligniteis a heterogenous, organoclastic sedimentaryrock mainly composed of lithified plant debris(Ward 1984), which was deposited in layers andmay have vertical and lateral facies changes.These changes reflect variations in vegetationtype, climate, clastic input, plant decomposi-tion rates, structural setting, water table, etc.Coal originated as a wet spongy peat, whichafter burial underwent compaction anddiagenesis (coalification). The coalificationprocess proceeds at different rates in differentstructural settings. Coal which has been sub-jected to low pressure and temperature is re-ferred to as a low rank coal, such as lignite.Coal which has been subjected to high pressureand temperature is referred to as a high rankcoal, such as anthracite. Bituminous coal isa medium rank coal. Thus the properties ofcoal are almost entirely a reflection of theoriginal depositional environment and dia-genetic history.

Some of the properties of coal depend onthe nature of the original plant material ormacerals (Stach 1982, Cohen et al. 1987). Theseproperties are measured using microscopictechniques. Large scale, subsurface changesare determined by studying the lithology of thecoal and coal-bearing strata in borehole coresand down-the-hole geophysical logs. These

data form the basis for any investigation intothe resource potential of a coal deposit.

The quality of a coal deposit must beassessed. This is undertaken by sampling out-crops, borehole cores (Whateley 1992), minefaces, etc. (section 10.1.3), and sending the coalto the laboratory. The coal is subjected toproximate analyses to determine the moisture,ash, volatile, and fixed carbon contents (Ward1984). Additional tests which are often re-quested are sulfur content, calorific value (CV),and specific gravity (SG). The quality of coalwill determine the end use to which it is put,e.g. steam coal to be burnt in a power stationto generate electricity or metallurgical coal forsteel making, although other uses are possible.

Since the energy crisis caused by the OPECoil price rises which started in 1973, there hasbeen a worldwide increase in the search foralternative fossil fuels. Lignite, bituminouscoal, and anthracite are important sources ofenergy. The exploration for and exploitation ofthese fuels is important for developing coun-tries which wish to reduce their reliance onimported oil by building coal-fired powerstations at or near their own coal mines.

In 1982 Golder Associates (UK) Ltd (GA)undertook a World Bank funded feasibilitystudy of the Soma Isiklar lignite deposit inwestern Turkey (Golder Associates 1983) forthe state-owned coal company Turkiye KomurIsletmeleri Kurumu (TKI). The lignite from theSoma Mine was used as feedstock for the age-ing 44 MW Soma A Power Station. It wasthe intention to increase the mine productionfrom 1.0 to 2.5 Mt yr−1, 2.0 Mt of which wereto come from the expanded surface mine. The

Page 312: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 295

increased tonnage was needed to feed thenewly constucted 660 MW Soma B Power Sta-tion. Excess lignite would be used for domesticand industrial purposes.

The GA report deals with the geologicalinterpretation of the deposit, assessment ofthe reserves in terms of tonnage and quality,assessment of the geotechnical aspects of themine area, and production of mine plans for theproposed surface and underground mines. Mostof this chapter is derived from that report, al-though the geostatistical reserve estimationshave been subsequently completed at LeicesterUniversity as student dissertations (Lebrun1987, Zarraq 1987). The existing surface minehad suffered a major footwall failure in 1980.The steeply dipping, shaly, footwall sedimentsunderwent noncircular failure on weak hori-zons over several hundred meters along strikeand buried the working face. One of the inten-tions of the study was to establish the reasonsfor this failure and to incorporate safety factorsinto the new mine design that would reducethe risk of a repeat of this type of failure.

It is the intention in this chapter to outlinethe way in which the geologists who were in-volved on this study collected the data, pre-sented the data for the mining and geotechnicalengineers, and calculated the reserve and qual-ity parameters. The geotechnical and miningaspects are also described. To simplify thischapter only the thick, lowermost seam in thearea, designated the open pit, is described. GAalso reported on the underground mining po-tential of the Soma Basin, which they reportedmight be economically mineable in the future.

13.1.1 Location

The Soma lignite deposit is in the Manisa Prov-ince of western Turkey, 10 km south of the

town of Soma (Fig. 13.1). A high (1100 m) ridgeseparates the proposed mine site from thetown. The site is on the south-facing slope ofthis ridge, with elevations ranging from 750 min the north to 310 m in the south.

13.1.2 Turkish mining rights

Mining and operating rights vary in differentparts of the world. Before any foreign or localcompany can start operations it is necessary tounderstand the local legal system. In Turkey allmineral and coal rights are deemed to be ownedby the State and are not considered to be part ofthe land where they are found. The Ministry ofEnergy and Natural Resources administers andimplements the mining laws and regulations,and grants exploration permits, exploitationpermits, and leases.

13.2 EXPLORATION PROGRAMS

13.2.1 Previous work

The lignite in the Soma area has been minedsince 1913, first for local domestic and indus-trial consumption and later as feed for theSoma A power station. TKI took over ligniteproduction on 1979.

The database for this deposit was derivedfrom four drilling programs (Table 13.1), fieldmapping, and a feasibility study conductedby TKI in 1981–82. The earliest study wasconducted by Nebert (1978), which includedlithological logs of 34 cored boreholes (Table13.1) and two geological maps of the Somaarea with cross-sections. Cored lignite wasanalyzed for moisture and ash contents, calo-rific value, and in some cases volatile matterand sulfur content. The holes were drilled at

Drilling Number Total Drillingprogram of holes meterage date

200 34 8790 1960100 50 10,825 1976300 9 1547 1981400 29 8631 1982

Total 122 29,793

TABLE 13.1 Summary of drillingexploration programs.

Page 313: Introduction to mineral exploration

296 M.K.G. WHATELEY

N

T U R K E Y

USSR

IRAN

IRAQ

SYRIA

Ankara

Mediterranean Sea

Black SeaIstanbul

STUDY AREA

Antalya

Soma

To powerstation

To Izmir

OPEN PITOPEN PIT

ISIKLARCONCESSION

Isiklar Stream

1000

750

500

500

750

250

surface contour(in meters)500

0 1 km

Izmir

39,000N

38,000N

37,000N

36,000N

35,000N

34,000N

33,000N

32,000N

31,000N

30,000N

29,000N

48,000E 49,000E 50,000E 51,000E 52,000E 53,000E

750

500

750

1000

SOMA

FIG. 13.1 Location diagram ofthe Soma study area.

Page 314: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 297

500-m centers to establish the resource poten-tial of the basin.

The second exploration program was super-vised by Otto Gold GmbH. Open-hole drillingwas used with predetermined intervals coredto include lignite seams. Lignite cores wereanalyzed for ash and moisture content, andcalorific value. The sulfur and volatile contentwere determined on a few samples. The pro-gram was designed to identify lateral anddown-dip lignite limits, as well as to undertakeinfill drilling at 250 m centers within the basin.

TKI initiated the third drilling program aspart of their preliminary mine feasibility study.They also used rotary drilling techniques withspot coring. Full proximate analyses and calo-rific value determinations were done on thesecores, and occasionally sulfur analyses. Theseholes were drilled to obtain additional informa-tion in areas of structural complexity or areaswith a paucity of data. TKI produced isopach,structure contour, isoquality, and polygonalreserve maps, as well as reserve tables, surfaceand underground mine plans, manpower sched-ules, etc.

A fourth drilling program was carried outduring the feasibility study. Five of the holeswere fully cored for geotechnical studies,while the remainder were rotary drilled andspot cored. Lignite cores were analyzed forproximates, calorific value, and specific grav-ity. Hardgrove grindability tests (section 13.6)and size analyses were also carried out onsamples from the existing open pit. Although122 boreholes were drilled, only 83 of themintersected the lowest lignite seam, or the hori-zon at which the equivalent of the lignite seamoccurred. This was due to common drillingproblems such as loss or sticking of rods in thehole, or burning the bit in at zones of seriousand sudden water loss. This resulted in an over-all density of 11 holes km−2, equivalent to a rec-tangular grid roughly 330 m × 250 m, althoughsome holes are closer and some farther apartthan this. This is considered to have given suf-ficient density to classify the lignite in termsof measured (= proved) reserves (USGS 1976).

13.2.2 Core recovery

One of the major problems of assessing ligniteor coal deposits is that of core recovery. If core

is lost during drilling there is no way of deter-mining the quality of the lost core and often thebetter quality, more brittle, bright sections arelost on these occasions. It is usual to have adrilling contract that requires the drillers torecover at least 95% of the core in the seam.Recoveries less than this (within reason) usu-ally require redrilling. Techniques that helpto improve core recoveries include the use oflarge diameter wireline (see section 10.3.2)or air flush core barrels (e.g. HQ series witha nominal hole diameter of 96.1 mm) and theuse of triple tube core barrels (Cummings &Wickland 1985, Berkman 2001).

13.2.3 Geophysical logging

Down-the-hole (DTH) geophysical logging (seesection 7.13) was first used in the oil industrybut soon found its way into coal exploration(Ellis 1988). It is now used routinely in theevaluation of coal deposits because geophysicallogs can help reduce drilling costs by enablingthe use of cheaper, rotary, open-hole drillingof, say, 80% of the holes. Good comparisonof open hole and cored holes is achieved withthe use of DTH logs. They can also helpin identifying the top and bottom of theseams, partings within the seams, lithologicalchanges, in checking on core recovery and thedepth of each hole, and ensuring that drillers donot claim for more meters than they actuallydrilled. Seams often have characteristic geo-physical signatures which, in structurally com-plex areas, can help with seam correlation.

Typical DTH geophysical logs used on coaland lignite exploration programs are naturalgamma, density, neutron, caliper, and resistiv-ity, as well as sonic and slim line dip meter logs(Ellis 1988). Shale, mudstone, and marl usuallyhave a high natural gamma response while coalhas a low response, with sharp contacts oftenbeing observed. Coarsening-upward or fining-upward sequences in the clastic sections of thelogs can also be inferred from the gamma logs.The density log, as its name implies, reflectsthe change in density of the rocks, with coaland lignite having low densities and shale andsandstone having higher densities. The neu-tron tool is used in estimating porosity, and theresistivity tool may be used to indicate bedboundaries. The caliper tool defines the size of

Page 315: Introduction to mineral exploration

298 M.K.G. WHATELEY

the hole. The sonic log is also used in estimat-ing the rock quality (fracture frequency), andthe dip meter can be used to interpret sedi-mentary structures (Selley 1989).

A geophysical log showing the typical re-sponses that the rocks in the Soma basin give,is described in section 13.3.2.

13.2.4 Sampling

It is essential to establish a procedure forsampling core and working faces so that conti-nuity is maintained throughout an explorationproject and into the mining phase. For example,at Soma the whole lignite sequence wassampled, including all parting material. Thecore was split and one half was retained inthe field. All lithological layers in the seamgreater than 30 cm thick were sampled. Thin-ner layers were included with adjacent layersuntil a minimum thickness of 30 cm wasobtained. The 30 cm limit was used as it wasconsidered by the study team that this was theminimum thickness of parting that could bemined as waste in the open pit (see alsosections 13.6.1 to 13.6.3). This ensured thatweighted average estimates for run-of-mine(ROM) lignite could be made.

13.2.5 Grouting

Where deep coal is likely to be mined by under-ground methods, all holes drilled prior to min-ing should be sealed using pressure grouting.This will reduce the potential water inrushhazard in underground mining.

13.3 GEOLOGY

13.3.1 Geological setting

The basement in western Turkey consists ofPrecambrian, Palaeozoic, and Mesozoic sedi-mentary and igneous rock (Campbell 1971,Brinkmann 1976) which have been subjected tovarious structural and metamorphic episodes.Turkey has a great variety of structures, thelargest of which is the North Anatolian Fault(NAF), a major strike-slip fault system (Fig.13.2). The NAF was formed in the lateSerravallian (Sengor et al. 1985). At this time

westerly strike-slip movement of the westAnatolian Extensional Province took place. Atthe same time a series of NE-trending grabens,such as the Soma Graben, began to form inwestern Turkey. These grabens began to fillwith Serravallian (Samartian) sediments, whichcontain thick lignite deposits.

13.3.2 Geology of the Soma Basin

The Soma Basin contains thick deposits ofMiocene and Pliocene sediments (Fig. 13.3),which range in age from Serravallian(Samartian) to Pontian (Gökçen 1982), but donot contain volcanic rocks. The Tertiarysediments rest unconformably on Mesozoicbasement rocks. The stratigraphy of the basinis summarized in Fig. 13.4. It contains twothick Miocene lignite seams designated theKM2 and KM3, but this case history deals onlywith the KM2 seam.

The Mesozoic basement forms rugged topo-graphy around the basin of limestone hillsrising to 350 m above the Tertiary sediments.Boreholes which reached the basement con-firmed the presence of limestone below thesediments in the basin. At Soma, the top ofthe basement appears to consist of debris flowdeposits.

The Miocene deposits have been divided intothree formations (Fig. 13.4). The basal TurgutFormation is predominantly immature sand-stone and conglomerate, but there is a grada-tional upward fining of the sediments to thelignite horizon. Increasing amounts of sandyclay, clayey silt, and carbonaceous clay appeartowards the top of the formation until thesegrade into the KM2 seam. The basal contact ofthe seam is gradational and is usually placedwhere the first recognizable lignite appearswith less than 55% ash content. The lignite ishard, black, and bright, has numerous cleats(typical close spaced jointing of coal andlignite), and breaks with a concoidal fracture.Using the ASTM classification, the calorificvalue of this material would more properlyresult in it being termed a sub-bituminous Bcoal, but it is referred to as a lignite in thelocal terminology.

There is an extremely sharp contact betweenthe KM2 and the overlying Sekköy Formation.This formation is a massive, hard marlstone

Page 316: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 299

shows a distinct change from the very high sig-nature of the marl to a medium response in thelimestone. The Yatagan Formation containsthe KM3 Lignite which is interbedded with aseries of thick calcareous clastic units, makingthe economic assessment of this unit moredifficult. The freshwater limestone is con-formably overlain by massive marlstone thebasal half of which contains thin, laterallyimpersistent, lignite seams. Gökçen (1982)considered these marlstones to belong to the

T U R K E Y

l

l

l

l

North Anatolian Fault

Bergama

KinikSTUDYAREA

EYNEZBASIN

Soma

DENISBASIN

Akhisar

28°00'E

39°00'S

27°30'E

Istanbul

AnkaraSoma

+

N

0 10 km

SOMABASIN

Quaternary

Neogene

Cretaceous

Jurassic Limestone

Palaeozoic basement

Granodiorite

Andesite

Ophiolite

+

FIG. 13.2 Generalized geology of the Soma area (taken from the 1:250,000 Izmir sheet regional geological mappublished by MTA). The inset shows the regional structural setting for the Soma Basin in relation to the majorstructural feature of northern Turkey, the North Anatolian Fault (NAF).

and geophysical logs identify this contactclearly. The KM2 has a distinctly low naturalgamma signature in contrast to the very highsignature of the marl. The sharp change frompeat (lignite) formation to the marlstone isbelieved to be the result of a change to an aridclimate (Sengor et al. 1985).

In cores and in the open pit, the contactwith the overlying freshwater limestone ofthe Yatagan Formation is unconformable andmarked by a color change. The gamma log

Page 317: Introduction to mineral exploration

300 M.K.G. WHATELEY

Upper Miocene Yatagan Formation, althoughGolder Associates (1983) placed the Miocene–Pliocene contact at the top of the freshwaterlimestone.

13.3.3 Structure of the Soma Basin

The Soma Basin is a fault-controlled graben,which has been subjected to regional tilting tothe southwest. The basin formed in response tostike-slip faulting along the North AnatolianFault Zone during the mid Miocene (Sengor etal. 1985). The faulting in the Soma Basin trendsNE–SW (Fig. 13.3). The major faults were de-fined by elevation differences in the structurecontour map of the top of the KM2 seam. Itis generally accepted that a disruption to thetrend of a set of structural contours, that areexpected to behave in a uniform way, mayindicate the presence of a fault (Annels 1991,Gribble 1993, 1994). Fault positions were inter-polated as lying between certain boreholes and,

as mining proceeds and more informationbecomes available, then the position ofthese faults will be defined more accurately.These faults have throws of up to 150 m anddisplacements of up to 300 m.

Faults of this size will affect the designof both surface and underground mines. Thefaults at Soma were used to define miningblocks, Blocks A, B, C, D, and E (Fig. 13.8).What is more difficult to determine is theamount of secondary faulting and fracturingformed in association with the major faulting.These minor, subparallel faults may havethrows of only a few meters. It is difficult toplot these faults from drilling results, but theymay have a significant effect on undergroundmining operations.

The sediments were deposited around thearcuate northern basin rim, and they dip to thesouthwest at an average of 20 degrees. Thesedips vary within each block, e.g. near the north-ern rim of the basin the dips are significantly

C

C′P1

P1

P1P1

P1 P1

P2

M3

M3 M2M1

P2

P2

M3 M2M3

P2 P2

M2 M2

M3

P2

M112

22

14

11

15

14

18

29

18

18

11

34

24 60

2236

15

18

23

40

0 1 km

Subcrop of KM2 seam

Dip and Strike15

Fault

P1Pliocene

Miocene

Lignite

Lignite

KM3

M3

M2

KM2

M1

P2

Area where KM2 ligniteis not developed

Basement

Outcrop of KM2 seam

N

32,000N

31,000N

30,000N

48,000E 49,000E 50,000E 51,000E 52,000E

CONCESSION BOUNDARY

FIG. 13.3 Simplified geological map of the Soma Basin. The detail of the limit of the lignite in the south-south-west is unclear as no information was available due to the great depth of the lignite. Block numbers refer tothe proposed mining blocks used in reserve and quality estimation procedures (see text).

Page 318: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 301

l

l

l

l

l

l ll

l

l

l

ll l

l l

l

l

l

l

l

10–4

05–

405–

140

90–2

5087

–170

96–3

000–

200–

23TH

ICKN

ESS

(m)

LITHOLOGY DESCRIPTION

Waste dumps

Talus

Marlstone

Lignite lenses (KP1)

Lignite (KM3)

Marlstone

Lignite and carbonaceous shale(KM2 member)

Conglomerate, sandstonesiltstone, and carbonaceous shale

Limestone

TURGUT(M1)

SEKKÖY(M2)

P1

P 2

FORMATION

GOLDERASSOCIATES

1983

GÖKÇEN1982

LITH

. U

NIT

AG

E

N3–

B3

N3–

B4

N3–

B2

N3–

B1

SA

RM

ATIA

NPA

NN

ON

IAN

PO

NTI

AN

ME

SO

ZOIC

MIO

CE

NE

PLI

OC

EN

EM

IOC

EN

E

FOR

M.

NYA

TAG

AN

TUR

GU

TS

EK

Y

Lignite (KP2)

Freshwater limestone

YATAGAN(M3)

Unconformity

FIG. 13.4 A simplified stratigraphical column for the Soma Basin. The sediments have not been dated with anycertainty. Comparison of traditional and recent microfossil dating (Gökçen personal communication) showsthat more detailed work is still required.

Page 319: Introduction to mineral exploration

302 M.K.G. WHATELEY

steeper. A more accurate assessment of the dipswas made by interpreting the structure contourmap of the top of the KM2 seam (Section 13.4).

13.3.4 Depositional model for the SomaBasin

A depositional model of the Soma Basin wasproposed after the isopach and ash contentmaps of the KM2 seam, detailed borehole logs,and stratigraphical sections were examined andinterpreted in the light of ancient and modernanalogs, in the manner described by McCabe(1984, 1987, 1991). The basin originated as aNE–SW fault-controlled graben in the base-ment rocks, which gradually infilled as thebasin subsided at varying rates. Steep peri-pheral gradients around the north of the basinat the time of initial deposition resulted inrapid transportation of sand and gravel tothe center of the basin by high energy runoff.These debris flow deposits alternate withpoorly sorted, coarse-grained sandstone.Gradual lessening of the gradient, with subse-quent lowering of the energy regime, resulted

in sediment load fractionation and depositionof finer sediments in the basin (sandstone andsiltstone). The overall result is a fining-upwardsequence from gravelly sandstone at the base tosiltstone and mudstone at the top (Fig. 13.4).Temporal variations in depositional conditionsare indicated by the repetition of these fining-upward sequences and rapid lateral variation.In the partially filled basin, conditions wereconducive to plant growth which resulted inthe formation of the thick, laterally variableKM2 lignite seam. The lignite is generallyshaly at the base but improves in quality (i.e.has a lower ash content) towards the top.

Two distinct sub-basins have been recog-nized, the eastern Demir Basin and the westernSeri Basin, separated by the Elmcik High(Fig. 13.5). The plant-forming ecosystem ormire (Moore 1987) formed around the northernrim of the basin, but carbonaceous materialwas transported to the more distal portions ofthe basin to the south and southwest. Fine-grained sediment was also transported into thebasin, and deposited mainly along the northernrim of the Seri Basin, where the lignite is

N

OPEN PIT

OPEN PIT

Basement

SERI BASIN

DEMIR BASIN

High ashlignite

High ashlignite

Good qualitylignite

Shale Highlybandedlignite

??

??

Shale

Shale

20

10

20

20

ELMACIKHIGH

40

30

30

32,000N

31,000N

30,000N

48,000E 49,000E 50,000E 51,000E 52,000E

0 1 km

Approximate seamisopach in10 m intervalsOutcrop of KM2 seam

Approximate position of lakemargin during KM2 deposition

10

Facies change

Facies change

4030

FIG. 13.5 A palaeogeographical interpretation of the Soma Basin at the time that the KM2 seam was beingformed.

Page 320: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 303

strongly banded having many nonligniticpartings. The overall result is a gradual facieschange from low ash lignite to carbonaceousshale basinward. There is a pronounced thin-ning of the lignite over the Elmacik High, sug-gesting that this area was an active high duringpeat formation.

A climatic change from warm humid condi-tions favoring peat formation to an arid climateresulted in a sudden and basin-wide changeof sedimentation to marl deposition. The finegrain size and the calcareous nature of the ma-terial suggests deposition by low energy inputinto a low energy water body. The resultingmarlstone (Sekköy Formation) is unconform-ably overlain by the freshwater limestoneof the Yatagan Formation. The KM3 ligniteoccurs at varying levels within the limestone,but is areally restricted. The KM3 has a sig-nificantly higher ash content than the KM2lignite. Alternating marlstone and limestonecontinues upwards, with occasional thin, later-ally impersistent seams developing.

13.4 DATA ASSESSMENT

In order to establish the reserve and quality cri-teria that were required for the mine design andfinancial analysis of the project, several mapshad to be constructed, such as structure con-tour maps, isopach and isoquality maps, as wellas stripping ratio maps and cross-sections. Allthe borehole logs were examined, the seamswere correlated, and the top and bottom of theKM2 seam were established. Lignitic materialwas frequently rejected from the base of theseam because of poor quality. The top of theseam was readily identified at the sharp contactof the lignite with the overlying marlstone, butthe basal contact was often drawn only afterthe quality data had been assessed. Once theseam thickness was established it was possibleto construct the maps.

Since this study was completed several inte-grated computer packages have become avail-able that calculate and plot contours quicklyand accurately (Whateley 1991), e.g. Datamine,Surpac, Borsurv, PCExplore, etc.). During thisstudy a planimeter was used to measure areas(e.g. between contours for tonnage calculations(sections 13.4.2 & 13.7)). Modern software

packages have various user-definable methodsof contour construction. Two contoured sur-faces (or digital terrain models (DTMs)) canbe superimposed and volumes calculated, e.g.DTMs of the top and bottom of a seam, orDTMs of the surface and the top of a seam foroverburden volume. The assessment of all thedata up to and including the mine designand the financial analysis can now be under-taken using these computer packages. Manualmethods (described in section 10.4) are stillwidely used by many exploration and miningcompanies.

13.4.1 Structure contour maps

Three structure contour maps were drawn forthis study: (i) on the top of the KM2 seam; (ii)on the bottom of the KM2 seam; and (iii) on thetop of the Mesozoic Basement. The KM2 seamhas the greatest areal distribution, a very welldefined upper contact, and the most detailedand reliable database. The elevation of the topof the KM2 seam was plotted manually at eachdata point and contours were constructed bylinear interpolation between points. The con-tours were drawn at 20-m intervals (Fig. 13.6).The strike of the lignite seam at depth was as-sumed to be similar to that of the overlyingsediments at the surface. Where the structurecontours differed significantly from this, a faultwas inferred. By checking against the surfacegeological map and cross-sections it was poss-ible to establish the fault pattern in the basin.The structure contours were then redrawnbetween the faults to give the pattern shown(Fig. 13.6). It was important to determine thestructure of the basin in order to assist the min-ing engineers with their design of the optimumopen pit, by avoiding areas of unstable groundnear faults and loss of lignite near these faultzones.

Once the structural pattern was established,it could be transferred to the structure maps ofthe base of the KM2 and the top of the base-ment. The structure contour map of the base ofthe KM2 was constructed by placing theisopach map of the KM2 on top of the structurecontour map of the top of the KM2. Where thetwo sets of contours intersected, the seamthickness was subtracted from the top of seamelevation to give the new elevations. These

Page 321: Introduction to mineral exploration

304 M.K.G. WHATELEY

contours were drawn at 20-m intervals, andwere only drawn in the area designed to be theopen pit, for use in mine design and to identifyareas of potential slope failure.

The structure contour map of the top of thebasement was also produced by linear inter-polation. Only a small number of boreholespenetrated the basement, so the contours wereconstructed at 50-m intervals. The structurecontour map of the top of the KM2 was thenplaced on the basement map and then the inter-sections of contours of the same value outlinedthe limits of the KM2 seam (Fig. 13.3).

13.4.2 Isopach maps

For this study, three isopach maps were drawnmanually: (i) for the KM2 seam (Fig. 13.7); (ii)for the overburden material; and (iii) for thesediments between the KM2 seam and thebasement (the Turgut Formation). The thick-ness of the KM2 was determined in each hole

between the sharp upper contact with the over-lying marlstone and the base of the seam. Theisopachs show the thickness of KM2 that couldbe extracted by surface and undergroundmining methods, and include waste partingswithin the lignite which could be separated aswaste during mining (section 13.4.4). Theisopach map was constructed at 5-m intervalsby linear interpolation. The seam ranges inthickness from 2 to 57 m, with an average inthe open pit area of some 17 m. The extremethickness of 57 m is the result of a boreholeintersecting a steeply dipping part of the seam.The seam isopach map formed the basis onwhich the reserve estimates were based, as wellas being used to calculate the structure con-tours at the base of the KM2 seam.

The isopach map of the waste material abovethe KM2 seam (the overburden) was con-structed by placing the map of the surfacetopography over the structure contour map ofthe top of the KM2 seam. Where the respective

C

C′

500

400

400

300

200

260

160120 120

400300

500

400

300200

100

0

400

300

200

100

0

100

500

400

300

200

100

0

100

32,000N

31,000N

30,000N

48,000E 49,000E 50,000E 51,000E 52,000E

Basement

200

Subcrop of KM2 seam

Outcrop of KM2 seam

Area where KM2 ligniteis not developed

Fault

Structural contour(20 m interval)

0 1 km

CONCESSION BOUNDARY

N

FIG. 13.6 A structure contour map of the top of the KM2 seam, Soma Basin.

Page 322: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 305

N

l

l

l

l

l

l

l

l

l

l

l l

l

l

ll

l

l

l

l

l

ll

l

l

l

l

l

l

ll

l

l

l

l

l

l

ll

l

l

l

l

l

l

ll

l

ll

l

l

l

l

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll l l

l

l

ll

ll

l

l

ll

ll

l

lll

ll

l

515

10

1510

10

15

20

25

1510

20 15

2530

3025

201510

255

30

20253035

10

20

l

l

l

ll l

l

32,000N

31,000N

30,000N

48,000E 49,000E 50,000E 51,000E 52,000E

29,000N

0 1 km

ISIKLAR CONCESSION BOUNDARY

OPEN PIT

OPEN PIT

10

ll

Isopach in 5 m intervals

Borehole site

Subcrop of KM2against basement

Outcrop of KM2 seam

Fault

Basement

Area where KM2 ligniteis not developed

10

5

FIG. 13.7 An isopach map of the KM2 seam. Location of all the boreholes drilled in the Soma Basin arealso shown.

contours intersected, the elevation of the top ofthe seam was subtracted from the elevation ofthe topographical contour and the resultingthickness value plotted. This method of con-structing the overburden map provided moredata points than would have been available hadthe borehole values alone been used, resultingin a more reliable map. The overburden abovethe KM2 seam varies from a few meters nearthe existing workings, to approximately 150 mat the deepest part of the proposed surfacemine. The map was used to construct the strip-ping ratio map needed in the mine planningexercise.

A detailed contour map was constructedduring the mine design phase of the project foruse by the mining engineers to calculate thevolume of overburden material which had tobe removed at various stages of mining fromthe various mining blocks (see section 13.3.3).Contours were constructed and a planimeterwas used to measure the area between adjac-ent overburden thickness contours whichwere then multiplied by the applicable vertical

overburden thickness. Products were summedto obtain the total volume for each block(Table 13.2). Pit slope volumes were calculatedin a similar manner assuming a final pit slopeangle of 45 degrees.

An isopach map of the Turgut Formation wasconstructed by subtracting the values of thestructure contours at the base of the KM2 seamfrom the structure contours at the top of thebasement. This isopach map was only drawnfor areas of potential open pit development toassist the geotechnical engineers in predictingthe areas of potential footwall failure in futureoperations (see section 13.5.1).

13.4.3 Stripping ratio maps

The stripping ratio map, or overburden ratiomap shows the overburden material in m3 as aratio per tonne of lignite, prior to applying se-lective mining criteria. This map (Fig. 13.8) wasconstructed by placing the overburden isopachmap on the isopach map of the lignite. Wherethe contours intersected, the thickness of the

Page 323: Introduction to mineral exploration

306 M.K.G. WHATELEY

TABLE 13.2 Estimated overburden volumes at the proposed Soma Isiklar open pit mine. For explanation ofM bank m3 see section 13.8.2.

Block Pit area Slopes Totalvolume

Area Thickness Volume Area Thickness Volume (M bank m3)(km2) (m) (M bank m3) (km2) (m) (M bank m3)

A 0.146 78.9 11.52 0.148 62.7 9.28 20.80B 0.412 90.1 37.12 0.189 54.6 10.32 47.44D 0.072 87.6 6.31 0.096 60.3 5.79 12.10E 0.882 84.8 74.79 0.357 70.1 25.24 100.03

Totals (mean) 1.512 (85.8) 129.74 0.790 (64.1) 50.63 180.37

2

4

68

10

13

5

7

9

1

3 5

3

5

7

35

1

3

5

79

C

C′

BLOCKE

BLOCKD

BLOCKC

BLOCKB

BLOCKA

0 1 km

N

150 m overburden isopach

Basement

Outcrop of KM2 seam

Area where KM2 ligniteis not developed

Fault

Subcrop of KM2 seam

Stripping ratio (m3: t)2

32,000N

31,000N

30,000N

48,000E 49,000E 50,000E 51,000E 52,000E

CONCESSION BOUNDARY

FIG. 13.8 A stripping ratio map showing the 150 m overburden isopach line used to determine the down-diplimit of the open pit mine.

The stripping ratio map provides a guide toareas where mining would be preferentiallystarted, i.e. where the ratios are lowest. Thefinancial study suggested that open pit min-ing could be carried out economically wherethe stripping ratio is less than 7:1. As a generalguide, the 7:1 stripping ratio in coal or lignitemining provides a limit to surface mining,

lignite was multiplied by a specific gravity of1.73 to obtain tonnes. The specific gravity wasobtained by averaging the values derived fromlaboratory analyses of lignite from the open pitarea of the mine. The product was divided intothe overburden thickness to calculate the strip-ping ratio. Contours were then constructed atunit intervals between 1 and 10.

Page 324: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 307

Ele

vatio

n ab

ove

sea

leve

l (m

)

Existinggroundsurface

Main haulage level

Old undergroundworkings

C′C

152152A 19 18 17 109 106

210A

M2

M3

M3

TM

P1

P1P2

KM2

KM3

M1

Distance (m)

49,0

00E

30,0

00N

30,5

00N

49,5

00E

31,0

00N

50,0

00E

31,5

00N

For rock types see Fig. 13.4

Ground water levellower than level shown

Artesian flow from borehole

Ground water level

0 200 400 600 800 1000 1200 1400 1600 1800 2000

700

600

500

400

300

200

100

0

SW NE

M1M2

M2

M2

FIG. 13.9 An example of a down-dip section used to help in correlating the stratigraphy and to illustrate thehydrogeological information derived from boreholes.

although different companies operate differentcriteria and this figure is variable.

13.4.4 Cross-sections

East–west, north–south and down-dip sec-tions (Fig. 13.9) were drawn at regular intervalsacross the basin. They were constructed byplotting the data from the structure contourand isopach maps. Boreholes which fell onor close to the section lines were also plotted.The east–west and north–south sections werecross-checked by comparing and correlatingthe intersection points. The sections wereused to verify the structural information andthe seam correlations. They were not used forreserve estimation (Reedman 1979), althoughthis is commonly done by some coal miningcompanies.

13.5 GEOTECHNICAL INVESTIGATION

13.5.1 Investigation program

It is always advisable to initiate geotechnicalinvestigations as soon as possible in an explora-tion program in order to avoid expensive

duplicate drilling of boreholes. In this study thefollowing program was carried out:1 Geotechnical logging of five fully coredboreholes. In addition to the geological de-scriptions, the logs included the rock qualitydesignation (RQD) (section 9.6) and evaluationof the point load strength (Brown 1981), whichcan be used to estimate the uniaxial com-pressive strength. Where a point load strengthtest was not available, the field descriptionwas used as a basis for estimating the strengthclass (see Table 10.8). All the cores were fullylabeled and photographed (Fig. 13.10). This isparticularly important where cores are prone todeterioration.2 Geotechnical testing of borehole cores andsamples collected from the existing open pit.3 Structural mapping of overburden exposuresin the existing open pit.4 Measurement of water levels in explorationboreholes and measurement of flows fromartesian boreholes.5 Sensitivity analyses of known failuresaltering the variables such as position ofthe piezometric surface or material strengthto establish typical shear strength values fordesign from back analysis of existing footwallfailures in the open pit.

Page 325: Introduction to mineral exploration

308 M.K.G. WHATELEY

FIG. 13.10 A photograph of a box of core drilled during this study. All core recovered during this study wasphotographed before the core had been tested and sampled by the geotechnical engineers and before the lignitehad been sampled. This ensured that a full record of the in situ core was available. All core box lids had adetailed descriptive label attached which gave the borehole number, box number, depths, date drilled, and ablack and white and color chart to ensure that each film was processed to the same standard.

These data were used in geotechnical ana-lyses to provide preliminary design guidelinesfor the proposed open pit.

13.5.2 Geotechnical conditions

Overburden

The overburden consists predominantly oflimestone and marlstone with lesser amountsof mudstone and siltstone. The limestone andmarlstone are moderately strong, with meanuniaxial compressive strengths of the orderof 80 MPa. The siltstone and mudstone areweaker rocks with compressive strengthsgenerally less than 30 MPa. The overburdenrocks are bedded and jointed and the bedding isgenerally parallel to that in the lignite seam

and dips range between 15 and 45 degrees to thesouthwest. Jointing is mainly vertical.

Lignite

Point load strengths of selected lumps of ligniteare about 1 MPa. However, the lignite is highlycleated and friable and the mass strength wouldbe considerably lower.

Footwall

The footwall of the KM2 seam consists ofvariable thicknesses of mudstone, siltstone andimmature sandstone overlying basement lime-stone. The immediate footwall has a high claycontent and is generally weak, with a high slak-ing potential. Typical residual shear strengths

Page 326: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 309

for this material are c = 0–30 kPa, and Π =14–17 degrees, where c = shear strength andΠ = angle of shear resistance. A series of backanalyses was carried out on exposed footwallmaterial in the existing open pit. These ana-lyses confirmed that the above shear strengthvalues were realistic for design purposes, andthat footwall stability is highly dependent ongroundwater conditions and the local dip of thefootwall.

13.5.3 Hydrogeological conditions

Standing water levels in exploration boreholeswere measured using piezometers. The levelswere generally close to the surface and, in thetopographically lower areas, flowing artesianboreholes occurred. In some cases, artesianflows were high (7 L s−1) and flows could bemaintained for long periods. One of the mainaquifers in the strata appeared to lie in thevicinity of the KM2 seam.

13.5.4 Implications for open pit mining

The overburden rocks are moderately strong,requiring drilling and blasting prior to excava-tion. Controlled blasting procedures would berequired on all final walls for good slope stabil-ity. Practical mining considerations wouldlimit the maximum wall slope to about 55 de-grees between haulroads and an overall slope of45 degrees for the final highwall of the pit, in-cluding provision for 20-m-wide internal haulroads. The overall slope angle of the advancingface is determined by the equipment clearanceand access requirements but should not exceed40 degrees for stability and safety. This couldbe steepened to 45 degrees at the final limitwhen controlled blasting procedures are used.The footwall dip means that footwall failurescould occur in the new pit. In order to mini-mize the risk, a strike advance mining methodwas recommended (section 13.8.1).

Dewatering the new open pit by means ofboreholes would probably be required, espe-cially prior to the initial box-cut excavation. Acontinuous program of borehole dewateringwill probably be required, to reduce waterpressure in the basement thus avoiding thepossibility of footwall heave. An excellentdescription of the processes involved in the

dewatering of a large open pit mine is given byCameron and Middlemis (1994).

13.6 LIGNITE QUALITY

Coal seams consist of multiple layers ofcarbonaceous material and noncombustiblerocks (waste), such as clay, shale, marlstone,sandstone, etc. These interbedded sedimentsare not continuous and at Soma can be regardedas lenses within the range of the borehole spac-ing. This inevitably means that difficultiesarise when seam quality and reserve quantitiesare being considered. Two alternatives areavailable when considering the approach to beadopted when estimating lignite quality (andreserves). The first is to assume that the zonewithin the lignite seam containing most of thelignite layers is mined in total without anyattempt at selectively mining waste partings.This gives the in situ lignite quality data. Thesecond is to consider mining the lignite layersselectively, aiming to produce a run-of-mine(ROM) product which is of acceptable quality.This gives the mineable lignite quality data.

The nonselective mining approach will re-sult in a ROM product which would not gener-ally meet the power station specification andwould be highly variable in composition. Thiscould be homogenized in a blending stockpileand upgraded in a washing plant, which iscostly and results in losses. Selective mining inthe open pit will be more expensive than bulkmining since time will be lost in movingmachinery and there will also be unavoidablelosses and dilution of lignite associated withthis method. At Soma it was necessary to adoptthe selective mining approach in the lignitequality and reserve estimations, because thelignite quality is low and further quality lossescaused by bulk mining would be unacceptable.

Lignite quality was assessed by evaluatingthe data provided with the borehole logs ob-tained during the first three drilling programs(Table 13.1). Additional quality data came fromsamples of core submitted to the MTA labora-tory in Ankara during the feasibility study. Thecore from the earlier programs was analyzed forash, moisture (on an as-received basis), andcalorific value. In some cases volatile matterand sulfur content were also determined. In the

Page 327: Introduction to mineral exploration

310 M.K.G. WHATELEY

final drilling program the lignite core wasanalyzed for proximate analyses on an as-received basis as well as specific gravity andcalorific value. As the power station stock-piles the lignite in the open, the as-receivedanalyses approximate more closely to the qual-ity of the lignite that is actually burnt. Lignitesamples which do not have the surface mois-ture removed by air drying are said to havebeen analyzed on an as-received basis. Sampleswhich are air dried until the mass of thesamples remains constant (all surface moistureis assumed to have evaporated) before beinganalyzed are reported on an air-dried basis(section 13.6.4).

These data were placed on a computer data-base to facilitate the assessment of the lignitequality, using down-hole-weighted averagingtechniques. The samples were weighted bysample length and specific gravity in order tocalculate the weighted average quality for eachcomposite. To ensure that a true weightedaverage was obtained, quality values had to beassigned to the partings within the seam. Mi-nor partings are often not sampled during theexploration phase of a program, but duringmining these partings will often be incorpo-rated with the lignite, thus reducing the qualityof the product. Partings were sampled in thelast drilling program and these values wereassumed to apply to similar rock types that hadnot been sampled earlier. Similarly, where thespecific gravity of the lignite had not beendetermined, a value was assumed (Table 13.3).Holes where the core recovery was less than75% were omitted as not being representative.

The possible expansion of the lignite minewas investigated for the purpose of increasing

the supply of lignite to the adjacent thermalpower stations. The lignite is crushed and feddirectly into the furnace as a powdered fuel.The solid fuel specifications for steam-raisingplants usually includes the Hardgrove Grindab-ility Index. This test gives an indication of theease with which a material will be crushed(Ward 1984). A low value indicates a hard rockwhile a high number indicates a relatively softrock. One would expect a high number (>63.2)for a lignite, but the values obtained in testsranged between 33 and 57. This probablyreflects the amount of parting material that isincluded with the lignite. Sieve analyses werealso carried out, and these showed that over60% of the ROM production in the open pit is+30 mm. The domestic and industrial marketsrequire a lump product, which the mine caneasily supply.

13.6.1 Borehole sampling

Cores from the seam were split and one halfwas retained in the field. The entire thicknessof the lignite seam was sampled including allparting material. All layers within the seamwhich were larger than 30 cm were sampledas individual samples. Thinner layers were in-cluded with adjacent layers until a minimumthickness of 30 cm was obtained. A maximumof 2 m of core was sampled at one time.

13.6.2 Selection criteria – surface miningoption

In appraising the viability of using surface min-ing methods the following steps were taken.The weighted average quality of the mineablelignite seam was obtained by classifying anysample with less than 55% ash content aslignite and samples with more than 55% ashas waste. The “rules” that were applied to eachsample to obtain the composite mineablethickness are shown in Fig. 13.11. The qualityof the mineable reserves was calculated on allmaterial within the seam except (i) waste part-ings greater than 50 cm and (ii) lignite piles lessthan 30 cm with partings on either side so thatthe total thickness of lignite plus partingsabove and below is greater than 50 cm. In addi-tion, dilution and loss of lignite during miningwas taken into consideration. At each interface

TABLE 13.3 Assumed ash and specific gravity valuesfor the rock types in the Soma Isiklar Basin.

Rock type Assumed ash Assumed SGcontent (%)

Lignite 1.40Clayey lignite 1.70Lignitic 75 2.00

limestoneMarlstone 75 2.30Clay 75 2.40

Page 328: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 311

but with no dilution or loss of lignite taken intoconsideration. Table 13.4 gives some examplesof the changes in quality which arose by apply-ing these selection criteria.

Once the areas to be mined on an annualbasis had been outlined, the quality in eacharea was estimated using the boreholes withina 450 m radius. The mineable qualities, exam-ples of which are seen in Table 13.4, were usedto determine this quality release schedule forthe open pit.

Inverse distance weighting (see section10.4.3) was reviewed and compared to a simpleweighted averaging technique. The two meth-ods indicated little difference in the qualities.The 450-m radius of influence was chosenfollowing a study of semi-variograms of thick-ness, calorific value and cumulations of thick-ness multiplied by calorific value. The rangeindicated on these semi-variograms wasapproximately 560 m. The distance of 450 mwas selected because it is just greater thanthe normal two thirds to three quarters ofthis range which is normally selected (Annels1991), and generally three or more boreholesfell within the radius.

13.6.3 Selection criteria – undergroundmining option

This section has been included to show thatdifferent mining methods require differentselection criteria. During the study, alternativeunderground mining methods suitable for theextraction of a medium-dipping, thick seam,were examined. In-seam mining (a) and cross-seam (b) mining were considered (Fig. 13.12).As a computer database was being used, it waspossible to try sensitivity analyses on the vari-ous selection criteria. These were used to assistin the selection of the best mining methodsand in the calculation of the ROM tonnagesand qualities. The weighted average qualityof the in situ and mineable lignite was calcu-lated using tabulated data examples which areshown on Table 13.5.

In-seam mining

The mineable lignite was determined oncewaste partings (material with >55% ash con-tent and/or <1800 kcal kg−1) greater than 1.5 m

Floor

Surface

Parting 0.30 m to 0.5 m

Lignite

Lignite

Sample

Sample

Sample

Overburden

(b)

Surface

Interburden > 0.5 m

Overburden

Lignite

Floor

Lignite

(c)

Floor

Lignite

Overburden

Lignite

Surface

0.30 mParting

(a)

FIG. 13.11 An illustration of the sampling “rules”that were applied to all the lignite core recoveredduring the study in the Soma Basin. (Modified afterJagger 1977.)

between the lignite and a parting, a dilution byvolume of 10 cm of parting was included re-placing a lignite loss of 10 cm. These interfaceswere taken into account at the top and bottomof the seam as well. The quality of the un-diluted mineable reserves was calculated in asimilar way to that of the mineable reserves,

Page 329: Introduction to mineral exploration

312 M.K.G. WHATELEY

TABLE 13.4 Examples of selectivemining evaluation in the proposedopen pit at Soma Isiklar, Turkey.

Borehole number

218 302

Lignite vertical thickness (m) 14.70 12.10Lignite in waste partings (m) 0.00 0.75Waste rejected from seam (m) 0.00 2.35Mineable vertical thickness (m) 14.70 9.00In situ CV (kcal kg−1) 3662 2400Mineable CV (kcal kg−1) 3595 2500In situ ash content (%) 24.5 39.5Mineable ash content (%) 25.2 37.6In situ SG 1.49 1.73Mineable SG 1.51 1.70Recovery by thickness (%) 100 74

TABLE 13.5 Examples of selectivemining evaluation in the proposedunderground mine at Soma.

thick were rejected. Occasionally thin lignitebeds within the thick waste partings were alsorejected. The top and bottom of the in situ andmineable lignite are the same. It was assumedthat there would be a 100% recovery of the lig-nite in the longwall slice, and 60% recovery oflignite and 40% dilution by waste in the cavedzones above the longwall zones. The mineablequality takes into account the lignite lossesand waste dilution which would occur duringcaving. It was considered that selective mining

could be implemented above and below a wasteparting greater than 1.5 m thick.

Cross-seam mining

The mineable lignite was determined once thetop and bottom waste material were excluded.As this method is less selective, dilution of thelignite is inevitable. This is accounted for byexpecting a 60% recovery of lignite and a 40%dilution by waste.

Borehole number

210 320

Lignite, vertical thickness (m) 18.40 11.80Waste rejected from seam 0.00 3.00Mineable vertical thickness (m) 18.40 8.80Number of waste partings 0 3Number of interface 2 8Dilution (m) 0.2 0.8In situ CV (kcal kg−1) 3524 2071Undiluted CV (kcal kg−1) 3524 2679Mineable CV (kcal kg−1) 3485 2435In situ ash content (%) 27.6 39.9Undiluted ash content (%) 27.6 29.7Mineable ash content (%) 28.1 33.8In situ SG 1.55 1.90Undiluted SG 1.55 1.70Mineable SG 1.56 1.78Recovery by thickness (%) 100 75

Page 330: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 313

Top waste

3 m cave2 m longwall

2 m longwall

Bottom waste

(a)

Horizontal Longwall lift

Top waste

Bottom waste

(b)

Roof of seam

Lign

ite s

eam

Floor of seam

Longwall

Cave

Cave

Longwall

Cave

Floor of seam

Roof of seam

< 3 m cave

Lign

ite s

eam

FIG. 13.12 A sketch of the two underground mining methods proposed for the deep lignite in the Soma Basin:(a) in-seam mining, in which the lignite is mined down-dip and (b) cross-seam mining, in which the lignite ismined horizontally.

resource of the basin but were considered to beinaccessible to open pit mining. All the ana-lyses were reported on an as-received basis. Theair-dried moisture content was calculated forseveral of the lignite samples from the finalseries of holes. The results showed that the in-herent moisture (air dried) was approximately8% and the surface moisture was approxi-mately 7%, totalling 15% on an as-receivedbasis. The moisture content does not preciselyrepresent that of the mined lignite because

13.6.4 Summary of lignite quality – surfacemining option

Once the limits of the potential open pit hadbeen established using the outcrop line and thestripping ratio limit down dip, it was possibleto calculate the average quality of the in situand the mineable lignite. The weighted averagequality data are shown in Table 13.6. The insitu lignite quality includes small areas oflignite which may be included in the lignite

Page 331: Introduction to mineral exploration

314 M.K.G. WHATELEY

TABLE 13.6 Average quality of thelignite in the Soma Basin.SG CV Ash (%) Moisture

(kcal kg−−−−−1) (%)

Open pitIn situ 1.70 3069 32.7 13.5Mineable 1.73 3103 33.0 13.5Block A 1.73 4143 20.9 13.8Block C 1.73 3104 32.5 14.4Block D 1.73 2516 39.1 14.0Block E 1.73 2816 36.6 12.9

UndergroundIn situ 1.66 3337 30.0 16.2Mineable 1.66 3467 27.5 16.2

of the presence of drilling fluid in the coresamples and also the additional moisture thatwould be picked up during mining.

The sulfur content of the KM2 seam wasanalyzed in a few cases. The results indicatethat the KM2 is a low sulfur lignite. Theweighted average is as follows: combustiblesulfur 0.53% and sulfur in the ash 0.42%, giv-ing a total sulfur at 0.95%.

13.6.5 Summary of lignite quality –underground mining option

The areas down dip of the 7:1 stripping ratioline defined that part of the basin that could bepotentially mined by underground methods.Within this area it was possible to calculate theaverage in situ lignite quality and the mineablelignite quality values (Table 13.6).

13.7 LIGNITE RESERVE ESTIMATES

The lignite reserves were calculated by assess-ing the data provided from borehole logs(Whateley 1992) obtained by drilling (Table13.1), within the limits of the open pit. Theselimits were established between the outcropline and the 7:1 stripping ratio line. On thenorthwest and the east sides of the pit, thelimit is determined at the depositional edge ofthe seam. This limit was determined by usingthe basement and basal KM2 structure contourmaps (section 13.4.1). The pit outline wasplaced on the isopach map of the total verticalthickness of the KM2 seam, and the area be-tween each isopach was measured with the aidof a planimeter (cf. section 13.4). The areas

were multiplied by the average vertical thick-ness between the isopach lines (the thicknessvalue at the midpoint) to obtain the in situvolume. The volume was multiplied by theaverage specific gravity to obtain the in situtonnage.

The mineable lignite tonnage was calculatedfrom the in situ tonnage by applying a recoveryfactor, examples of which are given in the finalrow of Table 13.4. The weighted average recov-ery for the open pit is 91%, but this varies fromas low as 45% in borehole 208 to 100% in manyof the remaining holes. The recoveries of mine-able lignite from the total lignite will changeduring mining depending on the local geology.Development drilling immediately in advanceof production will determine these recoveries.The reserve figures were calculated as follows:in situ reserves 49.4 Mt, mineable reserves40.6 Mt, and kriged estimate of in situ reserves46.3 Mt (Lebrun 1987).

13.7.1 Comparison of estimation methods

Zarraq (1987) undertook a comparison of re-serve estimation methods using the Soma data.He calculated the in situ reserves for the wholebasin using polygons, manual contours, andkriging. His reserves estimates are:

Method Tonnage

Polygons 117.8 MtManual contouring 102.5 MtKriging (150 × 150 m blocks) 109.4 MtStatistical mean 117.5 Mt(area × average thickness)

Page 332: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 315

The confidences as a percentage of the meanfor the open pit area are as follows: thickness(A) 15.9%, SG (B) 5.9%, and global confidence(G) 17.0%. These results indicate that the re-serves calculated from thickness and SG aresufficiently well known that they can be classi-fied as proved within the confidence limits.

13.8 SURFACE MINE EVALUATION

This evaluation project called for a productionrate of 2.13 Mt ROM lignite each year. Afterrejection of some waste dilution by rotarybreakers, the delivered output is expected to be2 Mt per year. This rate of production was con-sidered appropriate to exploit the mineablereserves over a mine life of 21 years. Reservesare contained in four fault-bounded blocks (A,C, D, and E), but it was not considered advis-able to include block B, because the lignitehas been partly extracted using now abandonedunderground workings. The presence of thefaults and the varying dip angle of the foot-wall dictated the mining method and box-cutlocations.

13.8.1 Selection of mining method

The hard massive marlstone which formsthe overburden will require blasting beforeremoval can take place and this rules out theuse of a dragline or a bucket wheel excavator.Rear-dump trucks and face shovels will have tobe used. Three alternative mining configura-tions may be considered using this equipment:(i) advance down the dip; (ii) advance up thedip; or (iii) advance along the strike. The thirdmethod, also known as terrace mining, wasconsidered the most suitable with the par-ticular geological and geotechnical conditionswhich exist at Soma (section 13.5.4).

Following this method a box-cut would beexcavated down the full dip of the deposit fromthe outcrop until an economic stripping limit,or practical mining depth limit, is reached.This study concluded that a maximum miningdepth of 150 m was feasible, although con-straints were generally related to the geologicalconfiguration and consideration of maintain-ing footwall stability, rather than to strippingratio economics. Advance could then be in one

The polygonal method has probably overesti-mated the reserve slightly because there aresome large polygons which have thick ligniteassociated with them, e.g. the polygon centeredon hole 208 has an area of 187,630 m2 with17.95 m of lignite, while the average area ofall the polygons is only 70,000 m2. Polygonsalso create artificial boundaries which do notexist in nature. The only way to reduce theestimation variance is to increase the sampledensity, which costs time and money. How-ever, this is a quick method of providing aglobal estimation.

Manual contouring smooths the data by vir-tue of the linear interpolation used to constructthe map. It appears to weight the large area oflow values excessively, which may explain thelower reserve figure. Although the kriged esti-mate of reserves does take the spatial relation-ships into account, the estimation varianceswere generally high. Kriging gives the best esti-mate for each block with the lowest estimationvariance, so the advantage of kriging is thatone knows how reliable each block estimate is(section 9.5.1).

13.7.2 Confidence in the reserves

The confidence in the mineable reserve esti-mate for the open pit area was calculated usingthe mean and standard deviation for thethickness and specific gravity derived fromthe boreholes drilled in the area designated tobe the surface mine. The formula

tSn

was applied where S = standard deviation, n =number of boreholes, and t = t value for n − 1degrees of freedom at the 90% confidence level.

The results of this calculation wereexpressed as a percentage of the mean. Thepercentages of the mean results were thenused in the following formula:

G = A2 + B2

where A = percentage of the mean for thick-ness, B = percentage of the mean for SG, andG = global confidence in the average expressedas a percentage.

Page 333: Introduction to mineral exploration

316 M.K.G. WHATELEY

the stripping ratio is more or less constant overthe life of the mine, and hence the mining costsare stabilized.

13.8.2 Mine design

Box-cut locations

Two box-cuts were proposed to take account ofthe adverse geological factors (Fig. 13.14). Thefirst was sited in blocks C and D, close to thecenter of the open pit reserve. This locationwas dictated by consideration of local strippingratios, seam dip, footwall clay thickness, andthe presence of faulting. The central site meansthat mining will eventually advance both tothe east and to the west. It was proposed thatthe second box-cut should be excavated at thewestern end of block A. Advance would betowards the west.

Slopes and access

A bench height of 12 m was proposed withexcavation from horizontal benches in theoverburden and lignite, with connecting ramps

Directi

on of adva

nceInternal spoil dump

End wall

To ROM hopperHigh wall

Advan

cing

face

Footwall

FIG. 13.13 A sketch of a truck-and-shovel, terrace-mining operation proposed for the surface mine in theSoma Basin.

or both directions along the strike line, depend-ing on the box-cut location within the proposedmining area. Waste disposal within the excava-tion is practical (and usually desirable) as thebox-cut excavation is enlarged. Horizontalbenches would be formed in the overburdenalong the advancing face and along thehighwall formed on the deep side of the excava-tion. As much spoil as possible from theadvancing face would be removed along thehighwall benches for disposal to form an inter-nal spoil dump within the excavation andbehind the advancing face; the remainder hav-ing to be dumped outside the pit. The frontof the spoil dump and the advancing face thenadvance in unison as mining continues alongthe strike of the deposit (Fig. 13.13). The finalhighwall is expected to have a maximum slopeangle of 45 degrees.

The application of this mining method willbring several advantages: (i) a minimum area offootwall clay will be uncovered at any time,thus reducing the risk of footwall failure; (ii)internal dumping of waste reduces transportcosts, helps stabilize the footwall, and beginsreclamation at an early stage of mining; and (iii)

Page 334: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 317

disposed of within the pit will need be trans-ported on haul roads at 48-m vertical intervalsformed in the highwall. Ramps between haulroads, and to the rim of the pit would be formedin the advancing faces. Spoil disposal outsidethe pit will be via three ramps to main haulageroutes on the pit rim. Lignite will be trans-ported to the ROM lignite plant by a mainhaul road from the pit to the south portal ofthe main rail tunnel (Fig. 13.14) through themountain and then by overland conveyer beltto the power stations near Soma (Fig. 13.1) tothe north.

Spoil disposal

One-third of the 180M bank m3 of overburden(Table 13.2) could be backfilled into the miningexcavation, with the remainder being disposedof in two dumps outside the pit. Bank m3 refersto the volume of rock in situ. Once mined therock “swells” to occupy a greater volume

N

BLOCKA

BLOCKB

BLOCKC

BLOCKD

BLOCKE

Existingopen pit

32,000N

31,000N

30,000N

48,000E 49,000E 50,000E 51,000E 52,000E

CONCESSION BOUNDARY

Existingopen pit

Final open pit walls

Fault

Initial box cuts

Direction of advance

Outcrop of KM2 seam

Rail tunnel portal

0 500 m

FIG. 13.14 The outline of the final open pit design determined by the 150-m overburden isopach.

at up to 8% grade located in the advancingfaces (Fig. 13.13). The highwall should havean overall slope angle of 45 degrees, formed bythe 12-m bench, and by berms and haulroadsinside the pit for internal waste dumping(Fig. 13.15). Controlled blasting will be re-quired for stable slopes.

Advancing faces will require an overall slopeangle of 10–20 degrees, depending on the cur-rent mining activity. A working width of 65 mwas proposed for each bench for independentdrilling and blasting, loading and haulageactivities. Lignite excavation would be fromsub-benches 4 m high, subdividing the 12-moverburden benches for ease of excavation andloading (Fig. 13.15).

Haul roads

A haul road width of 20 m will be needed toprovide room for trucks to pass with adequateside and center clearances (Fig. 13.15). Spoil

Page 335: Introduction to mineral exploration

318 M.K.G. WHATELEY

FIG. 13.15 Details of (a) the proposed open pit bench design, (b) the haul road, and (c) the lignite-miningoperation.

(a)Haul road

4 m

20 m

12 m

Haul road

Haul road

Lignite

45°

56°

Bench

Berm

4 m

Operating level

Haulage level 20°

Lignite

Parting

Footwall

(c)

1.5

2.0 2.0 2.0

2.5

BermDrain

(b)

0 5 m

20 m

5.0 5.0

Page 336: Introduction to mineral exploration

13: SOMA LIGNITE BASIN, TURKEY 319

although the increase in the void spaces resultsin a mass reduction m−3.

13.8.3 Mining equipment

Large items of open pit equipment were notmanufactured in Turkey at the time of thisstudy. Calculations of equipment utilization,mechanical availability, and productivity weremade to decide what equipment would haveto be imported. The overburden equipmentrequirements were assessed to be three rotaryblasthole rigs (250-mm-diameter holes), andelectrically powered rope shovels for over-burden stripping, assisted by front-end loaders.Hydraulic face shovels were recommendedfor lignite loading. Rear-dump trucks wereselected for both overburden and lignite trans-port. Supporting equipment included trackedand wheeled bulldozers and a conventionalfleet of pit and haul road maintenance vehicles.

13.9 SUMMARY

Available data on the reserve tonnage andquality of the Soma deposit were adequate fora feasibility evaluation, which suggested thatthe reserves would support a projected life of21 years at the proposed rate of productionand specified quality. Additional data wererequired on the groundwater regime before

forecasts could be made of the effects ofgroundwater pressure and flows on mining.This highlights the need for careful planningof any exploration drilling program to max-imize the data that can be gained from anydrilling.

13.10 FURTHER READING

Ward (1984) covers a wide range of topics andhis book is intended for senior students andprofessional geologists. He discusses coal ex-ploitation, mining, processing, and utilizationin a clear manner. Other papers on these sub-jects can be found in the Bulletin de la Sociétégéologique de France, 162 (1991). The SpecialPublications of the Geological Society editedby Annels (1991) and Whateley and Spears(1995) give some excellent case histories of coaldeposit evaluation. Stach’s Textbook of CoalPetrology (Stach 1982) is well referenced andgives detailed information on the microscopicproperties of coal. The sedimentology of coaland coal-bearing strata is covered in a seriesof papers in a Geological Society of LondonSpecial Publication edited by Scott (1987).Further papers on the sedimentology of coal-bearing strata are to be found in the IAS SpecialPublication edited by Rahmani and Flores(1984), and in the International Journal of CoalGeology.

Page 337: Introduction to mineral exploration

320 C.J. MOON & M.K.G. WHATELEY

14Witwatersrand

Conglomerate Gold –West Rand

Charles J. Moon and Michael K.G. Whateley

14.1 INTRODUCTION

Alluvial concentrations of heavy minerals(placer deposits) are very important sources ofgold, such as the rich deposits in the Yukon, theMother Lode area of the forty-niners in north-ern California, and the modern-day rush in theSierra Pelada area of the Amazonian Brazil.Even today much Russian production comesfrom the Lena and Magadan alluvial depositsin northeast Siberia. However modern alluvialgold output is dwarfed by production fromdeposits in quartz pebble conglomerates thatappear to be fossilized placers. These depositsare restricted in geological time to the lateArchaean and early Proterozoic and dominatedby one basin, the Witwatersrand Basin in SouthAfrica. Other gold-producing quartz pebble con-glomerates occur in the Tarkwa area of Ghana,Jacobina and Moeda in the São Francisco Cratonof Brazil, while the Elliot Lake area in Ontariohas been a significant producer of uraniumfrom similar rocks and carries minor gold.

This case history concentrates on the Wit-watersrand (Afrikaans for ridge of white watersand often abbreviated to Rand or Wits) Basinthat has been responsible for the productionof 37% of all gold produced in modern timesas well as significant uranium. This basin wasthe source of 399 t of gold metal in 2002, about16% of annual world gold production. Miningin the Witwatersrand Basin is from large tabu-lar orebodies that form the basis of some ofthe largest metalliferous underground mines inthe world and which reach more than 3500 mbelow surface. This combination of depthand scale has been a great challenge to mining

engineers and it has produced a long andinteresting exploration history involving largeexpenditures on deep drilling and geophysics.

14.2 GEOLOGY

Mineralized conglomerates are found in avariety of settings within the Kaapvaal Craton(Fig. 14.1) but the vast majority of productionhas come from the upper part of a 7000-mthick sedimentary sequence termed the Wit-watersrand Supergroup. These sediments weredeposited within a 350 × 200 km basin thatformed relatively soon after the consolidationof the Kaapvaal Craton from a series of green-stone belts and granitic intrusives. The Wit-watersrand sedimentation was part of a muchlonger depositional history which began withsedimentation in a probable rift setting ofbimodal volcanics and limited sediments,termed the Dominion Group (Stanistreet &McCarthy 1991). Recent age determinationsdate volcanics within the succession at 3074± 6 and underlying granites at 3120 ± 6 Ma(Armstrong et al. in De Beer & Eglinton 1991).The basal sediments are conglomeratic and sup-ported minor gold production and substantialuranium operations. These Dominion Groupsediments form structural remnants overlainby Witwatersrand Supergroup sediments.

14.2.1 Stratigraphy of the WitwatersrandSupergroup

The Witwatersrand Supergroup is divided intotwo, the lower West Rand Group and the upper

Page 338: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 321

Granite – greenstone domes

Metamorphic belts

SWAZILAND

MOZAMBIQUE

SOUTH AFRICA

BOTSWANA

LIMPOPO BELT

MO

ZAM

BIQ

UE

BE

LT

NAMAQUA - NATAL BELT

Indian Ocean

LESOTHO

WitwatersrandBasin

Pretoria

K A A P VA A L C R AT O N

Z I M B A B W E A N C R AT O N

N

0 200 km

B

M

S

P

ZIMBABWE

Greenstone belts (Swaziland Sequence)Primary gold provinces

Dominion Group and WitwatersrandSupergroup Placer gold province

BMSP

Barberton Mountain Land

Murchison Range

Sutherland Range

Mount Maré (Picksburg)

FIG. 14.1 Regional tectonic setting of the Witwatersrand Basin. Mesozoic and younger cover has beenremoved. (After Saager 1981.)

Page 339: Introduction to mineral exploration

322 C.J. MOON & M.K.G. WHATELEY

Central Rand Group, which contains most ofthe gold deposits.

The West Rand Group consists of aboutequal proportions of shale and sandstone and250 m of volcanics, the Crown lavas (Tankardet al. 1982). A number of iron-rich shales arepresent in the lower part of the group and thesegive a magnetic response used in exploration.The depositional environment was littoral orsubtidal on a stable shelf as indicated by thegeneral upward coarsening nature of the macro-cycles. However there are a number of upwardfining cycles with conglomerate at their base,which represent more fluvial conditions andcontain mineralisation. Despite their alluringnames, the Bonanza, Coronation, and PromiseReefs (reef is used as a synonym for mineral-ized conglomerate throughout the Witwaters-rand) have produced only 35 t of gold, althoughtheir mineralogy is similar to reefs in the Cen-tral Rand Group (Meyer et al. 1990). Age dateson detrital zircons give a lower age limit fordeposition of 2990 ± 20 Ma and an upper limitof 2914 ± 8 Ma on the Crown Lava (De Beer &Eglington 1991).

The later stages of the deposition of the Wit-watersrand Supergroup reflect major changesin the style of sedimentation together withchanges in the shape and size of the basin ofdeposition (Stanistreet & McCarthy 1991). TheCentral Rand Group consists predominantlyof coarse-grained subgreywacke with lessthan 10% conglomerate, silt, and minor lava(Tankard et al. 1982). The overall depositionalsetting is of fan delta complexes progradinginto a closed basin, perhaps containing a lakeor inland sea. The change in depositional stylereflects the increasing importance of tecton-ism, certainly of folding, probably of faulting.Burke et al. (1986) suggested that the overalltectonic setting is that of a foreland basin re-lated to collisional tectonics in the area of theLimpopo River to the north. Dating of detritalzircons shows that the Eldorado Formation,at the top of the group, is younger than 2910± 5 Ma and older than overlying Ventersdorplavas dated at 2714 ± 8 Ma (Robb et al. 1990).

The fan delta complexes mainly have dis-tinct entry points into the basin as indicated bypalaeocurrent data, pebble size, and composi-tion and isopach maps; these correspond withthe location of the major goldfields. At present,

eight major goldfields are recognized, fromsouth to northeast, the Orange Free State (OFSor Welkom), Klerksdorp, West Wits Line (orFar West Rand), West Rand, Central Rand, EastRand, South Rand, and Evander (Fig. 14.2). Theareas between the goldfields are known asgaps, notably the Bothaville and PotchefstroomGaps. Recent exploration has shown that thereis significant mineralisation in these gaps, par-ticularly the Bothaville Gap, where the Targetmine is currently being developed. Howeverthe south side of the basin seems to be essen-tially barren.

Regional correlation between the differentreefs is difficult as they are developed onunconformities and some reefs are developedin only one goldfield, as might be anticipatedfrom the discrete sediment entry points. How-ever broad correlations are possible as shownin Fig. 14.3. The reefs can be divided into twotypes, sheet conglomerates and channelizedconglomerates (Mullins 1986). Sheet reefs havegood continuity and are gravel lags overlainby mature sand, whereas the channelized reefshave well-defined channels filled with lenti-cular bodies of sand and gravel. Examples ofchannelized conglomerates are the KimberleyReef and the Ventersdorp Contact Reef (VCR),whereas the Vaal Reef, Basal Reef, and theCarbon Leader are sheet-like. Where bothsheet-like and channelized conglomerates aredeveloped on the same unconformity, as in thecase of the Steyn Reef in the OFS Goldfield, thechannelized conglomerates are more proximalto the source area and change into sheet-likeconglomerates over a distance of 20 km. Thestratigraphical location of the major reefs isshown in Fig. 14.3.

14.2.2 Overlying lithologies

Little of the known extent of the Witwater-srand Basin sediments is exposed and mostis covered by later rocks, both Precambrianand Phanerozoic: lavas and sediments of theVentersdorp Supergroup, and sediments of theTransvaal and Karoo Sequences.

The uppermost major reef in the Wit-watersrand Basin, the Ventersdorp ContactReef, is preserved by a cover of Ventersdorplavas and appears to have formed at a time of vol-canism and block faulting. It is economically

Page 340: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 323

important in the West Wits Line and Klerks-dorp areas, where it is heavily channelized.Although other placers have developed higherin the Ventersdorp succession, and undersimilar conditions, they are of little economicimportance. The remainder of the Ventersdorpsequence consists largely of tholeiitic and highmagnesia basalts with lesser alluvial fan andplaya lake sediments.

The Transvaal Sequence was deposited fromaround 2600 to 2100 Ma over a wide area ofthe Kaapvaal Craton. The important parts ofthe sequence as far as exploration and miningare concerned are the basal conglomeratic unit,known as the Black Reef, and the overlying

dolomites of the Malmani Subgroup. The BlackReef has been mined in a number of areas, not-ably on the East Rand and in the Klerksdorpareas. Mineable areas correlate with areaswhere the Black Reef cuts Central Rand Groupmineralisation, and it seems certain that theBlack Reef mineralisation is the result of re-working or remobilization of gold from below(Papenfus 1964). The dolomites overlie most ofthe mining area on the West Rand and Klerks-dorp areas and cause considerable problems asthey contain large amounts of water and areprone to the formation of sinkholes. One sink-hole swallowed the primary crusher at WestDriefontein in 1962.

O.F.S.

Klerksdorp

Ventersdorp

Wolmaransstad Viljoenskroon

PotchefstroomGap

Wesselsbron

Winburg

Welkom Steynsrus

Kroonstad

Vredefort

Sasolburg

Leslie

Johannesburg

Randfontein

Carletonville

Delmas

Cooke Section

S.Deep

E.R.

Witpoortjie horst

Evander

Bank break

W.W.L.

Western Deep Levels

Vaal ReefsComplex

St Helena G.M.

Krugersdorp

Basement rocks

Dominion Group

Central Rand Group

West Rand GroupWitwatersrandSupergroup

Reclamation of slimesfor gold and uranium

Palaeocurrent directions

Current and recent gold anduranium mines

N

0 50 km

FIG. 14.2 Witwatersrand Basin with later cover removed, E.R. indicates location of East Rand Basin (Fig. 3.7).Note that the mine names are those referred to in the text and many of the current mines have new names.The new target mine is ∼40 km north of Welkom. (Modified from Camisani-Calzolari et al. 1985.)

Page 341: Introduction to mineral exploration

324 C.J. MOON & M.K.G. WHATELEYWELKOM KLERKSDORP

W.W.L.CARLETONVILLE

WEST RANDKRUGERSDORP

3000 m

2000 m

1000 m

0

–1000 m

–2000 m

–3000 m

–4000 m

WIT

WA

TE

RS

RA

ND

SU

PE

RG

RO

UP

Individual

VCRMassive

Composite

WE

ST

RA

ND

GR

OU

PC

EN

TR

AL

RA

ND

GR

OU

P

VE

NT

ER

SD

OR

PS

UP

ER

GR

OU

PK

LIP

RIV

ER

SB

ER

GG

RO

UP

HO

SP

ITA

L H

ILL

GO

VE

RN

ME

NT

JEP

PE

ST

OW

NJO

HA

NN

ES

BU

RG

TUR

FFO

NTE

INS

UB

GR

OU

P

Formation Placers Formation Placers Formation Placers Formation Placers

LeaderBasal/Steyn

B

EA

VCR

E3.5

Vaal

Coronation

Bonanza

Carbon Leader

Kloof

VCR

Main

Promise

Major production

Crown

Eldorado

Welkom

St. Helena

Virginia

Venterspost

Klerksdorp

GoldEstates

Crown

Welgegund

Bonanza

OrangeGrove

Crown

O. Grove

Brixton

Promise

Roodepoort

Kimberley

(Westonaria)

Venterspost

Venterspost

Elsburg

Kimberley

Randfontein

Elsburg

Shale

Reef

Lava

Quartzite

Kimberley

FIG. 14.3 Stratigraphical columns of the Witwatersrand Supergroup in the main gold-producing areas. Notethe general correlations between goldfields and major producing conglomerates. (After Tankard et al. 1982.)

Page 342: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 325

FIG. 14.3 (CONTINUED)

EVANDERHEIDELBERGBENONIJOHANNESBURG

Major production

Main

MainLeader

South

Bird Kimberley

Bird

Main L. Main L. Main L.

Kimb.

Formation PlacersFormation PlacersFormation PlacersFormation Placers

?

?

?

?

Mondeor

(Westonaria)

Kimberley

Krugersdorp

Crown

Witpoortjie

Promise

Parktown

O. GroveO. Grove

Promise(not fullyexposed)

Florida

Kimberley

CrownCrown

Promise

O. Grove

OrangeGrove

Promise

Coronation

Leandra

Evander

Kimberley

Coronation

Promise

Page 343: Introduction to mineral exploration

326 C.J. MOON & M.K.G. WHATELEY

The youngest major sedimentary sequencefound on the Witwatersrand consists of sedi-ments of the Karoo Supergroup deposited fromCarboniferous to Triassic times in glacial, mar-ginal marine to fluvial conditions. These areimportant in two respects. The Ecca Group sedi-ments are coal-bearing and besides providing acheap source of energy have probably generatedthe methane that is found in a number of minesand which poses a threat of underground explo-sions. The Karoo sediments also have a highthermal gradient and give rise to a blanketingeffect which raises the overall thermal gradientand reduces the ultimate working depths inboth the OFS and Evander Goldfields.

Witwatersrand sediments have been intrudedby a variety of dykes, which cause considerabledisruption to mining. The oldest intrusivesare dykes of Ventersdorp age and the youngestlamprophyres of post-Karoo age but the mostimportant are dykes of “Pilanesberg” age, upto 30 m wide, on the West Rand which dividethe dolomites into separate compartments forpumping purposes (Tucker & Viljoen 1986).

14.2.3 Structure and metamorphism

Although the Witwatersrand sequence hasclearly been deformed and metamorphosed,there has been little emphasis on making asynthesis until the last few years. An exceptionis faulting of the reefs which has been exam-ined in detail as this is a major control on theday-to-day working of the mines. The type ofdeformation varies through the basin but themost important for mining is block faultingwhich forms the boundaries of mining blocks.The main activation of these faults was inmid Ventersdorp times when there was alsothrust faulting at the margin of the OFS Gold-field (Minter et al. 1986). The major faultswere, however, reactivated in post-Transvaalbut pre-Karoo times, although the exact tim-ing is unclear. In the east of the basin thesediments are affected by the intrusion of theBushveld Complex which has tilted them.Recent dating suggests that a major basinwidefracturing event, including the developmentof pseudo-tachylite in the Ventersdorp ContactReef, is associated with the Vredefort eventat ~2023 Ma that is probably the result of ameteorite impact (Frimmel & Minter 2002).

The occurrence of a pyrophyllite–chloritoid–chlorite–muscovite–quartz–pyrite assemblagewithin the pelitic sediments indicates a re-gional greenschist metamorphism at temper-atures of 350 ± 50°C that has been dated at~2100–2000 Ma, approximately the same ageas the intrusion of the Bushveld Complex(Phillips et al. 1987, Phillips & Law 2000,Frimmel & Minter 2002).

14.2.4 Distribution of mineralisation andrelation to sedimentology

Gold and uranium grades appear to be con-trolled by sedimentological factors on both asmall and a large scale. The diagram in Fig. 14.4shows a typical relationship between gradeand sedimentology for a channelized reef,with higher grade areas restricted to the mainchannels. The delineation of these payshootsmay be the only way to mine certain con-glomerates economically (Viljoen 1990). Thedistribution of facies can also be mapped inthe sheet-like placers as shown for the CarbonLeader in Fig. 14.4b. High grade gold and ura-nium deposits are associated with the carbonseams and thin conglomerates, particularly inthe West Driefontein area. Mapping of thesesedimentary structures is also important asthey form homogenous geostatistical domainsfor the purpose of grade calculations. In generalthe oligomictic placers have higher grades thanthe polymictic conglomerates, which containshale and chert as well as vein quartz clasts,presumably as a result of greater reworking.

On a smaller scale, variations in sediment-ary structures control grade distribution. Thisis exemplified by Fig. 14.5 which shows thedistribution of gold through the channelizedA reef (Witpan Placer) in the OFS Goldfield.In this case gold is concentrated at the topand base of the conglomerate unit. Gold is alsoassociated with heavy mineral bands withinthe unit. In the sheet conglomerates gold isclosely associated with carbon, if present, andthe basal few centimeters may contain mostof the gold in the reef. The remainder of thereef, although mostly waste, must be minedto expose the working face using a minimumstoping height of 80 cm. Jolley et al. (2004) pro-vided a contrasting explanation for the detaileddistribution of gold in the conglomerate, which

Page 344: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 327

Alluvial fan Fault scarp

Scouredpediment

N

0 5 km

Perennial channel(well mineralized)

Flood plain(poorly mineralized)

PROXIMAL

MIDFAN

DISTAL

(a)

Bank Break

(b)

Carbon seam type

Thin conglomeratic type

Thin sandy type

Thick multiple band type

Erosion channel

Reworked deposits

Inferred palaeocurrent direction

FIG. 14.4 (a) Sketch showing a reconstruction of depositional conditions in a channelized reef and relation tograde. Generalized from the Composite Reef. (From Tucker & Viljoen 1986.) (b) Facies of the Carbon Leaderon the West Wits Line. Driefontein is the mine immediately west of the Bank Break. (From Viljoen 1990.)

Page 345: Introduction to mineral exploration

328 C.J. MOON & M.K.G. WHATELEY

W EAu (cm. g t–1)

Au (log g t–1)

3

2

1

0

Hei

ght (

m)

0 10 20 30Distance (m)

Aandenk conglomerate

Witpan placer F

F

F

FIG. 14.5 Relation between grade and location in the Witpan placer, OFS Goldfield. The Witpan placer isstratigraphically above the B Reef of Fig. 14.3. The exact grades were omitted for commercial reasons.(After Jordaan & Austin 1986.)

they advocated were controlled by small scalefractures related to thrust fracture networks.

14.2.5 Mineralogy

The conglomerates mined (Fig. 14.6) are pre-dominantly oligomictic with clasts of vein

quartz. The matrix consists largely of sec-ondary quartz and phyllosilicates, includingsericite, chlorite, pyrophyllite, muscovite, andchloritoid. The opaque minerals have beenthe object of intense study and more than70 are known; reviews are given by Featherand Koen (1975) and Phillips and Law (2000).

FIG. 14.6 Underground photographof a typical conglomerate. TheUE1a Upper Elsburg at CookeSection, Randfontein EstatesGold Mine. Courtesy of theSedimentological Section of theGeological Society of South Africa.

Page 346: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 329

Pyrite is the predominant opaque mineralwith important, but lesser amounts, of gold,uraninite, carbon, brannerite, arsenopyrite,cobaltite, galena, pyrrhotite, gersdorffite, chro-mite, and zircon.

Pyrite forms 3% of a typical reef and ispresent largely as rounded grains. The largerof these are known as buckshot pyrite, beingsimilar to the shot used in cartridges. Pyriteoccurs as grains with recognizable partings andas porous grains, often filled with chalcopyriteor pyrrhotite. Some concretionary grains arealso present. At least some of the pyrite wasformed by pyritization of other minerals androcks, such as ilmenite or ferruginous pebbles.Pyrrhotite probably formed largely as a resultof the metamorphism of pyrite and is mostfrequent near dykes. Uraninite is presentmainly as rounded grains of apparently detritalorigin. Gold occurs largely in the fine-grainedmatrix between pebbles as plate-like, euhedral,or spongy grains. The majority of gold is inthe range 0.075–0.15 mm and is not oftenvisible. Very fine-grained gold is found in thecarbon seams. The majority of gold shows signsof remobilization and replacement relationswith the fine-grained matrix. It also is com-monly associated with secondary sulfides andsulfarsenides. Typically gold grains containabout 10% silver.

Uraninite is the dominant uranium mineralalthough the titanium-rich uranium mineral,brannerite, predominates in the Vaal Reef,presumably forming as a result of a reactionbetween ilmenite and uranium in solution.Carbon, or more correctly bitumen, is commonin a number of the sheet conglomerates andis closely associated with high uranium andgold grades. Hallbauer (1975) suggests thatit formed from lichen and fungi, althoughmore recent work shows that it is a mature oiland some authors (e.g. Barnicoat et al. 1997)have argued for a source from outside thebasin.

Chromite is widely distributed in thereefs and shows a sympathetic relationshipwith zircon. Both appear to be detrital. Tworare detrital minerals are diamond andosmium-iridium alloys. Authigenic sulfidesare patchily distributed and include cobaltite,gersdorffite, sphalerite, and chalcopyrite.This assemblage is associated with remobi-lized gold.

14.2.6 Theories of genesis

The origin of the mineralisation has beena matter of bitter debate between thosewho believe it to be essentially sedimentary,developed either by normal placer-formingmechanisms, the palaeoplacer theory, or byprecipitation of gold in suitable environ-ments, the synsedimentary theory; and thehydrothermalists who believe that the con-glomerates were merely the porous sedimentswithin which gold and sulfides were depositedfrom hydrothermal solutions derived eitherfrom a magmatic source or from dehydrationof the West Rand Group shales. The full back-ground to the debate can be found in Pretorius(1991), Phillips and Law (2000), and Frimmeland Minter (2002).

The key arguments in favor of a placer originare the strong spatial correlation between gold,uraninite, and unequivocally detrital zircon,and the intimate relationship between theheavy minerals and the sedimentary structuresand environment. Placer advocates also pointto the equal intensity of mineralisation inporous conglomerates and less porous pyriticquartzites. The problems for the placerists aremainly mineralogical: the very small particlesize of the gold, its hackly shape, and low fine-ness (i.e. high % Ag), which are more typical ofhydrothermal deposits, the presence of pyrite,uraninite, and the absence of black sands typ-ical of modern placers. The placerists invokemetamorphism to explain the shapes of thegold grains and explain the lack of magnetiteas the result of intense reworking (Reimer &Mossman 1990) or transformation into pyriteafter burial. The problem of detrital pyrite anduraninite is somewhat more difficult to explainas these minerals are unstable in most mod-ern environments, although they are knownfrom glacially fed rivers, such as the Indus inPakistan (for discussion see Maynard et al.1991). However it seems very probable that theArchaean atmosphere contained less oxygenthan at present and that both uraninite andpyrite would be stable under these conditions.Recent direct dating of gold and pyrite fromthe Basal Reef using the Re/Os technique hasgiven ages of 3016 ± 110, indicating a detritalorigin for the gold, as these ages are similarto the zircon ages and older than the ages ofdeposition of the sediments (Kirk et al. 2002).

Page 347: Introduction to mineral exploration

330 C.J. MOON & M.K.G. WHATELEY

By contrast, gold from the Ventersdorp Contactreef appeared to have ages reset by metamorph-ism. Osmium concentrations are compatiblewith derivation from a mantle source.

Another major problem is the source of thelarge quantity of gold and uranium in the basin,far more than is likely to have been derivedfrom any Archaean greenstone belt, certainlymore than from the most productive SouthAfrican greenstone belt, Barberton. While itis certain that greenstone belts were part of thesource area, as evidenced by the occurrence ofplatinum group minerals in the Evander area, itseems likely that the gold and uranium couldalso have been derived from altered granites.Studies by Klemd and Hallbauer (1986) andRobb and Meyer (1990) have demonstratedthe presence of hydrothermally altered graniteswith gold and uranium mineralisation in theimmediate hinterland of the WitwatersrandBasin. A neat twist to the placer theory is thesuggestion by Hutchinson and Viljoen (1988)that the gold could have been derived fromreworking of low grade exhalative gold depositsin the West Rand Group.

The principal arguments for a hydrothermalorigin are that the gold is crystalline, sometimesreplaces pebbles, may be contained within andreplace pyrite that has replaced pebbles, and isassociated with a suite of hydrothermal oreminerals and alteration products (sericite andchlorite) typical of hydrothermal orebodies theworld over. Previous important advocates ofthe hydrothermal origin were Graton (1930),Bateman (1950), and Davidson (1965). Thehydrothermal theory has been recently givena new breath of life by Phillips et al. (1987)and Barnicoat et al. (1997), who pointed out thelack of metamorphic studies and demonstratedthat the shales show evidence of metamorph-ism to greenschist facies similar in many waysto the conditions of formation of greenstonebelt gold deposits. They also suggested that insome areas faulting may have controlled golddistribution. Regional stratabound alterationzones have also been recognized which cor-relate spatially with major gold reefs (Phillips& Law 2000).

The modified placer theory, which empha-sizes the control on the occurrence of oreminerals by placer-forming mechanisms butaccepts some modification by metamorphism,

has been widely used in exploration. If youare lucky enough to make an underground visityou should try and discuss the merits of thedifferent theories.

14.3 ECONOMIC BACKGROUND

14.3.1 Profitability

The economics of mining in the Witwatersrandhave been largely governed by the price ofgold. Unlike most other commodities the goldprice is not controlled by supply and demandfor industrial use but by the perception of itsworth. As gold is transportable and scarce themetal has been used as a way of moving andstoring wealth. Its price was regulated for longperiods of time by intergovernmental agree-ment, enforceable by the large reserves heldby reserve banks such as the US hoard at FortKnox. These large stocks were initially heldwhen currencies were directly convertible intogold. This convertibility, the gold standard,was abandoned by the UK in 1931 and by theUSA in 1933. From 1935 to 1968 the gold pricewas fixed at $US35 t oz−1 but in 1968 PresidentJohnson allowed the gold price to float forprivate buyers and in 1971 President Nixonremoved the fixed link between the dollarand gold and left market demand to determinethe daily price. The 1970s saw extreme fluctua-tions in the value of gold as it was used as ahedge against inflation with sharp increasesin price (see Fig. 1.8) following the increases inthe oil price in 1973 and 1980. The gold priceremained in the $US300–400 range in the restof the 1980s, steadily depreciating against aninflating dollar as production has increased.The sharp increase in the gold price in 1931 setoff the major exploration programs that led tothe discovery of two new goldfields in SouthAfrica. That of the 1980s led to the prospectingboom of 1985–89.

Long-term plots (Fig. 14.7) of the overallprofitability of the Witwatersrand minesreflect these changes in the gold price, changesin the South African rand–US dollar exchangerate, and increases in working costs. The over-all picture has been relatively healthy withspectacular profits in the early 1980s. Sincethen working costs have escalated sharply as

Page 348: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 331

and complexity of mining. The major politicalchanges culminating in the 1994 electionsmeant that the gold mines have not beenwilling or able to maintain wage differentialsbased on skin color or migrant labor andoverall labor costs have increased sharply. Thisled to a sharp decline in the numberof operating mines and a 48% decline in thenumber of employees between 1988 and 1998.Reductions in unit costs will only comethrough increases in productivity, in particularthrough mechanization of mining. The taxa-tion arrangements under which the minesoperate are complex but favored the treatmentof large tonnages of low grade ore. In 1999 theformula was y = 43 – 215/x where x is the ratioof taxable income from gold to gross goldrevenue and y the tax rate.

14.3.2 Organization of mining

Mining on the Witwatersrand requires largeamounts of capital and has been dominated bylarge companies since its inception. The majoroperators in 2004 were Anglogold Ashanti,Gold Fields, Harmony Gold Mining, and SouthDeep. These companies have evolved fromthe major mining houses that dominated theSouth African mining industry until 1994:Anglo American Corporation, JohannesburgConsolidated Investments (JCI), Gold Fields ofSouth Africa (GFSA), Rand Mines, Gencor,and Anglovaal. After 1994 Anglo Americanmoved its headquarters to London and createda gold holding company, which has progress-ively increased its share of non-South Africanoperations. A number of loss-making or short-life operations were sold, although Anglogoldstill produces 106 t from South Africa. GoldFields, created by the merger of Gold Fieldsof South Africa’s and Gencor’s gold interests,is pursuing a similar path, although it operatesthe major mines at Kloof, Driefontein, andJoel. Harmony is a relatively new companycreated from the remnants of Rand Minesand short-life mines sold by other major pro-ducers, as well the major developing Targetmine. South Deep, discussed in detail below,is a joint venture between Western Areas, asuccessor company to JCI, and Placer Domeof Canada. To add to the confusion since 1994,many mines have been split and renamed, e.g.

FIG. 14.7 Plots of (a) profit margins, (b) goldproduction, and (c) gold production, tonnes milled,and grade for the Witwatersrand Basin. (From Willis1992, updated from Handley 2004.)

the overall recovered grade has declined,largely as a result of the exhaustion of highergrade deposits (Fig. 14.7), and South Africahas suffered sustained high inflation rates.In particular, working costs have increaseddue to higher wages and the increasing depth

Ore milled (tx107)Gold production (kgx105)Recovery grade (g t–1)

3456789

1011121314

(c) 15

1900

1910

1920

1930

1940

1950

1960

1970

1980

1990

80

70

60

50

40

30

20

10

Pro

fit m

argi

ns (

%)

AV 1902 – 1971(a)

(b)

1914

1884

1944

1974

1999

0

200

400

600

800

1000

Tonn

es

1960

1965

1975

1980

1985

1990

2000

2000

1995

1970

Page 349: Introduction to mineral exploration

332 C.J. MOON & M.K.G. WHATELEY

Western Deep Levels is now split into Savuka,Mponeng, and Tau Tona, and Vaal Reefs intoGreat Noligwa, Tau Lekoa, Kopanang, andMoab Khotsong.

14.3.3 Mineral rights

Mineral rights on the Witwatersrand werelargely privately held, either by gold miningcompanies or in unmined areas by farmersor descendants of the original landowners.The value of these assets is usually appreci-able and purchase values may be as high asR10,000 ha−1; options to purchase are negoti-ated, usually lasting for 3 or 5 years (Scott-Russell et al. 1990). The competitive natureof exploration means that options were oftennot easy to obtain and option costs mightrocket as soon as news leaks out of a company’sinterest. In many cases the mineral rights weredivided into various shares and agreement wasrequired by all parties. However the system is,at the time of writing (2004), undergoing radi-cal change. South Africa is moving to a systemin which the state will control explorationrights in a similar fashion to most of those inNorth America and Australia. Under the pro-visions of the 2002 Mineral and PetroleumResources Development Act new explorationareas will be granted for 5 years. Existing min-

eral rights, either of mining or exploration, canbe converted into the new system but only ifthey are actively being used.

14.3.4 Mining methods

The generally narrow mining widths, availabil-ity of cheap labor, moderate dip of the reefs,and great depths of most mines have led to thedevelopment of a very specific mining method,known as breast stoping. In this method(Fig. 14.8), development is kept in the footwallof the reef and a raise is cut between levels.From this raise gullies are cut at a 45-degreeangle to the raise and the stope developed in aChristmas Tree manner. The stopes advanceby jackleg drilling and blasting of a 2-m lengthof stope on each shift. Ore is scraped outfrom the advancing face along the gully usingmechanical scrapers and the stope is cleanedcarefully, often using high pressure jets, asmuch gold is contained in dust in the cracks inthe floor of the stope. The scraper transfers theore to footwall boxholes whence it is trammedto the main ore pass and the primary under-ground crusher. If the stope takes in morethan one level then it is known as a longwall.The hanging wall of the advancing stope issupported by hydraulic props, later reinforcedby timber packs.

Dip of reef

1000

0

2000

3000

Dep

th (

m)

Ventilation shaftProcessing plant

Main shaftSURFACE

LEVEL 1

LEVEL 2

LEVEL 3

Subverticalshaft

ShaftX – cut

Systematicpack support

Stope face

Gullies

Ore pass to level belowRAISE

RAISE

Footwall X – cut

Ore pass

X – cutto reef

Overlying rocks

REEFREEF

Directionof advance

FIG. 14.8 Diagram showing a typical Witwatersrand mine. (From International Gold Mining NewletterApril 1990.)

Page 350: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 333

Mechanized mining of narrow stopes isnot easy and mechanization has been mostsuccessful in wide stopes, such as those inthe Elsburg reefs on the West Rand, which areoften 2 m or thicker. Trackless mining wasfirst introduced in 1984, and is seen as a way ofreducing the working costs, using specializedmachinery instead of the labor-intensive con-ventional methods. Trackless mining involvesthe substitution of tracks, locomotives, and rail-cars by free-steering, rubber-tyred (or caterpillartracked) equipment for drilling, loading, andtramming of the ore, waste, men, and mater-ials. The substitutions involved are shown inTable 14.1.

The potential benefits of such conversionsare lower operating costs, higher productivity,greater flexibility, and improved grade controland safety (particularly in stopes, on haulagesand declines, and at grizzlies). There is a con-comitant change in the workforce, with thelarge semi- or unskilled workforce replaced bya much smaller, highly skilled workforce.

There are a number of variations in theway trackless mining is carried out, but thegeneral use of room (bord) and pillar miningusing Tamrock and Atlas Copco drillingrigs and LHD (load, haul, dump) vehicles iscommon to most. There are two stages ofproduction, namely primary development andstoping followed by secondary extraction ofpillars. During primary development and stop-ing, access ramps are driven on reef betweentwo levels approximately 150 m apart. Stopingcommences with the development of roomsapproximately 17 m apart off the access ramp.Pillars 10 × 7 m are left between rooms. Drill-ing of blastholes is by means of boom drill rigs.The reef is blasted out into the heading makingthe broken ore accessible to the LHD vehicles.

During this stage of mining 71% of the oreis extracted, the remaining 29% is left in thepillars. Secondary partial extraction of thepillars takes place once primary stoping is com-plete. The pillars are reduced to a final size of5 × 5 m. The area is supported by roof bolts,and the broken ore is again removed by LHDvehicle.

A major consideration at great depths isrock stability. The quartzites have consider-able strength and deform in a brittle manner,causing rockbursts (explosive disintegration ofrock). These events cause disruption to produc-tion and are a major cause of fatal accidents.Rockburst frequency can be reduced by backfilling of stopes and careful monitoring of stressbuild ups.

14.4 EXPLORATION MODELS

Exploration has been carried out on a variety ofscales ranging from delineating extensions ofknown pay shoots within a mine to the searchfor a new goldfield. As most of the shallowareas have been explored one of the key deci-sions to make before optioning land is to decidethe depth limit of economic mining and explo-ration. The deepest mining planned in detail atpresent is about 4000 m but options have beentaken over targets at up to 6000 m and holesdrilled to 5500 m. According to Viljoen (1990)good intersections have been made at 5000 mand although there is of course a stratigraph-ical limit, mineralisation is present at greatdepths. The major constraints are the stabilityof mine openings, rock temperature (at 6000 mit is likely to be >85°C) and the costs of miningat these depths. At present, temperature prob-lems are overcome by circulating refrigerated

TABLE 14.1 Equipment substitution in trackless mining.

Operation Conventional mining Trackless mining

Drilling Pneumatic and hydraulic Hydraulic drill rigshand-held drills

Stope cleaning Scraper winches LHDsGully cleaning Gully winches Trucks or LHDsTramming Locomotives and railcars Trucks or LHDsLight haulage and railcars Locomotives Utility vehicles

LHDs, Load, haul, dump vehicles.

Page 351: Introduction to mineral exploration

334 C.J. MOON & M.K.G. WHATELEY

air and chilled water and amounts of these haveto be increased substantially at 5000–6000 m.Currently the deepest single lift shaft is lessthan 3000 m (South Deep) and mining to 5000–6000 m would require two further shafts under-ground, termed secondary or subvertical shaftsand tertiary shafts. Therefore every tonne ofore and materials would have to be transhippedat least twice. At 3000 m depth it can take aminer one and a half hours to reach the min-ing face and consequently production time isreduced. Mining at 6000 m may require a revi-sion of shift patterns.

Most exploration has taken place withinthe known extent of the Witwatersrand Basinalthough there have been a number of searchesfor similar rocks outside it. The latter searchhas particularly concentrated on areas wherelithologies similar to those in the basin areknown either in outcrop or at depth. The pos-sibility of West Rand Group correlatives isusually based on magnetic anomalies and theexploration model is that of a basin withinmineable depth and preferably closer to post-ulated source areas. Examples of Witwater-srand lithologies at depth are known from deepdrillholes and from surface outcrops to theNorthwestern and Northern Natal-Kwazulu.

Within the basin, exploration is aimed atdetecting the continuation of known reefs ornew reefs within known goldfields or of detect-ing new goldfields in the intervening (gap)areas. The 1980s saw a great deal of explora-tion in the gap areas based on the premise thatprevious deep drilling has missed potentialgoldfields due to lack of understanding of thestructure or controls on sedimentation. Thefundamental information required is an under-standing of the subsurface stratigraphy and thepotential depths of the target. Upthrown blocksof the Central Rand Group are also a target.

The detection of extensions to known gold-fields, either down dip or laterally, is a lessdifficult task as considerable information isavailable from the other mines. In this case itis possible to collate the grade distributionwithin an individual reef and project pay-shoots across structures as shown in Fig. 14.9.In this example sedimentation is controlledby synsedimentary structures with the highsforming areas of nondeposition or of low grade.The target areas in the structural lows can be

clearly established by structural contours andisocons of grade distribution.

In 1990, JCI were spending 5% of theirexploration budget on the search for newbasins, 7% on exploration for new goldfields,12% on basin edge deposits, and 26% on down-dip extensions. The remaining 50% was spentin the search for extensions to known mines(Scott-Russell et al. 1990). In 2003 spendingwas very limited with Anglogold spending onlyUS$2.4 million on exploration.

Environmental constraints on explorationhave, in the past, been minimal as theWitwatersrand Basin mainly underlies flatfarmland, largely used for maize and cattle pro-duction. In spite of the obvious problems gener-ated by the extensive use of cyanide, there hasbeen little concern about pollution. Forstnerand Wittman (1979) however demonstrate thatthere is major contamination of surface watersassociated with oxidation of pyrite and thathigh Zn, Pb, Co, Ni, and Mn are present as aresult of this oxidation and the cyanidationprocess. Cyanide itself is not a problem as itbreaks down under the influence of ultravioletlight and levels are within those specified inthe South African Mines and Works Act. Alsoof immediate concern has been the dust gener-ated by the erosion of the slimes dams. Thiscan be countered by revegetation (Friend 1990).The rehabilitation of surface workings has nowbeen given legal force in the Water Act and theMinerals Act of 1991.

14.5 EXPLORATION METHODS

The original discovery of gold-bearing con-glomerate was made in 1886 on the farm Lang-laagte by two itinerant prospectors, GeorgeWalker and George Harrison, who were restingon their way to the Barberton greenstone gold-field (Werdmuller 1986). While engaged inbuilding a house for the owner they went for awalk and Harrison, who had been a miner inAustralia, recognized the potential of the iron-rich conglomerate. He crushed and panned asample that gave a long tail of gold. They wereno doubt encouraged to prospect by the previ-ous activities of Fred Struben, who was granteda lease to prospect veins in 1885 (Robb & Robb1995).

Page 352: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 335

F

Explorationtargets

F

F

F

F

GRADEWest Rand regional anticline or culmination

Major syndepositional anticline with local direction of plunge

Major fault showing ground loss

Mining lease areas

W.R.C.M.

L.E.G.M.

E.C.D.G.M.

D.R.D.

R.E.G.M.

Doornkop section ofRandfontein Estates

High

Medium

Low

F

N

0 2 km

FIG. 14.9 Integration of regional grade distribution and structure to generate areas for drilling, West Rand.(After Viljoen 1990.)

14.5.1 Surface mapping and down-dipprojection

After the original discovery the surface expos-ures were quickly evaluated by surface pro-specting and the outcrop of the conglomeratesmapped. In contrast to some expectations theconglomerates were not a surface enrichmentbut continued down dip. However this pres-

ented a major problem as the surface iron oxidesturned to pyrite at around 35 m and, as all pro-spectors knew, this meant poor recoveries ofgold by the recovery process then used, amalga-mation of the gold with mercury. Fortunatelythe cyanide process of extraction had beendeveloped by three Scots chemists in 1887 andwas immediately demonstrated to be effect-ive. In contrast to ores from greenstone belts,

Page 353: Introduction to mineral exploration

336 C.J. MOON & M.K.G. WHATELEY

Witwatersrand ore gives good recoveries due tothe presence of gold as small free grains withvery little or no gold contained in arsenopyrite,pyrite, or tellurides. The first major trial ofcyanidation on tailings from the amalgama-tion process from the Robinson Mine yielded6000 oz (190 kg) of gold from 10,000 tons. Thesuccess of cyanidation prompted new interestin the possible depth continuation of the reefs.This depth continuation was tested by drillingand the first drillhole in April 1890 intersectedthe South and Main reefs at 517 ft (157 m) and581 ft (177 m) with grades of 9 oz 12 penny-weights (dwt) t−1 (328 g t−1) and 11 dwt t−1

(18.8 g t−1). Further drillholes, notably the RandVictoria borehole (Fig. 14.10), suggested in1895 that the reef decreased in dip from the 60–80 degrees at outcrop to 15 degrees at 2350 ft(715 m) and that grades persisted at depth. Thelarge potential of the area down dip was recog-nized by the mining houses with capital tofinance mining at depth, and much of the areadown dip was pegged by Goldfields and CornerHouse companies (Fig. 14.10). This down-diparea provided the basis for a number of majoroperations of which the largest, Crown Mines,eventually produced 163 Mt of ore and 1412 tof gold, an average grade of 8.7 g t−1 Au.

For the 40 years after the initial discovery ex-ploration and mining concentrated on the

areas down dip from the outcrop of conglo-merates on the Central, West, and East Rand.Exploration involved drilling and develop-ment down dip from outcrop; by 1930 theVillage Main Mine had reached 710 m belowcollar. In addition, although the average reefwas payable, it was noticed that there wereconsiderable variations in grade. Pioneeringstudies by Pirow (1920) and Reinecke (1927)demonstrated that these higher grade areasformed payshoots and could be correlated withlarger pebbles and thicker reefs. Reinecke pro-posed that the control on the payshoots wassedimentary and paralleled the direction ofcross-bedding.

14.5.2 Magnetics and the WestWits Line

The next stage in the exploration historywas the search for blind deposits under laterTransvaal and Karoo cover to the west of theWest Rand Fault which forms the westernlimit of the West Rand Goldfield. Althougha number of geologists were convinced of thecontinuation of the Main Reef to the west ofthe known outcrop and it had been tested bydrilling in the period 1899–1904, results wereunconvincing. A shaft sunk in 1910 to testthese reefs had to be stopped at 30 m due to

BezuidenvilleB /H (1898)

0 2 km

Turf ClubB /H (1901)

Rand VictoriaB /H (1892)

JOHANNESBURG

Main ReefOutcrop

MAIN REEF

±1500 m level

±1800 m level

Consolidated Gold Fields

The Corner House

Stoping on Main Reef Group

N

FIG. 14.10 Exploration and mining on the Main Reef, about 1900. (After Viljoen 1990.)

Page 354: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 337

Granite

Witwatersrand Supergroup

Chuniespoort Group

Pretoria Group

Outcrop of Main Reef

Subcrop of Main Reef

Probable position of subcrop of Main Reef

Magnetic Shales of the West Rand G. outcrops

Magnetic Shales of the West Rand G. subcrops

Dykes Shafts BoreholesFaults

1

12

23

3

1 2 3

1 2 3

4

5

67

1 2

1 3234

45

6

6

7

7

89

78

9

98

654

32

1

12

34 567

89

1234

5

678

9

12

3

456

78

9

987

6

Ven

ters

post

D

yke

N26°00'S

26°15'S

26°30'S

27°15'E 27°30'E 27°45'E

Gats Rand Range

Eastern Section

Roodepoort Fault

Moo

i Riv

er

Wes

t R

and

Fau

lt

Water Tower SlatesContorted BedLower Hospital Hill ShalesMiddle Hospital Hill ShalesUpper Hospital Hill ShalesPromise BedsWest Rand ShalesGovernment Reef ShalesJeppestown Shales

Bank Fault

MiddlevleiStation

0 5 km

112233445566778899

Magnetic anomalies caused by

CMI

E4

Witp

oortjie Fault

FIG. 14.11 Exploration on the West Witwatersrand line 1936. Drillhole E4 is the Carbon Leader discovery holeon the farm Driefontein, now (West) Driefontein Mine, and CM1 the discovery hole on the farmBlyvooruitzicht. (After Englebrecht 1986.)

an uncontrollable inflow of water from theTransvaal dolomites.

Wildcat drilling was prohibitively expensiveand the search for the subcrop of the Main Reefwas an unsolved puzzle for the period 1910–30.The solution to this problem was found bya German geophysicist, Rudolph Krahmann.One day in 1930 at a motor rally he noticed thatthe rocks of the West Rand Group containedmagnetite and he thought that it would bepossible to trace these shales under cover usingnewly developed magnetometers. After check-ing the theory over exposed West Rand Groupsediments five major and four minor anomalieswere found within the sequence (for fullerdetails see Englebrecht 1986 or the very read-able account of Cartwright 1967). The depthpenetration of the method was soon testedwith backing from Gold Fields of South Africawho had also acquired the rights to the

cementation process that allowed shaft sinkingthrough water-bearing formations. Financealso became more easily available with theincrease in the gold price from £4.20 to £6 perounce. The initial 11 drillholes (Fig. 14.11)down dip from the Middlevlei inlier not onlyintersected what was believed to be the equi-valent of the Main Reef but also a further con-glomerate at the base of the Ventersdorp lavas,termed the Ventersdorp Contact Reef. Theseholes gave rise to the Venterspost and LibanonMines, but the real excitement came with re-sults further to the west between the majorBank Fault, subparallel and west of the WestRand Fault, and Mooi River (Fig. 14.11). Drill-ing on the farm Driefontein began with threeholes designed to cut the Main Reef. The firsttwo holes were to the north of the subcrop ofthe Main Reef against the dolomites whilethe third cut poor values within the Main Reef

Page 355: Introduction to mineral exploration

338 C.J. MOON & M.K.G. WHATELEY

but intersected a carbon seam with visiblegold some 50–60 m below the Main Reef. Thesignificance of this was not appreciated atfirst as core recovery in the very fragile carbonseams was incomplete due to grinding of core.This carbon seam later became known as theCarbon Leader Reef. In the original calcula-tions of viability a grade of 13.6 g t−1 over 1.1 mwas assumed, based on previous working onthe Main Reef. When the first shaft was sunkby Rand Mines on the farm Blyvooruitzicht, asmall piece of ground that Gold Fields did notown, the nature of the discovery soon becameclear. Underground sampling around the shaftarea in 1941 returned a grade of 69 g t−1 Auover 70 cm compared with a drillhole grade of23.1 g t−1 over 55 cm. The mine managementat first refused to believe the results and hadthe shaft resampled only to yield the sameresults. Similar grades were obtained whenthe GFSA mine at West Driefontein openedin 1952. Exploration to the west of the Bankbreak, along an area known as the West WitsLine, largely by GFSA, led to the developmentof 10 mines. Although magnetics led to theestablishment of the subcrop of the Main Reef,the structural complexity of the area was notappreciated until the 1960s when the angularunconformity between the Main Reef and theVentersdorp lavas was described. In the northand east of the area, the upper parts of theCentral Rand Group have been removed by pre-Ventersdorp erosion. In addition, parts of theCarbon Leader have been removed by erosionand distinct barren areas can be recognized(Englebrecht et al. 1986).

When the importance of the finds on theimmediate subcrop of the West Wits Linebecame apparent companies were quick to ap-preciate the down-dip potential. In particularAnglo American Corp. made a deliberate deci-sion in 1943 to explore for deposits at 3000–4000 m and obtained the mineral rights tothe area known as Western Deep Levels (Fig.14.2). This is currently the deepest mine inthe world and was based on 12 drillholes withan average intersection in the Carbon Leader of3780 cm-g t−1 and additional mineralisation inthe VCR. Production in 2003 was ~41 t of gold.

Although the magnetic method was out-standingly successful on the West Wits Line,detailed studies in other areas of the basin

have encountered problems in interpretation.In particular the lateral and vertical faciesvariations of the beds are not understood.Magnetic anomalies can also be due to beds,other than Witwatersrand sediments, such asmagnetite-rich layers in basement (Corner &Wilsher 1989).

14.5.3 Gravity and the discovery of theOrange Free State Goldfield

The increase in the gold price in the early1930s encouraged companies to explore areaseven more remote than the West Wits Line. Inparticular Anglo American made a discovery(now the giant Vaal Reefs Mine complex) inthe Klerksdorp area, immediately north of theVaal River, which forms the border betweenthe Transvaal and Orange Free State provinces(Fig. 14.2). This discovery encouraged othercompanies, particularly Union Corporation, toexplore in this area and also to the south ofthe Vaal River, where it was theorized thatthe Central Rand Group would extend undercover. In the Klerksdorp area Union Corpora-tion had demonstrated for the first time thatgravity methods could be used to predict thedepth of the Central Rand Group as theirdensity was significantly lower at 2.65 kg m−3

than the Ventersdorp lavas at 2.85 kg m−3. Thistechnique was selected to improve interpreta-tion south of the Vaal River.

The area to the south of the Vaal River,particularly that around the village ofOdendaalrust (north of Wesselbron, Fig. 14.2),had attracted the attention of a number ofprospectors as it had been demonstrated asearly as 1910 that gold could be panned fromconglomerates that are stratigraphically partof the Ventersdorp Supergroup. A small ven-ture took up the submittal from the originalprospector and drilled a hole on the farmAandenk to test the conglomerates. Althoughit was obvious from the beginning that theconglomerates were stratigraphically part ofthe Ventersdorp Supergroup, the venture per-sisted and intersected the top of the CentralRand Group quartzites in 1934 at a depth of2726 ft (831 m). The hole eventually bottomedat 4046 ft (1234 m) with a best intersection of120 in-dwt t−1 (521 cm-g t−1) when they ran outof funds.

Page 356: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 339

On the basis of the results from the Aandenkborehole and the success of the gravity methoda company, Western Holdings, was floatedby South African Townships to prospect inthe Southern Free State in collaboration withUnion Corporation. Intensive geophysical sur-veys showed a significant gravity anomaly overthe farm St Helena flanked to the west by mag-netic anomalies (Fig. 14.12). The first hole SH1,drilled in 1938, indeed cut Central Rand groupquartzites at 302 m and was drilled for another1850 m cutting three reefs at 200–300 in-dwt t−1

(860–1300 cm-g t−1). Further drilling to the westconfirmed the location of the magnetic shales.By this stage a number of other companieswere drilling in the area but the stratigraphicalcontrol of the mineralisation was unknown.After routine assaying of a conglomerate one ofthese companies obtained an assay of 6.9 g t−1

over 1.05 m in what is now known to be themajor gold-bearing conglomerate. Within a fewmonths Western Holdings intersected payablevalues of the same reef in hole SH7 at 59 g t−1

over 1.5 m at a depth of 348 m. These were thefirst intersections of the Basal Reef (Fig. 14.3),which has been the main producing placer inthe Orange Free State Goldfield. Proving theexact positions of the reefs was howeverdifficult as the goldfield is cut by a number ofmajor faults and interpretation of gravity datawas not as simple as first thought. Gravityanomalies could reflect the depth of the Cen-tral Rand Group quartzites but they also couldbe due to faulting, folding, and changes in thelithology of the Ventersdorp Volcanics. How-ever experience showed that the Central RandGroup anomalies could be recognized by theirsharp gradients, whereas changes due to theVentersdorp were much more gradual. Furtherexploration and exploitation of the goldfieldwas delayed by World War II but resumption ofexploration after the war confirmed the poten-tial of the area including a phenomenal inter-section on the farm Geduld of 3377.5 dwts t−1

(5775 g t−1) over 6 inches (15 cm). Although notthe active partner in the original explorationventure, Anglo American Corp. became thedominant operator in the OFS goldfield by tak-ing over Western Holdings and several othercompanies immediately after 1945.

Similar techniques were successful in thediscovery by Union Corporation of the Evander

Sub-basin, about 100 km east-south-east ofJohannesburg (Fig. 14.2). Although the areahad been known to contain concealed CentralRand Group rocks, it was only the advent ofairborne magnetic and ground gravity sur-veys that unravelled the structure. Persistencewith a drill program eventually proved thatthe Kimberley Reef was the only payable reefand led to the establishment of four mines(Tweedie 1986).

14.5.4 Reflection seismics, gaps, andupthrown blocks

One of the major advances of the 1980s was theuse of seismic reflection surveys in detailingthe structure and location of the Central RandGroup rocks at depth. Although use of themethod was mooted by Roux in 1969, it wasthe impact of improved technology and lackof work in the oil industry that prompted arenewed trial by oilfield service companies incollaboration with the major mining houses.Initial logging of boreholes showed that therewas a contrast in seismic velocities as confirmedby initial surface surveys (Pretorius et al. 1989,1997). Although it is not possible in reconnais-sance surveys to define the individual reefs, thesurveys are able to define major contacts, par-ticularly the base and top of the Central RandGroup and the contact between the Transvaaldolomites and the Pretoria shales overlyingthem. By 1989 Anglo American had coveredthe basin based on more than 10,000 line-km of2D survey and further coverage concentratedon the detailed follow-up of target areas. Othermining houses use seismic surveys extensivelyand JCI (Campbell & Crotty 1990) have used3D surveys in planning mining layouts and thelocation of individual longwall faces at theirSouth Deep prospect (section 14.7.2). The costof 3D surveys limits their use; Campbell andCrotty report that the survey at the South Deepprospect cost R2.2M, the cost of one and a halfundeflected drillholes.

Pretorius et al. (1989) defined three differ-ent types of seismic survey used by AngloAmerican during the 1980s:1 Reconnaissance 2D surveys of stratigraphyand structure on public roads concentrating onthe gap areas where the target beds are relat-ively shallow.

Page 357: Introduction to mineral exploration

340 C.J. MOON & M.K.G. WHATELEY

FIG. 14.12 Gravity and magnetic anomalies over the St Helena Gold Mine, OFS Goldfield. Note how thediscovery hole (SH7) is on the flank of the main anomaly. (From Roux 1969.)

SH7

SH1X Y

Contour intervals 0.5 milligals

X Y

St Helena Gold Mine Karoo

Ma

gn

et ic

anomaly

7.0

6.5

6.05.5

6.05.55.0

4.03.53.02.52.0

4.5

1.5

4.5

4.0

3.5

4.54.0

3.5

4.5

-0.50.0

2.0

1.0

1.0

2.0

3.0

4.0

3.0

St Helena G.M.

PresidentBrand G.M.

PresidentSteyn G.M.

4.03.02.0

Ground magneticprofiles

Gravity profiles

0

4

8

0100200300400

Airborne magneticprofile

0 2km

Ventersdorp Conglomerate

Central Rand Group

West Rand Group

Granite

Magnetic bed

Karoo Supergroup

Ventersdorp Sediments

Ventersdorp Lava

Mill

igal

sN

0 2 km

Ventersdorp Lavas Central Rand Group West Rand Group

Page 358: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 341

2 Reconnaissance 2D of deeper areas withinthe basin where targets might be upthrown tomineable depths by faulting.3 Detailed 2D surveys of extensions to re-serves on existing mines.

Few examples of seismic sections have beenpublished, but the paper by Pretorius et al.(1989) provides examples of the use of thetechnique in reconnaissance exploration. Thearea shown in Fig. 14.13a is on the westernedge of the basin and the seismic surveyclearly delineates the upper and lower con-tacts of the Central Rand Group and faultingwithin the Transvaal sediments. The pro-jected depth of the VCR at the upper contactis over 5000 m and drill targets can be assignedon this basis. The contrast between theseismic section and the previous interpreta-tion of the gravity data is striking (Fig. 14.13a)In the original gravity interpretation the3 mgal anomaly superimposed on the strongregional gradient was attributed to a horst ofCentral Rand Group sediments at a depthof around 2 km whereas the seismic sectionshowed that there was no major faulting,that the Central Rand target was at 5 km,and the gravity low was the result of changesin lithology of the Pretoria Group sediments.Thus seismic and gravity data can be success-fully integrated.

Figure 14.13b shows the result of a recon-naissance seismic reflection survey in themiddle of the basin that detected an upthrownblock of Central Rand Group sediments thatare at mineable depth (<3000 m); this was con-firmed by later drilling.

The third type of survey is exemplified bythe seismic survey which defined the struc-

ture at the margins of an established mine(Fig. 14.13c). From this section an additionalarea of Central Rand Group sediments at mine-able depths was established.

14.5.5 Drilling

One of the major financial constraints onexploration is drilling. A 6000-m-deep holedrilled from surface may take 18 months tocomplete. A typical 15-hole program on a leaseof 7000 ha at 3–4000 m target depth wouldtake 20 years to complete if the holes weredrilled sequentially. This posed considerableproblems, not only of capital, but also of deci-sion making, as mineral rights options are only3 or 5 years in length and a decision to exercisethe option must be taken at the end of theperiod. Clearly programs must have severalrigs drilling simultaneously or drilling mustbe accelerated. Each drillhole will be deflectedto provide a number of reef intersectionsand improve confidence in grade estimation(Fig. 14.14, Table 14.2).

Perhaps the key decision to be made isapproximately how many intersections needto be made before a confident estimate can bemade of a lease’s value or lack of prospectivity.A quantitative approach to the problem hasbeen given by Magri (1987), whose results aresummarized in Fig. 14.14. Using an approachsimilar to Krige’s t estimator to predict themean of log-normally distributed data, he hasproduced confidence limits of estimates ofthe mean grade for reefs of immediate interestto his company (Anglovaal) based on know-ledge of the grade distribution in mined blocks.These clearly show that making further drill

Depth Number of deflections(m)

1 2 3 4 5 7 9

1000 93 118 131 144 158 188 2201500 150 176 190 205 220 253 2882000 220 249 267 286 306 348 3932500 303 337 359 382 407 457 5113000 408 446 473 501 530 592 6583500 545 590 624 660 698 777 8624000 714 768 812 858 905 1004 11114500 907 962 1007 1053 1102 1204 1315

TABLE 14.2 Costs of drillholes anddeflections, in 000 rand 1987 value(1R ===== US$0.3). (From Magri 1987.)

Page 359: Introduction to mineral exploration

342 C.J. MOON & M.K.G. WHATELEY

FIG. 14.13 (a) Use of reflection seismics in discriminating between two possible gravity models for targets onthe western margin of the Witwatersrand Basin. The left-hand model used a block of upthrown Central RandGroup (A) to explain the gravity low while the right-hand model, confirmed by seismics, invoked near-surfaceTransvaal sediments.

GLI

VCR

West Rand Group / Intrusive (∆p = 0.10)

Central Rand Group (∆p = 0)

Bou

guer

ano

mal

y (m

gals

)50

40

30

20

10

0

Observed Gravity FieldCalculated Gravity Field

GLI

Bou

guer

ano

mal

y (m

gals

)

0

10

20

30

40

50

60

Observed Gravity FieldCalculated Gravity Field

Density Contrast (g cm –3)∆p Density Contrast (g cm –3)∆p

0 10 km

A

0

4

8

12

16

Dep

th (

km)

0 10 km

0

4

8

12

16

Dep

th (

km)

Area of seismic control

Pretoria Group (∆p = 0.15)

Malmani Subgroup (∆p = 0.15)

Klipriviersberg Group (∆p = 0.10)

Central Rand Group (∆p = 0)

West Rand Group (∆p = 0.10)

Basement & Dominion Group (∆p = 0)

Upper Silverton Fm. (∆p = 0)

Pretoria Group (∆p = 0.15)

Transvaal Dolomite (∆p = 0.22)

Klipriviersberg Group (∆p =0. 14)

West Rand Group (∆p = 0.15)

Basement & Dominion Group (∆p = 0)

0 5 km

Dep

th (

km)

15

10

5

0

Basement

West Rand & Dominion Groups

Central Rand Group

Klipriviersberg Group

Malmani Subgroup

Timeball Hill Fm.

Hexpoort Fm.

Silverton Fm.

(a)

Page 360: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 343

14.5.6 Underground exploration

Considerable underground exploration is nec-essary before a mining layout can be designedthat will minimize losses of production fromfaulting and dykes. Figure 14.15 shows the lay-out of a typical underground drilling programin a faulted area.

14.5.7 Re-treating old mine dumps

One of the most significant innovations onthe Witwatersrand has been the major projectsto re-treat old slimes dams and sand tips.The reasons for the occurrence of the goldin waste are many; gold may be enclosed inpyrite or cyanide-resistant minerals that werenot crushed or the operation of the plant wasnot optimal for the mineralogy treated. The

intersections after a certain cut-off, whichvaries for each reef, provides little extrainformation. In the example shown this is12 or 13. In the same way it is possible tooptimize the number of deflections from eachdrillhole – usually four are made. Magri sug-gested that the optimum approach to drillingan area is to decide on a minimum number ofholes necessary to give a picture of an area,say 10 for a 7000-ha lease and review the datawhen drilling is complete. If the grade of themean less confidence limit for the prospectis greater than the pay cut-off, then the pro-gram should be continued. If not, it should beabandoned.

The application of drilling rigs used foroil and gas exploration has helped to speedexploration and to increase the accuracy ofdeflections from the mother hole.

(b)

0

Dep

th (

km) 5

10

0 5 km

Karoo Sequence

Hekpoort Fm.

Timeball Hill Fm.

Malmani Subgroup

Pniel Group

Klipriviersberg Group

Central Rand GroupWest Rand &Dominion GroupsBasement

Upper West Rand Group

Bonanza Fm.Lower West Rand& Dominion GroupsBasement

(c)

Pniel Group

Platberg Group

Klipriviersberg Group

Central Rand Group

0 5 km

0

Dep

th (

km)

5

10

FIG. 14.13 (b), (c) Typical uses of seismic reflection data: (b) detection of upthrown blocks; (c) detection ofunsuspected upthrown blocks adjacent to a lease area. (From Pretorius et al. 1989 with permission)

Page 361: Introduction to mineral exploration

344 C.J. MOON & M.K.G. WHATELEY

Reef

Length of deflection 100 m

Angle of deflection 1°

One set of deflectionsper borehole

0

50

100

150

200

250

Con

fiden

ce li

mits

rel

ativ

e to

mea

n =

100

5 10 15 20 25

12345

15

1 One intersection per borehole2 Two intersections per borehole3 Three intersections per borehole4 Four intersections per borehole5 Five intersections per borehole

Number of boreholes

FIG. 14.14 Schematic diagram of deflections andplot of confidence limits for increasing number ofdrillholes and deflections per hole. There isconsiderable variation between placers, the caseshown is the Vaal Reef. (From Magri 1987.)

FIG. 14.15 Section showing typical undergrounddrilling to determine reef location and fault throws,Vaal Reefs Mine. The length of the section isapproximately 70 m. (After Thompson 1982.)

LOCATION 2# 74 LEVEL NE

73.5 46 x/c

74.5 46 x/c

Fault Diamond dri l lhole Reef

Drive

allowed profitable reworking of this waste, andby 1986 about 11 t of gold per year were pro-duced from this source.

14.6 RESERVE ESTIMATION

Reserve estimation on a gold mine occurs intwo phases, namely those resources calcul-ated during exploration and those reservesestimated during exploitation. Ore reservescalculated during exploitation refer to thoseblocks of ore that have been developed suf-ficiently to be made available for mining.Usually, a raise would have been driven andsampled. The samples selected for use in thereserve calculation are those which are abovethe cut-off assigned according to the mine’svaluation method (Janisch 1986, Lane 1988).The valuation of a developed reserve block hasbeen traditionally carried out using the aver-age grade of all the samples taken along theraises and the drives. This is steadily givingway to kriging methods (Krige 1978, Janisch1986).

In exploration, resource estimation is notso precise. It is essential to identify the blockswhich could be mined. These blocks wouldmost probably be bounded by faults. Hencethe importance of establishing the structureof an area from the borehole information, usingstructural contour maps and a series of cross-sections. More recently this information hasbeen supplemented with information from 2Dand 3D seismic surveys (Campbell & Crotty1990, Pretorius et al. 1997).

average gold content in the waste is probablyabout 0.3 ppm but waste from the older mineson the East and Central Rand contains appro-ximately 0.7 ppm. Advances in metallurgicaltechniques and increases in the gold price have

Page 362: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 345

The average thickness and grade for eachblock must be calculated. Each borehole hasan average of four deflections which intersectthe reef, often in random directions (Fig. 14.14),and this gives an unbiased sampling of thereef within a radius of about 5 m of the originalhole. When an inclined hole intersects adipping reef, only an apparent dip of the reefcan be seen in the core, so corrections must bemade, using the following formula, to calculatethe true thickness, to be used in the tonnagecalculations:

W = cos b {C(sin d − [cos(90 − e)tan b]}

where W = true reef thickness, b = reef dip,C = apparent reef thickness, d = boreholeinclination, and e = azimuth of reef dip.

A Sichel’s t average grade value (Sichel 1966)is calculated for each cluster of acceptablesample values. Each value thus calculated iseither assigned to the block into which theborehole falls (modified polygonal method;Chapter 9), or the averages of all the holes,also calculated using the Sichel’s t estimator,is used as the average grade of the reef for theentire area being assessed.

Usually a mine has a minimum mineablereef thickness, often of 1 m. Where reef inter-sections are less than 1 m the grade is recalcu-lated to reflect the additional waste rock thatwould have to mined to achieve the minimummineable stoping height and this height (in thisexample, 1 m) is used in tonnage calculations.Often, to establish whether the minimummining height is going to be economic to mine,an accumulation or grade thickness product iscalculated, such that ax = Wx × gx. This accu-mulation is the product of the true thicknessW in the sample at x and the average grade, gx.In the gold mines, thickness is quoted incentimeters, the grade in g t−1, and the pro-duct in cm g t−1. Where a sample is shown tohave 35 g t−1 over 10 cm, the accumulationis 350 cm g t−1. To recalculate the grade overthe minimum mining width of 100 cm (1 m)the accumulation is divided by the miningwidth and in this case the grade over 100 cm is3.5 g t−1. If the cut-off grade for this particularmine happened to be 3 g t−1, then this blockwould be considered payable and the block

included in the reserve calculations using the100-cm mining width.

The tonnage is calculated for a block whichhas a borehole at the centre using the followingformula:

M = W × A × r

where M = mass in tonnes, W = average truethickness of the reef (with a minimum of 1 m),A = area of the block, and r = specific gravity.

The tonnage for a particular area is thencalculated by summing the tonnage for eachblock.

14.7 EXPLORATION AND DEVELOPMENT INTHE SOUTHERN PART OF THE WEST RANDGOLDFIELD

As discussed in section 14.4, the easily discov-ered, near-surface deposits have been known,and often exploited, for many years. The objectof more recent exploration is to find the deeper,more stratigraphically, sedimentologically andstructurally complex deposits. In areas withactive mining, careful analysis and reinterpre-tation of geological data, including data sup-plied in company annual reports, can lead tonew discoveries being made in areas previouslythought to be barren. The area discussed inthis case history is the southern extension ofthe West Rand Goldfield mined immediatelydown dip from the outcrop of the Central RandGroup. The original mines are RandfonteinEstates, West Rand Consolidated, Luipardsvlei,and East Champ d’Or (Fig. 14.16). Four depositshave been discovered to the south of thesemines and to the east of a major structuralbreak known as the Witpoortjie Horst. Theyare the Cooke and Doornfontein sections ofRandfontein Estates, Western Areas, and SouthDeep. The exploration of Cooke Section willbe discussed in detail as it is one of the fewdetailed, published, exploration case historiesfor the Witwatersrand (Stewart 1981) and thatof South Deep as it shows the application ofmethods developed in the 1980s.

Development of Cooke Section only startedin the mid 1970s after almost a decade ofexploratory drilling, but nearly 90 years after

Page 363: Introduction to mineral exploration

346 C.J. MOON & M.K.G. WHATELEY

FIG. 14.16 Surface geology of the West Rand. (From Stewart 1981.)

0 5 km

Wi tp

oor t ji e

Horst

Krugersdorp

Roodepoort

Randfontein

Westonaria

Panv lak te Dyke

SDPA

Wes

tern

Are

as

K loof

Liba

non

Vent

ersp

ost

Coo

ke

S

ect i

onSouth

Roodepor t

DurbanRoodepor t

Deep

WestRandCons

Randfonte inEstates

Timeball Hill Formation

Silverton FormationDaspoort FormationStruben Kop FormationHekpoort Lava Formation

Malmani DolomiteKromdraai Member

DiamictiteKlipriviersberg Lava

Turffontein SubgroupJohannesburg SubgroupJeppestown SubgroupGovernment SubgroupHospital Hill Subgroup

GreenstoneGranite

Igneous Plug

Dykes

Major Faults

KarooSupergroup

TransvaalSupergroup

VentersdorpSupergroup

WitwatersrandSupergroup

SwazilandSupergroup

Central Rand Group

West Rand Group

Vryheid Formation

Luipaardsvlei

East Champ D'Or

Pan

vlak

teFa

ult

N

Page 364: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 347

the discovery of gold in the region. Explana-tions for this late discovery are that these gold-bearing rocks are concealed beneath youngersediments, the structure is more complex thanin other areas, the gold-bearing conglomeratesare highly variable, and these conglomeratesare in a part of the sequence that had not beforeyielded economic quantities of gold. This latterpoint, in conjunction with the fact that theconglomerates which normally produced goldin other parts of the basin were not here consid-ered to be economic, meant that other areas wereexplored in preference to the Cooke Section.

14.7.1 Cooke Section – previous exploration

The area had been geologically mapped byMellor (1917). This map formed the basis ofexploration until 1962 when Cousins extendedthis mapping southward. Mapping (Fig. 14.16)showed that Witwatersrand Supergroup rockscrop out with an approximate E–W strikefrom the East Rand through Johannesburg toKrugersdorp in the west (Fig. 14.2). In the westthe strike of the Witwatersrand rocks startsto swing to the southwest, but the outcrop iscovered by younger sediments of the TransvaalSupergroup and the Witwatersrand rocks areupthrown on to a horst, the Witpoortjie Horst,by major faulting (Fig. 14.17).

Attempts were made to trace extensions ofthe West Rand by drilling from as early as thebeginning of this century and resulted in the re-cognition of the West Wits Line (section 14.5.2).Further drilling was carried out in the area tothe south of Randfontein Estates between 1915and 1950 (Fig. 14.18). This work showed thatthe Witpoortjie Horst had a nearer N–S strikethan previously thought. This drilling alsoprovided more detailed information about thevariations in the generalized stratigraphicalcolumn (Fig. 14.19). For example, the MainReef zone was expected to be highly product-ive, but nothing significant was intersectedin the Cooke Section. Unfortunately, the Mid-dle Elsburg Reefs, which were subsequentlyshown to be the main gold-bearing reefs ofthis area, were only poorly developed in theseholes. One hole was drilled in what is nowknown to be a high grade part of Cooke Sectionbut did not penetrate the reef as it was cut outby a dyke.

Pan

vlak

teFa

ult

Hor

st

40°

45°10°

35°

10°

25°

45°

Wi tp

oor t ji e

Gap Hors t

N

0 5 km

Ventersdorp Supergroup

Central Rand Group

West Rand Group

FIG. 14.17 Sub-Transvaal geology of the West Rand.Note the absence of Central Rand Group sedimentson the Witpoortjie Horst (From Lednor 1986.)

Ten holes were drilled in the early 1950s(Fig. 14.18) in Cooke Section. Some wereintended to obtain information about thepotential of the Main Reef zone on the horstwhile two of the holes were drilled to test thepotential of the reefs on the eastern, down-thrown side of the Witpoortjie Horst. Holeswere also drilled farther south, and as theyimmediately intersected good gold values inthe Elsburg Reefs, on what is now WesternAreas Gold Mine, further drilling on the CookeSection was deferred until 1961.

Before the 1950s, borehole cores were onlyassayed for gold content. With the advent ofinterest in uranium in the 1950s, the possibil-ity of extracting the uranium in addition to thegold was considered. Sampling and assayingof conglomerates for uranium showed thatthe increase in uranium price had made theuranium an attractive proposition which could

Page 365: Introduction to mineral exploration

348 C.J. MOON & M.K.G. WHATELEY

FIG. 14.18 Drilling on the West Rand. (From Tucker & Viljoen 1986.)

Pan

vlak

te F

aul t

Pan

vlak

te F

aul t

Pan

vlak

te F

aul t

Pan

vlak

te F

aul t

GB2

(a) (b)

(c) (d)

N

0 5 km

Southern limitof WitwatersrandSupergroup outcrop

Phase 1 DrillingPre 1915

N

0 5 km

Phase 2 Drilling1915 – 1950

N

0 5 km

Phase 3 Drilling1950 – 1960

N

0 5 km

Phase 4 Drilling1960 – 1970

Page 366: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 349

ll

l

l

l

l ll

ll

ll

ll l

l

l

l

l

l

l

l

l

l

ll

l

ll

l

l

l

l

l

l

l

l

ll

l

l

l

l

lll

ll

ll

l

l l

l

l

ll

l

l

l

l

l

ll

l

ll

l

l

l

l

l

ll

l l

l l l

l

l

l

ll

l

l

ll l l

l

lll

l

l

l

l

l

l

l

l

l

l

l

l l

l

l l

l

ll

l

l

l

l

ll

l

l

l

l

l

l

l

l l

l

l

l

ll

l

ll

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

ll

l

l

l

l

l

l

l

l

l

l

l

l

l

l

ll

l

l

l

l

ll

l

l

l

l

l

l

l

ll

ll l

ll

ll

ll l

l

l

l

l

lll

ll

ll

l

ll

l

l

l

l

l

l

l

l

l

l

ll l

lll

l ll

l

l l

l

l l

l

l

ll

l l

l ll

lll

lll

ll

ll

l l l

ll l l

l l

l

lll

l

l

l

l

l

l

ll l

ll

l

ll

ll

l

l

ll

l

l

l

l

l

l l

l

ll

l

l

l

ll

ll l

l

l

l

l

l

ll

l

l l

l

l

l

l

l l

ll

l

l

l

l l

l

l

l l

ll

l l

ll

l

l

l

l

l

l

ll

l

l

ll

l

ll

l

ll

l

l

l l

l

ll

l

l l

l

l

l

l

ll

l

l

ll

l

ll

l

ll

ll

l

ll

ll

l

l

l

l

l

l

l

l

l l

l

ll

1500

1000

500

0

AC

B

DEFG

A

B

C

DE

A

B

A

B

E8E7

E6E5

E4E3 E2E1

E9

K11(LE3)

K10(LE2)

K9 (LE1)K8

K7

K6

K5

K4

K3

K2

K1

UE1

UE2

UE3

UE4

UE5

UE6

UE7

L VCRMASS

VCR Polymictic and oligomictic conglomerate

Massive conglomerate

Interbedded conglomerates and coarse-grained quartzites

Coarse-grained quartzite with conglomerate bands

Occasional conglomerate bands

Grey - green quartzitebecoming more siliceous andfiner grained towards base

Argillaceous quartzite with siliceous phases

Argillaceous quartzite with black and yellow specks

Argillaceous khaki quartzite

Siliceous quartzite with basal conglomerate

Siliceous coarse-grained quartzite withwell developed conglomerate at base

Argillaceous quartzites with numerousconglomerate bands

Conglomerates

Gray siliceous quartzites alternating with narrowargillaceous bands; occasional conglomerates

Argillaceous coarse-grained quartzitewith occasional siliceous bands

Coarse-grained quartzite with conglomerate bands

ConglomerateArgillaceous quartzite or shaleArgillaceous coarse-grained quartzite withbasal conglomerateSiliceous quartzite and basal conglomerateArgillaceous quartzites

Coarse-grained argillaceous quartzites

Basal conglomerateArgillaceous coarse-grained quartziteCoarse-grained argillaceous quartzitewith occasional conglomeratesBasal conglomerate

Medium-grained yellow quartzite

Upper transition of Kimberley shale

ModderfonteinMember

Waterpan Member

GemsbokfonteinMember

PanvlakteMember

GemspostMember

VlakfonteinMember

VENTERSPOSTFORMATION

ELSBURGFORMATION

WESTONARIAFORMATION

ROBINSONFORMATION

BOOYSENSFORMATION

WIT

WA

TE

RS

RA

ND

SU

PE

RG

RO

UP

CE

NT

RA

L R

AN

D G

RO

UP

TU

RF

FO

NT

EIN

SU

B–

GR

OU

PJH

BS

UB

–GP

Hei

ght

(m)

FIG. 14.19 Detailed stratigraphical column for the Central Rand Group in the Randfontein Estates area.(After Randfontein Estates Gold Mine personal communication.)

Page 367: Introduction to mineral exploration

350 C.J. MOON & M.K.G. WHATELEY

make a deposit with marginal gold values pay-able. The existing data were reviewed, and theresults of one of the holes (GB2) drilled in thedownthrown block showed gold and uraniumvalues which were considered worth followingup (Fig. 14.18).

Geophysics

Gravity and airborne and ground magnetic sur-veys were carried out across the Cooke Sectionto obtain data that would assist in the interpre-tation of the structure. Despite the success ofthese methods elsewhere in the WitwatersrandBasin (section 14.6), neither method proveddirectly successful in this area and drilling wasused to further evaluate the Cooke Section.The density difference between the Witwater-srand quartzites and the overlying Venters-dorp lavas was masked by the thick sequenceof younger Transvaal Supergroup dolomites,which made the gravity method unsuccessful.Magnetics were used to try to establish thepresence of the magnetic shale in the WestRand Group. Unfortunately the combinedthickness of the overlying Central Rand Groupand Transvaal Supergroup rocks was sufficientto effectively mask the magnetic signal fromthese shales.

Drilling

One hole drilled near the hole GB2 hadrevealed promising gold grades. This new holeintersected 1.32 m of Elsburg Reef at 968 mwhich assayed 20 g t−1 Au and 690 ppm U3O8.This success led to the planning of a full drill-ing program in which 88 holes were completedbetween 1960 and 1970 (Fig. 14.18). The holes,drilled on a grid, were spaced approximately800 m in a N–S (strike) direction and as close as300 m in an E–W (dip) direction.

At that time it was standard practice todrill these holes using narrow diameter (NXor BX) diamond bits. The unconsolidated over-burden was drilled using larger diameter, per-cussion rigs (also known as drilling a pilothole). Pilot holes are now drilled to significantdepths, with much time and cost saving (Scott-Russell et al. 1990). The pilot hole would havesteel casing inserted to prevent collapse of theunconsolidated material into the hole. Once

coring commenced in the hard rock, this con-tinued until the final depth was reached and acomplete section of core would be available forthe geologist to log. To save time and expense,four deflections were usually drilled from theoriginal hole to obtain additional reef intersec-tions. On average each hole took 4–5 monthsto drill.

All the core was taken to a central loca-tion where it was logged. After logging, mostof the Transvaal Supergroup dolomites and theVentersdorp Supergroup lavas were discarded,but the Witwatersrand Supergroup rocks werestored, in case they were needed at some futuredate for further study. The entire successionof Witwatersrand Supergroup rocks was thensampled, at 30-cm intervals where gold anduranium grades were expected, and at 60-cmintervals where no assay values were expected.Fire assaying was used to determine the goldcontent and radiometric methods were usedto determine the U3O8 content, with suitablecheck assays undertaken both within onelaboratory and between laboratories.

Interpretation of results

The unusual sequences intersected by thefirst few boreholes made it difficult to interpretthe stratigraphy and the structure. Work com-menced on compiling a detailed stratigraphicalcolumn relevant to this area. It was found thatthe initial logging had not been done in suffi-cient detail to compile a specific stratigraphicalcolumn. Cores were retrieved from the storeon their trays and two holes were laid out sideby side for re-examination. In this way detailedcomparisons could be made and this wasrepeated many times. Marker bands, mainlyfine-grained quartzites, were identified whichcould be used for correlation, and, after manymonths and many attempts, a stratigraphicalcolumn was produced which was tested againstnew holes. The difficulties can be gaugedfrom Fig. 14.20 which shows the correlationbetween three holes (note the numerousunconformities). The process was further com-plicated by the presence of faults and dykes.Eventually a reliable stratigraphical columnwas generated and each hole could then bedivided into reef zones. At this stage it waspossible to study the grade and geometry of

Page 368: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 351

200

180

160

140

120

100

80

60

40

20

0

Dep

th (

m)

PV 5 PV 16 PV 31

M

M

K 4/5

L

L K 2/3

K 11

L

ML

H

K 11

L

M

M

H

UE 2UE 1

UE

2UE 1

E9E / c

E9E8

E1

E8

WitwatersrandSupergroup

Ventersdorp Supergroup

Fine-grained quartzite

Dense quartzite

Conglomerate

Dyke

Shear possible fault

High grade

Moderate grade

Low grade

H

M

L

FIG. 14.20 Logs of three drillholes indicating thedifficulties of correlating between holes. For thelocation of holes see Fig. 14.21. (After Stewart 1981.)

0 1 km

PV29

PV8

PV23PV9

PV37

PV26

PV7

PV12PV22

PV15 PV10

PV40PV13

PV18

PV6

PV27

PV41

PV35

PV34

PV11PV39PV17PV33

PV28PV14

PV16

PV31PV5

22 m

30 m

38 m

45 m

PV38

22m

15m

Limit of E9E/c Conglomerate

PV36

N

FIG. 14.21 Typical isopach map used to interpretdrilling, in this case the E9c to UE1a unit.(After Stewart 1981.)

individual reef zones, using grade distributionand isopach maps, and to determine the struc-ture of the area, using structure contour maps.The sedimentological controls on the gold anduranium distribution mean that additionaldata can be gained by comparing a reef isopachmap with the overlying inter-reef interval(Fig. 14.21) to identify possible depositionaltrends, or by investigating pebble size, andconglomerate to quartzite or gold to uraniumratios.

The drilling showed that the cover of Trans-vaal Supergroup dolomites varied in thicknessfrom 100 m at the northern boundary of CookeSection to approximately 600 m in the south.Below the dolomites older rocks were faulted.The dominant feature is the N–S PanvlakteFault, which forms the eastern boundary of theWitpoortjie Horst. This horst was active duringthe later part of Witwatersrand sedimentation.Most of the Witwatersrand rocks had beeneroded from the horst during the pre-Transvaalpeneplanation. To the east of this fault, an

Page 369: Introduction to mineral exploration

352 C.J. MOON & M.K.G. WHATELEY

almost entire succession of Witwatersrandrocks, and some Ventersdorp Supergroup lavas,were preserved on Cooke Section (Fig. 14.22).A gentle, N–S trending anticline exists, whichcauses the sediments to dip from just a fewdegrees up to 45 degrees to the east.

Of all the recognized reefs on Cooke Sec-tion (Figs 14.20 & 14.22) only three compositeunits, or packages, are currently consideredeconomically important, the UE1a–E9Gd, theE8, and the K7–K9. The former is the mostimportant reef package that is currently beingmined with only minor stoping on the E8 andE9 reefs.

Development of Cooke Section

In early 1973, shaft sinking was completed onthe Cooke Section No. 1 Main Shaft at a finaldepth of 867 m. Towards the end of the yearstoping began in the area adjacent to the shaft.By the end of 1973 some 15,544 t of ore hadbeen mined and hoisted to the surface. This

material had an average grade of 17.7 g t−1 Au.Cooke Section No. 2 Main Shaft was commis-sioned in 1977, and in 1980 the Cooke SectionNo. 3 Shaft was started. It was finished, at afinal depth of 1373 m, in 1983.

Cooke Section has evolved into a large rela-tively low grade producer. By 1992 around 5 Mtof ore were being mined at an average gradeof 8 g t−1 of gold, although this had dropped to2.87 Mt at 4.9 g t−1 Au in 2002. The incomefrom uranium has become of very minor im-portance with the decline in the price of thatmetal. Trackless mining now accounts for 70%of production and the mine is one of the mostefficient on the Witwatersrand. Cooke Sectionhas been a major success due to the natureof the reefs and its inception at a time of thesharp increase in the price of gold. Tracklessmining has, however, not been successful onthe Doornkop Section, immediately north ofCooke Section, due to excessive dilution andthe mine was mothballed in 1999. It is howeverbeing reactivated in 2004.

A B

UE 3UE 3

A5UE1 A

E8E1K11

156155150

140128123148106104 Level

Kimberley Reef ZoneVentersdorpAge ‘A’ Dyke

K9

K7E1

K11

A5E8

A B

Cooke 1

Cooke 2

Cooke 3 0 500 m

Weathered / leached zone

Dolomite

Shales and quartzite

Ventersdorp Lava

UE5 Conglomerate Zone

E1K11

Lithological contact betweenthe Elsburg and Kimberley Reefs

Black Reef

Fault

Middle Elsburg Reef Zone

K7

K9

1500

1000

500

0

Hei

ght

abov

e se

a le

vel (

m)

W ECooke 3

N

FIG. 14.22 Section through Cooke Section Mine. (After Randfontein Estates Gold Mine personalcommunication 1991.)

Page 370: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 353

14.7.2 South Deep

This prospect is one of the major undevelopeddeposits in the Witwatersrand Basin with re-sources of more than 1000 t of contained gold.In situ reserves are 216 Mt at 8.4 g t−1 Au ina resource of 295 Mt of 10.1 g t−1 Au basedon 5 g t−1 Au cut-off (South Deep 2001). Thepotential of the area was first recognized dur-ing the drilling of Western Areas in the mid1960s. However the depth (>2500 m) and theexploration of other prospects in the areadelayed systematic exploration until the late1970s. It was quickly recognized that the de-posit was unusually thick and provided oppor-tunities for mechanized wide orebody miningas well as conventional stoping. The reef pack-ages of interest are the Massive and IndividualElsburg Reefs as well as the VentersdorpContact Reef (Figs 14.3 & 14.19). The thickerMassive and Individual Elsburg Reefs lie tothe east of a NNE trending feature known asthe upper Elsburg shoreline. The IndividualElsburg Reefs, that are up to nine in number,merge within 750 m of the shoreline to form acomposite reef which can be up to 70 m thick.To the west of the shoreline only the Venters-dorp Contact Reef is developed. The initialexploration was by conventional drilling fromthe surface (Fig. 14.23) and about 200 inter-sections were used in the feasibility study.

In order to make evaluation more accurate a3D seismic survey was shot between August1987 and September 1988. A nominal grid of25 × 50 m reflectors was used to generate a1:10,000 Ventersdorp Contact Reef (VCR)structure contour map. All faults over 250 m instrike length and 25 m throw are shown; thesefeatures can be extrapolated from the VCR tothe Elsburgs and form the basis of mine layoutsand ore reserve determinations that are moreaccurate than could be estimated from drilling(Campbell & Crotty 1990).

Wide orebody mining

In the South Deep Project Area, to the westof the shoreline, near the palaeoshoreline, thereefs are narrow and will be exploited by nar-row orebody stoping techniques, as describedabove. To the east of the shoreline, the reefsincrease in thickness from 2 m to over 100 m

(Fig. 14.24), and it is in these areas that themechanized, wide orebody mining systemswill be used. Room and pillar mining can beused to mine reefs up to 8 m thick, after mininga conventional stope to distress the rock butfor greater thicknesses, variations on longholestoping such as vertical crater mining will beemployed (Tregoning & Barton 1990, Maptek2004, South Deep 2004). Vertical crater miningcan be developed initially in the same way asroom and pillar mining. Two stopes will beestablished at the top and bottom of the pro-posed mining block (Fig. 14.25). Each reef blockis planned to be 150 m wide in a strike direc-tion and 250 m wide in a dip direction. Theupper excavation is called the drilling drive andthe lower, the cleaning drive. Retreat miningby means of slots up to 40 m high (betweenthe two levels), 50 m long, and up to 7 m widewill be practised. The stopes are to be clearedusing remote-controlled LHDs, and a vacuumor hydraulic cleaning system used for sweep-ing before backfilling. The cemented, crushedwaste rock, backfill will be placed in themined-out slots as a permanent support.

Developments 1990–2004

South Deep has an advantage over other newprospects in that the target reefs are a continua-tion of those mined on the south section ofWestern Areas and test mining of the area hasbeen possible without the cost of sinking a newshaft. A twin haulage way has been driven fromWestern Areas’ south shaft to the site of SouthDeep’s main shaft, a distance of about 2000 mat a cost of around R70M, and was completedin 1991 (Haslett 1994). Underground drillingfrom the haulage (Fig. 14.26) and developmenthas shown that grades obtained from surfacedrilling err on the low side but has confirmedthe structure deduced from the seismic survey.Mining is underway to extract the reef in thearea around the main shaft. Besides produc-ing revenue from around 24,000 t per month(2.2 t of gold in 2002), the development willdestress the shaft pillar. Development wasalso undertaken to enable underground testingof the wide orebody by drilling holes at 30-mintervals. At the time of writing (2004) bothconventional (Fig 14.27) and trackless miningwere underway outside the destressed area and

Page 371: Introduction to mineral exploration

354 C.J. MOON & M.K.G. WHATELEY

N

0 1 km

MD38

MD15

MD22MD16

MD24

DP15

MD46MD42

MD45

MD44KMF4

DP12

DP8

DP2

DP7

KMF4

KMF1

KMF2

K1

DP13

KDP1

DP14

DP11

DP9

DP10

DP3

Wrench Fault

Wrench Fault

Wat

erpa

n F

ault

Pan

vlak

te F

ault

Wes

t R

and

Faul

tW

est

Ran

d Fa

ult

South Deep Project Area

Western Areas G.M.

800 m

1000 m

1000 m

1200 m

1400 m

1600 m

800 m

1200 m

1000 m

800 m

NE

W

Drift and intersection point

Elevation below meansea level

Borehole collar

South Deep Shaft

Fault

Upp

erE

lsbu

rgS

hore

line

FIG. 14.23 Drill plan and simplified structure contours on VCR of the South Deep Prospect Area.(After South Deep 1990.)

Page 372: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 355

(a)

(b)

W E

SHORELINE

VCR narrow stoping

Narrow stoping to 2 – 5 m width

Vertical crater mining room and pillar

Room and pillar mining

Vertical crater mining

EC top destressing stoping

Narrow stoping, room and pillar

VCR

Upp

er E

lsbu

rg R

eefs

MB

MI

MAED

EC150

100

50

0

Dep

th (

m)

0 500 1000 1500 2000Distance (m)

Dep

th (

m)

200

100

0

FIG. 14.24 Generalized section showing the relation of proposed mining methods to reef thickness.(After Tregonning & Barton 1990.)

40 m

ED ZONE

Cemented fill

Final slot being mined450 m

150 m

150 m

150 m

Developmentof new block

VCM stoping(slot)

Strike drives (4 m × 4 m)

Distress cut EC top

FIG. 14.25 Perspective view of proposed vertical crater mining. (After Tregonning & Barton 1990.)

Page 373: Introduction to mineral exploration

356 C.J. MOON & M.K.G. WHATELEY

Boreholes completed

Proposed boreholes

Development

Proposed development

Mineral rights heldby Western Areas G.M.

1

23

45

6

78911

12 10

X Y

South Deep Project Area

Shore

line

95 T

win H

aula

ge

Western Areas G.M.

South DeepMain Shaft

Wat

erpa

n F

ault

Panv

lakt

e Fa

ult

Fault

X YShoreline

0 200 400 800 1000Distance (m)

VCRMB

MI MAED

EC

EB Quartziteand stringers

EA Quartzite

200

100

0

Dep

th (

m)

0 1 km

N

12 11

10 9 8 7 6 5 4 2

FIG. 14.26 Plan and section showing initial development at South Deep from Western Areas gold mine.Note the drilling to confirm grade and reef location. (After South Deep 1992.)

production was 1.37 Mt at 8.4 g t−1 Au (WesternAreas 2003). The transition from conventionalto trackless mining has taken considerable timeand has been accompanied by the retrench-ment in 1999 of 2560 staff.

Development of the deposit has had acomplex history since 1990. The initial devel-

opment was financed by floating a company(South Deep Exploration Limited) on theJohannesburg Stock Exchange. This companywas then merged with the Western Areascompany in 1995 and subsequently the hostmining house (JCI) was broken into constitu-ent parts by commodity. In 1999 50% of the

Page 374: Introduction to mineral exploration

14: WITWATERSRAND CONGLOMERATE GOLD 357

FIG. 14.27 Underground view ofworkings on the Massive Elsburgsat South Deep, 2000.

Geophysical methods have proved very suc-cessful in delineating the stratigraphical set-ting of the blind deposits. Magnetic and gravitymethods made a major contribution in the1930s to 1960s, as have reflection seismics inthe 1980s.

Definition of mineable areas depends onan understanding of the sedimentological andstructural controls on ore distribution as re-vealed by the very limited information avail-able from deep drilling. Evaluation of thislimited information led to the developmentof geostatistical methods that provide reliableestimates of grade and tonnage before a deci-sion was made to commit large expendituresto shaft sinking. The discovery of the SouthDeep deposit shows that large deposits remainto be discovered, although they are likely to bevery few in number and difficult to develop.

project was bought for $US235 million byPlacer Dome who then assumed managementcontrol. Shaft sinking began in 1995 and by2004 the twin shaft complex had reachedbottom (2994 m) and was being equipped. Theoverall cost of the project is estimated atR6.52 billion (~$US970M) and is anticipatedfor completion in 2006.

14.8 CONCLUSIONS

The Witwatersrand Basin with its largetabular orebodies is a unique gold occurrence.Exploration has evolved from the testing ofdown-dip extensions from outcrop through therecognition of blind districts, such as the WestWits Line, to the definition of deep targetsthat can support a large underground mine.

Page 375: Introduction to mineral exploration

358 A.M. EVANS & C.J. MOON

15A Volcanic-associated

Massive Sulfide Deposit –Kidd Creek, Ontario

Anthony M. Evans and Charles J. Moon

15.1 INTRODUCTION

Before reviewing the exploration history andgeology of Kidd Creek it is necessary to de-scribe and discuss the general features anddifferent types of massive sulfide deposits,as a good knowledge of these is essential indesigning an exploration program to find moredeposits of this type. The term volcanic-associated massive sulfide deposits is some-thing of a mouthful and in this chapter willoften be abbreviated VMS deposits.

15.2 VOLCANIC-ASSOCIATED MASSIVE SULFIDEDEPOSITS

15.2.1 Morphology

These deposits are generally stratiform, lenticu-lar to sheetlike bodies (see Fig. 3.8) developedat the interfaces between volcanic units orat volcanic–sedimentary units. Though oftenoriginally roughly circular or oval in plan,deformation may modify them to blade-like,rod-shaped or amoeba-like orebodies, whichmay even be wrapped up and stood on end, andthen be mistaken for replacement pipes such asthe Horne Orebody of Noranda, Quebec. Themassive ore is usually underlain by a stock-work that may itself be ore grade and whichappears to have been the feeder channel upwhich mineralizing fluids penetrated to formthe overlying massive sulfide deposit. It mustbe remembered that slumping may upset thesimple model shape mentioned above andindeed the whole of the massive ore can slide

off the stockwork and the two become com-pletely separated. The massive ore is oftencapped by an exhalative, sulfide-bearing, ferru-ginous chert which may have a significantlygreater areal extent than the orebody and hencebe of considerable exploration value.

15.2.2 Classification

These ores show a progression of types. Therehave been a number of attempts to classifythem based on the metals produced and thehost rocks. Perhaps the most comprehensiveis that of Barrie and Hannington (1999), whoclassify the deposits into five types based onhost rock composition. From most primitiveto most evolved these are: mafic, bimodal–mafic, mafic–siliciclastic, bimodal–felsic, andbimodal–siliciclastic. The sizes and grades ofthese deposits are summarized in Table 15.1.

To some extent corresponding with theabove environmental classification a geochem-ical division into iron (pyrite deposits used forsulfuric acid production), iron-copper (Cyprus-type), iron-copper-zinc (Besshi-type), and iron-copper-zinc-lead (Kuroko- and Primitive-type)is often employed (Table 15.2, Fig. 15.1). Usinga different approach from the simple chemicalone Lydon (1989) has shown that, if each pointon a ternary diagram of the grades of the depos-its is weighted in terms of tonnes of ore metalwithin the deposit, then clearly there are justtwo major groups, Cu-Zn and Zn-Pb-Cu (Fig.15.2). Indeed, as Lydon stresses, there are fewso-called copper deposits without some zinc.

Yes, reader, the situation is confusing! –we may indeed be dealing with a continuous

Page 376: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 359

TA

BLE

15.1

Tot

al a

nd

aver

age

grad

e an

d to

nn

age

for

VM

S ty

pes,

exc

ludi

ng

Ch

ina

and

ex-S

ovie

t B

lock

cou

ntr

ies.

Fig

ure

s ar

e fo

r co

mbi

ned

min

ed a

nd

min

eabl

e re

serv

es a

nd

reso

urc

es. (

From

Bar

rie

& H

ann

ingt

on 1

999.

)

Typ

en

Tot

al to

nnag

eT

otal

Cu

Tot

al P

bT

otal

Zn

Tot

al A

uT

otal

Ag

in b

illi

onin

mil

lion

in m

illi

onin

tonn

esin

mil

lion

in to

nnes

tonn

esto

nnes

tonn

esto

nnes

×××× × 1

02 ×××× ×

103

Mafi

c62

0.18

3.7

0.04

1.3

2.31

2.6

Bim

odal

-mafi

c28

41.

4524

.32.

044

.312

.91

38.2

Mafi

c-si

lici

clas

tic

113

1.24

16.2

0.6

9.7

4.03

9.2

Bim

odal

-fel

sic

255

1.29

7.1

13.2

54.2

14.1

812

0.0

Bim

odal

-sil

icic

last

ic97

2.50

21.5

24.0

55.1

4.11

60.0

Tot

al81

16.

66

Ave

rage

siz

eA

vera

ge C

uA

vera

ge P

bA

vera

ge Z

nA

vera

ge A

uA

vera

ge A

gin

mil

lion

grad

e in

wt

grad

e in

wt%

grad

e in

wt%

grad

e in

g t−−−− −1

grad

e in

g t

−−−− −1

tonn

es

Mafi

c2.

82.

040.

101.

822.

5620

.0B

imod

al-m

afic

5.1

1.88

0.75

4.22

1.52

36.5

Mafi

c-si

lici

clas

tic

11.0

1.74

1.83

2.43

0.84

19.8

Bim

odal

-fel

sic

5.2

1.44

1.64

5.63

2.06

92.8

Bim

odal

-sil

icic

last

ic23

.71.

101.

844.

161.

1384

.4

Num

ber

Num

ber

Num

ber

Num

ber

Num

ber

of d

epos

its

of d

epos

its

of d

epos

its

of d

epos

its

of d

epos

its

>>>> >100

Mt

50–1

00M

t20

–50

Mt

10–2

0M

t5

–10

Mt

Mafi

c0

03

17

Bim

odal

-mafi

c1

69

1620

Mafi

c-si

lici

clas

tic

31

107

10B

imod

al-f

elsi

c0

312

1929

Bim

odal

-sil

icic

last

ic9

45

611

(b)

(c)

(a)

Page 377: Introduction to mineral exploration

360 A.M. EVANS & C.J. MOON

TA

BLE

15.2

Vol

can

ic-a

ssoc

iate

d m

assi

ve s

ulfi

de d

epos

it t

ypes

– “

con

ven

tion

al”

term

inol

ogy.

(Mod

ified

from

Hu

tch

inso

n 1

980.

)

Typ

e

Bes

shi

(= K

iesl

ager

)C

u–Z

n±A

u±A

g

Cyp

rus

Cu

(+Z

n)±

Au

Ku

rok

oC

u–Z

n–

Pb±

Au

±Ag

Pri

mit

ive

Cu

–Zn

±Au

±Ag

Cla

stic

sed

imen

tary

rock

s

Con

tin

ent-

deri

ved

grey

wac

kes

an

dot

her

tu

rbid

ites

Min

or o

r ab

sen

t

Shal

low

to

med

ium

dept

h c

last

ics,

few

carb

onat

es

Imm

atu

regr

eyw

ack

es, s

hal

es,

mu

dsto

nes

Plat

e te

cton

icse

ttin

g

Epi

con

tin

enta

l or

back

-arc

Oce

anic

rif

tin

g at

accr

etin

g m

argi

n

Bac

k-a

rc r

ifti

ng

Mu

ch d

ebat

ed:

fau

lt-b

oun

ded

trou

gh, b

ack

-arc

basi

n?

Kno

wn

age

rang

e

Ear

ly P

rote

rozo

icP

alae

ozoi

c

Ph

aner

ozoi

c

Ear

ly P

rote

rozo

icP

han

eroz

oic

Arc

hae

an–e

arly

Pro

tero

zoic

Vol

cani

c ro

cks

Wit

hin

pla

te(i

ntr

apla

te)

basa

lts

Oph

ioli

tic

suit

es,

thol

eiit

ic b

asal

ts

Bim

odal

su

ites

,th

olei

itic

bas

alts

,ca

lc-a

lkal

ine

lava

san

d py

rocl

asti

cs

Full

y di

ffer

enti

ated

suit

es, b

asal

tic

torh

yoli

tic

lava

s an

dpy

rocl

asti

cs

Dep

osit

iona

len

viro

nmen

t

Dee

p m

arin

ese

dim

enta

tion

wit

hba

salt

ic v

olca

nis

m

Dee

p m

arin

ew

ith

th

olei

itic

volc

anis

m

Exp

losi

ve v

olca

nis

m,

shal

low

mar

ine

to c

onti

nen

tal

sedi

men

tati

on

Mar

ine

<1k

m d

epth

.M

ain

ly d

evel

oped

ingr

een

ston

e be

lts

Gen

eral

cond

itio

ns

Rif

tin

g

Ten

sion

al, m

inor

subs

iden

ce

Rif

tin

g an

dre

gion

alsu

bsid

ence

,ca

lder

afo

rmat

ion

Maj

or s

ubs

iden

ce

Page 378: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 361

FIG. 15.1 Stratigraphicaldistribution of volcanic-associatedmassive sulfide deposits in theNoranda area, Quebec. The nameson the diagram are those of variousvolcanic formations each of whichmay include both flows andpyroclastic deposits. The relativesizes of the orebodies are indicated.The largest, at the Horne Mine (H),is over 60 Mt. (After Lydon 1989.)

Acid tuff

Ferruginous chert zone

Baryte deposit

Stockwork

Kuroko or black ore,galena-sphalerite-baryte

Gypsum

Clay

Explosion breccia

Rhyolite dome

Acid tuff breccia

Oko or yellow ore,pyrite-chalcopyrite

spectrum that we have not yet fully recognized,as Lydon’s work suggests (Fig. 15.1), but mean-while mining geologists will continue to usethe terms listed in Table 15.2 because in thepresent state of our knowledge, they implyuseful summary descriptions of the deposits,helpful in formulating exploration models.

It is important to note that precious metalsare also produced from some of these deposits;indeed in some Canadian examples of thePrimitive-type they are the prime product.

15.2.3 Size, grade, mineralogy, and textures

Data from the five types defined by Barrie andHannington (1999) are given in Table 15.1. Themajority of world deposits are small and about80% of all known deposits fall in the size range0.1–10 Mt. Of these about a half contain lessthan 1 Mt (Sangster 1980). Average figures ofthis type tend to hide the fact that this is adeposit type that can be very big or rich or bothand then very profitable to exploit. Examplesof such deposits are given in Table 15.3. Largepolymetallic orebodies of this type form thehigh quality deposits much sought after bymany small to medium sized mining compan-

Cu

Cu–Zngroup

Zn–Pb–Cugroup

ZnPb

Thousands of tonnesper 1% area of diagram

>3000

500–2999

1–499

FIG. 15.2 Schematic section through a Kurokodeposit. (Modified from Sato 1977.)

Page 379: Introduction to mineral exploration

362 A.M. EVANS & C.J. MOON

TABLE 15.3 Tonnages and grades for some large or high grade VMS deposits.

Mined ore +++++ reserves

(Mt) Cu% Zn% Pb% Sn% Cd% Ag g t−−−−−1 Au g t−−−−−1

Kidd Creek, Ontario 155.4 2.46 6.0 0.2 r r 63 –Horne, Quebec 61.3 2.18 – – – – – 4.6Rosebery, Tasmania 19.0 0.8 15.7 4.9 – – 132 3.0Hercules, Tasmania 2.3 0.4 17.8 5.7 – – 179 2.9Rio Tinto, Spain* 500 1.6 2.0 1.0 – – r rAznalcollar, Spain 45 0.44 3.33 1.77 – – 67 1.0Neves–Corvo, Portugal† 30.3 7.81 1.33 – – – – –

2.8 13.42 1.35 – 2.57 – – –32.6 0.46 5.72 1.13 – – – –

* Rio Tinto, originally a single stratiform sheet, was folded into an anticline whose crest cropped out and vast volumes weregossanized. The base metal values are for the 12 Mt San Antonio section.† The three lines of data represent three ore types and not particular orebodies of which there are four. Ag is present in somesections. r, indicates this metal is or has been recovered, but the average grade is not available.

ies (see “Metal and mineral prices” in sectionin 1.2.3). The lateral continuity and shape ofVMS deposits usually makes them attractivemining propositions either in open pit or asunderground operations using trackless equip-ment. Mechanization is particularly import-ant in countries with high labor costs, such asCanada, Sweden, or Australia.

The mineralogy of these deposits is fairlysimple and often consists of over 90% ironsulfide, usually as pyrite, although pyrrhotiteis well developed in some. Chalcopyrite,sphalerite, and galena may be major con-stituents, depending on the deposit class,bornite and chalcocite are occasionally import-ant, arsenopyrite, magnetite, and tetrahedrite–tennantite may be present in minor amounts.With increasing magnetite content these oresgrade to massive oxide ores (Solomon 1976).The gangue is principally quartz, but occa-sionally carbonate is developed and chloriteand sericite may be locally important. Theirmineralogy results in these deposits havinga high density and some, e.g. Aljustrel andNeves-Corvo in Portugal, give marked gravityanomalies, a point of great exploration signi-ficance (see section 7.4).

The vast majority of massive sulfide depositsare zoned. Galena and sphalerite are moreabundant in the upper half of the orebodieswhereas chalcopyrite increases towards thefootwall and grades downward into chalcopy-rite stockwork ore (see Figs 3.8 & 15.7). This

zoning pattern is only well developed in thepolymetallic deposits.

15.2.4 Wall rock alteration

Wall rock alteration is usually confined to thefootwall rocks. Chloritization and sericitizationare the two commonest forms. The alterationzone is pipe-shaped and contains within it andtowards the center the chalcopyrite-bearingstockwork. The diameter of the alteration pipeincreases upward until it is often coincidentwith that of the massive ore. Metamorphoseddeposits commonly show alteration effects inthe hanging wall. This may be due to the intro-duction of sulfur released by the breakdownof pyrite in the orebody. Visible chloritic orsericitic alteration is almost entirely confinedto the zone around the stockwork and is onlyexposed at the surface in cases of near-surface,steeply dipping deposits. The alteration halocan be used as a proximity indicator in drillholes that miss the orebody as can the chem-ical content of the halo (see below).

In several mining districts laterally wide-spread zones of alteration have been reported asoccurring stratigraphically below the favorablehorizon. Such zones extend for up to 8 km andvertically over several hundred meters. In asuccessful attempt to expand the explorationtarget by mapping a larger part of the causativehydrothermal system, Cathles (1993) carriedout a regional whole rock oxygen isotope sur-

Page 380: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 363

0 500 m

N

D6

D3

D5D4

D1

D8D9

D2

D7

Sulfide deposit

Gypsum deposit

Volcanic breccia

Stockwork ordissemination depositStructure contours on the lavadomes, mapped and assumed

FIG. 15.3 Data of representative VMS districtsplotted on a Cu-Pb-Zn ternary diagram weighted intonnes of ore metal contained in the deposit andthen contoured. (Modified from Lydon 1989.)

vey to reveal the extent of the oxygen isotopealteration in the Noranda region.

15.2.5 Some important field characteristics

Association with volcanic domes

This frequent association is stressed in the lit-erature and the Kuroko deposits of the Kosakadistrict, Japan, are a good example (Fig. 15.4).Many examples are cited from elsewhere, e.g.the Noranda area of Quebec, and some authorsinfer a genetic connexion. Certain Japaneseworkers, however, consider that the Kurokodeposits all formed in depressions and that thedomes are late and have uplifted many of themassive sulfide deposits. In other areas, e.g.the Ambler District, North Alaska, no closeassociation with rhyolite domes has beenfound (Hitzman et al. 1986).

Cluster development

Although the Japanese Kuroko deposits occurover a strike length of 800 km with more than100 known occurrences, these are clusteredinto eight or nine districts. Between these dis-tricts lithologically similar rocks contain onlya few isolated deposits and this tends to be thecase, with a few notable exceptions, for mas-sive sulfide occurrences of all ages. Sangster(1980) calculated the average area of a cluster tobe 850 km2 with 12 deposits containing in total94 Mt of ore. The Noranda area (Fig. 15.1) isabout 170 km2 and contains about 20 depositsaggregating about 110 Mt. The four Neves-Corvo deposits lie in an area of 6 km2 and have250 Mt of possible ore; more mineralisationoccurs around Algare just 2.5 km to the north(Leca 1990).

In some districts VMS deposits and asso-ciated domes appear to lie along linear fractureswhich may have controlled their distribu-tion. Synvolcanic basinal areas are importantorebody locations in the Iberian Pyrite Belt.

Favorable horizons

The deposits of each cluster often occur withina limited stratigraphical interval. For Primit-ive and Kuroko types, this is usually at the topof the felsic stage of cyclical, bimodal, calc-alkaline volcanism related to high level magma

chambers rather than to a particular felsicrock type (Leat et al. 1986, Rickard 1987). Forexample, the deposits of the Noranda area,Quebec (Fig. 15.4), lie in a narrow stratigraph-ical interval within a vast volcanic edifice atleast 6000 m thick. A very good account of thestratigraphical relationships is given in Kerrand Gibson (1993).

Page 381: Introduction to mineral exploration

364 A.M. EVANS & C.J. MOON

FIG. 15.4 Distribution of dacitelava domes and Kuroko deposits,Kosaka District, Japan. (Modifiedfrom Horikoshi & Sato 1970.)

Waite Rhyolite

Amulet Rhyolite

Rusty Ridge Andesite

Northwest Rhyolite

Flavrian Andesite

Waite Andesite

Flavrian Lake Granite

Amulet Andesite H

Joliet Rhyolite

1 km

Brownlee Rhyolite

summary in Robb (2004) as well as the reviewvolume of Barrie and Hannington (1999).

15.2.7 Volcanic facies

The association of massive sulfide with par-ticular parts of the volcanic assemblage haslong been recognized. Brecciated rhyoliticvolcanics are known as “mill rock” in somemining camps because of their proximity toprocessing plants. The breccias are attributedto one of three processes: (i) they formed closeto structural magmatic vents which alsovented hot springs; (ii) they were the result ofhydrothermal explosive events; and (iii) thebreccias formed as a result of caldera forma-tion with which the ore-forming process wasassociated (Hodgson 1989). Whichever of theseprocesses is correct it is clear that an under-standing of the development of the volcanicpile is important. This knowledge can be ob-tained by detailed mapping, core logging, andcomparization with modern volcanic models.The most useful summary of these models isprovided in the book by Cas and Wright (1987),and the reader is referred to this text as well asthe paper of Gibson et al. (1999).

15.2.8 Exploration geochemistry

Exploration geochemistry has been widelyand successfully used in the search for VMSdeposits. In Canada the prospective areas have

Structural controls

Many massive sulfide deposits occur withinrecognized cauldron subsidence areas localizedalong synvolcanic faults and extracauldron de-posits may lie within graben flanked by vents(Kerr & Gibson 1993). Detailed mapping andexpertise in the interpretation of volcanic phe-nomena are obviously essential in the searchfor deposits in this terrane.

Exploration gelogists working in a favorableterrane should study all the literature on itsdeposits to refine their exploration model. Thiswill help them to decide on finer points suchas whether any newly found deposit is morelikely to be flat lying or steeply dipping, little orbadly deformed, little or much metamorphosedand to adjust their geochemical, geophysical,and drilling programs accordingly.

15.2.6 Genesis

For many decades VMS deposits were con-sidered to be epigenetic hydrothermal replace-ment orebodies (Bateman 1950). In the 1950s,however, they were recognized as being syn-genetic, submarine-exhalative, sedimentaryorebodies, and deposits of this type have beenobserved in the process of formation fromhydrothermal vents (black smokers) at a largenumber of places along sea floor spreadingcenters (Rona 1988).

There is a voluminous literature on thesubject and the reader is advised to read the

Page 382: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 365

been glaciated and have poor surface response,but rock geochemistry is widely used to deline-ate drill targets at depth. In the southernhemisphere surface geochemistry has been suc-cessful even in areas of intense weatheringwhere EM surveys have been ineffective due toconductive overburden.

Rock geochemistry

The rock geochemical signature of VMSdeposits has been investigated by a number ofworkers but is well summarized by Govett(1989) and Franklin (1997). In a review of alarge number of case histories he documentsthe very consistent response, regardless of ageand location. Aureoles are usually extensive,more than 1 km laterally and 500 m vertic-ally with footwall anomalies more extensivethan those in the hanging wall, especially inproximal deposits. In Archaean deposits sig-nificant hanging wall anomalies appear to beabsent. As discussed above Fe and Mg areenriched and Na and Ca depleted in nearly alldeposits (Fig. 15.5). Copper tends to be enrichedin the footwall and is depleted in the hangingwall; Mn shows the reverse behavior. BothK and Mn may be relatively depleted near thedeposit and enriched further away. The ore ele-ments, such as Zn, are commonly enriched.Copper may be depleted especially in the im-mediate wall rocks. The behavior of other traceelements, in particular Au, Cl, Br, As, Sb, maybe distinctive but data from case historiesare incomplete. Some authors have regressedtrace elements against SiO2 to compensate forfractionation. This is not always effective assome deposits, such as those in the Lac Dufaultarea, Noranda, have undergone addition of SiO2

during alteration.Cherty exhalites have been widely used as

indicators of hydrothermal activity but theirgeochemical signatures are varied. Kalogero-poulos and Scott (1983) have used a majorelement ratio, R = (K2O + MgO) × 100/K2O +Na2O + CaO + Mg to indicate proximity toKuroko mineralisation in Japan. Trace andmajor elements, in particular zinc and man-ganese but also volatiles such as arsenic, areenriched sporadically in these horizons butapplication has had varying degrees of success.A key requirement of any successful applica-tion is that the stratigraphical position of the

samples must be well understood. It is of littleuse comparing ratios of samples from differentexhalative horizons.

Geochemistry has also been used on aregional scale to attempt to discriminate min-eralized sequences of volcanics. However thevariation associated with mineralisation isusually less than that caused by fractionationand trace element data must be corrected forfractionation. Studies of immobile elements(Ti, Zr, REE), that have not been markedlyaffected by hydrothermal alteration, may indi-cate the original chemistry of the volcanics.Campbell et al. (1984) have suggested thatmineralized volcanics have flatter rare earthelement (REE) patterns than barren volcanicswhich tend to be enriched in light REE andhave negative Eu anomalies. Few regionalrock geochemical studies have been published,but Govett (1989) suggests that many VMSdeposits are associated with regional copper de-pletions. Both Russian and Canadian workers(Allen & Nichol 1984) suggest that analysis ofsulfides, separated either by gravity or select-ive chemical extraction, within volcanics, maygive regional anomalies of up to 10 km alongstrike from VMS deposits.

Surficial geochemistry

The extreme chemical contrast of the sulfideswith volcanics gives rise to high contrast tran-sition metal anomalies in soil and streamsediments. A large number of case histories areavailable.

15.2.9 Geophysical signatures

One of the key elements in the explorationmodel is the geophysical signature of thedeposit. The majority of VMS discoveries haveresulted from geophysics, notably in theglaciated areas of the Canadian Shield. VMSdeposits respond to a number of geophysicaltechniques, as they contain much sulfide thatoften shows electrical connection, permittingthe use of EM techniques, and the bodies aredense, allowing the use of gravity methods.

Mapping methods

Airborne magnetic and electromagnetic (EM)systems are widely used for mapping, both of

Page 383: Introduction to mineral exploration

366 A.M. EVANS & C.J. MOON

Mg O1.0

2.03.0

1.0

2.0

3.0

3.0

1.0

2.0

K2O

2.0

3.0

3.0

4.0

2.0

Ca O0.6

0.5

0.4

0.3

0.20.1

0.4

0.30.2

0.3

Na2O

3.0 4.0

2.0

1.0 4.0

3.0

2.0

3.0

2.01.0

0.1

Drillholes

Massive sulfides

0.1

0 500 m

FIG. 15.5 Major element anomalies in the footwall of the Fukasawa Mine, Japan. (From Govett 1983, data fromIshikawa et al. 1976.)

Page 384: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 367

general geology and of favorable lithologies.The high resolution of modern aeromagneticsystems resulted in the obsolescence of groundmagnetics for all but the smallest projects. Inaddition it is possible to merge aeromagneticdata with remotely sensed information. Anexcellent example of the use of magnetics inthe detailed mapping of Archaean terranes isprovided by Isles et al. (1989) for the Yilgarnblock in Western Australia. They highlight theuse of multi-client surveys as a cost-effectivemethod of mapping major lithologies andstructural changes. Magnetic units such asbanded iron formations are apparent as mag-netic highs in contrast to the magneticallyless disturbed areas of acid volcanics and sedi-ments. Sometimes secondary maghemite canform in these deeply weathered environments,such as parts of the Yilgarn, and confuse themagnetic signature, although this can be help-ful in detecting paleochannels as shown in Fig.15.6. In Canada both the provincial and federalgovernments have flown large parts of the pro-spective areas and the data is available at lowcost. Much of sub-Saharan Africa has also beenflown although availability of data is varied(Reeves 1989).

Direct detection of sulfides

Electrical methods have been very successful,particularly in Canada. Up to the 1970s this in-volved the use of airborne EM systems. Theirlimited depth penetration (originally <150 m)and the saturation coverage of prospectiveareas has resulted in the delineation and drill-

ing of virtually all near-surface conductors.More recent geophysical exploration is aimedat finding deposits at depth and has encourageda switch to down hole methods.

Pemberton (1989) has provided a summaryof EM responses from a number of CanadianVMS deposits. Figure 15.7 shows the detectionof a VMS deposit in the Noranda camp usingdownhole pulse EM from a hole drilled some100 m from the deposit at 700 m depth.

Airborne geophysics has been ineffectivein intensely and deeply weathered areas, suchas most of the Yilgarn Shield of WesternAustralia. The conductive nature of the over-burden is responsible for the generation ofnear-surface responses which mask signaturesfrom the VMS. Options that have been invest-igated include increasing transmitter power,shortening integration times, and increasingthe number of time channels used (Smith andPridmore 1989).

A more general problem is the excellent EMresponse of graphitic sediments. It is extremelydifficult, if not impossible, to differentiategraphite from massive sulfides and, as graphiticsediments are often found associated withbreaks in the volcanic sequence, this has re-sulted in many thousands of meters of wasteddrilling. However, some VMS deposits, suchas the Trout Lake deposit in Manitoba, arecompletely enclosed in graphite and would bemissed if graphitic sediments were excludedfrom consideration.

15.2.10 Integration of exploration techniques

Perhaps the most important part is the incor-poration of the factors mentioned above intoan integrated exploration program. A recentexample of the successful use of integration isthe discovery of the Louvicourt deposit nearVal d’Or, Quebec (Fig. 15.8). The program wasaimed at discovering large deep extensionsto the near-surface Louvem Mine which hadproduced for about 10 years. Drilling of sur-face holes to 700 m, costing $C50,000 each,was undertaken on 300 m centers, to testprospective horizons at depth (Northern Miner1990). The spacing of 300 m was chosen as onlylarge deposits were thought capable of support-ing a mine at 600 m depth. Hole 30 intersecteda stringer zone containing 0.86% Cu over

100 m

Palaeochannel

Conductive shear zone

Alluvial coverMaghemite

WEATHERED ZONE

Preferentialweathering

Sulfideor

Kimberlite

UNWEATHEREDBASEMENT

FIG. 15.6 Schematic diagram showing the problemsof using magnetics and EM in deeply weatheredareas, such as the Yilgarn in Western Australia.(After Smith & Pridmore 1989.)

Page 385: Introduction to mineral exploration

368 A.M. EVANS & C.J. MOON

STRINGERZONE

5 –20% sulfides

FLAVIAN GRANITE

AMULET RHYOLITE

AMULET ANDESITE

R 79 -2

1300

1200

1100

1000

900

800

700

600

500

400

300

200

100

0D

epth

(m

)

0 500 m

FIG. 15.7 Downhole pulse EM survey at the Ribago deposit, Noranda, Quebec. The survey clearly detects thesulfides 130 m from the hole. (After Pemberton 1989.)

intersected further stringer zone mineralisa-tion, leading to massive sulfides grading 11.8% Cu over 36 m in hole 42 and 2.8% Cu and5.7% Zn over 43 m in hole 44. Resources areestimated at 28 Mt of 4.3% Cu, 2.1% Zn,27.8 g t−1 Ag and 1.03 g t Au. The total cost ofthe discovery is put at $C2.5M.

118 m but more importantly downhole EM sur-veys showed a significant off-hole conductorand rock geochemistry showed a major ele-ment anomaly. This suggested a major zone ofhydrothermal alteration associated with a dif-ferent horizon (the Aur horizon) than that ofthe Louvem Mine. A further phase of drilling

Page 386: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 369

N S

HA

NG

ING

WA

LL

A

ND

ES

ITE

S

PYR

OC

LAST

IC

SEQ

UEN

CE

WH

ITE

FR

AG

ME

NT

BR

EC

CIA

TUFF

SE

QU

EN

CE

CA

PR

OC

K F

LOW

S

ALTE

RAT

ION

ZO

NE

AUR

ZO

NE

57

42B

6049

46

47

?

53B

53

54A

71

66

64A

6.98 0.961.37 0.07

39.82and

0.59 7.890.55 0.01

3.20

4.34 0.090.50 0.01

11.65

0.19 4.650.42 0.014

18.26

%Cu %Znoz.Ag oz.Au

Meters

3.81 6.492.78 0.08

22.35

2.26 0.070.14 0.015

65.55incl.

12.52 0.070.87 0.013

8.05

900

Dep

th (

m)

0 150 m

750

600

450

150

300

100°N 90°N 80°N

FIG. 15.8 Section of the Louvicourt deposit, Val d’Or . Note the blind nature of the deposit. (After NorthernMiner Magazine 1990.)

unremediated, give rise to acid mine drainage.Many deposits also contain significant con-centrations of toxic metals such as arsenic,cadmium, and mercury. The containment oftailings from flotation separation of sulfides isalso expensive. These environmental problemsand their associated costs have downgraded

15.2.11 Environmental impact

The highly sulfidic nature of the depositsmeans that they are likely to give rise to envir-onmental problems (Taylor et al. 1995). In par-ticular the iron and base metal sulfides areunstable when brought to the surface and, if

Page 387: Introduction to mineral exploration

370 A.M. EVANS & C.J. MOON

the appeal of VMS deposits for some miningcompanies.

15.3 THE KIDD CREEK MINE

15.3.1 Regional setting

Kidd Creek lies within the Abitibi Belt ofthe Superior Province of the Canadian Shield(Figs 15.9 & 15.10). With an exposed area of1.6 million km2 the Superior Province is thelargest Archaean Province in this shield. Therocks are similar to those found at Phanerozoicsubduction zones with differences due tohigher mean mantle temperature and possiblygreater crustal (slab?) melting at mantle depth(Hoffmann 1989). The province is composed ofsubparallel ENE trending belts of four contrast-ing types: (i) volcanic–plutonic terranes, whichresemble island arcs; (ii) metasedimentarybelts which resemble accretionary prisms(Percival & Williams 1989); (iii) plutonic com-plexes; and (iv) high grade gneiss complexes.In the southern part of the Superior Provincelinear, ENE trending, greenstone–granite beltsalternate, on a 100–200 km scale, with meta-sedimentary dominated terranes.

0 400 km

A

Grenvi

l le O

rogen

New QuebecOrogen

Hudson Bay

SUPERIOR

PROVINCE

FIG. 15.9 Location of the Superior Province of theCanadian Shield. A, position of the Abitibi Belt.

ONTARIO QUEBEC

Timmins

KIDD CREEKMINE

Noranda

Grenville O

rogen

6

5

4

7

8

3

9

10

11

2 150°

48°

80° 75°

0 100 km

FIG. 15.10 Distribution of volcanic complexes in the Abitibi Belt. Approximate outlines of the complexesincluding cogenetic intrusions and sediments are shown by the heavy pecked lines. The complexes withsignificant developments of VMS deposits are 2, 4, 6, 8, and 9. (Modified from Sangster & Scott 1976.)

Archaean greenstone belts the world overappear to be divisible into a lower dominantlyvolcanic section and an upper sedimentary sec-tion. The lower section may be divided into alower part, of primarily ultramafic volcanics,and an upper volcanic part in which cal-calkaline or tholeiitic, mafic to felsic rocks pre-dominate. The ultramafic (komatiitic) part also

Page 388: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 371

contains some mafic and felsic rocks and isthus bimodal. The tholeiitic or calc-alkalinepart is dominated by basaltic, andesitic, andrhyolitic volcanics. With increase in stratigraph-ical height in greenstone belts there is, interalia, (i) a decrease in the amount of komatiiteand (ii) an increase in the relative amount ofandesitic and felsic volcanics. The associatedsediments are mainly exhalative chert and ironformation with some thin slates. The uppersedimentary section is dominantly clastic.

There are regional differences in the volcaniccharacter of greenstone belts that appear to berelated at least in part to their age and ero-sion level. Thus the oldest belts of southernAfrica and Australia are typically bimodal andenriched in ultramafics and mafics (whichmake them prime areas for nickel explora-tion), whereas the younger Canadian beltshave a higher proportion of calc-alkaline andtholeiitic rocks and have so far been the bestground for finding VMS deposits. Large areas ofthe Superior Province formed at 2.7–2.6 Ga.

The Abitibi Belt is 800 km long and 200 kmwide making it the world’s largest, single, con-tinuous Archaean greenstone belt. It containsa relatively high proportion of supracrustalrocks of low metamorphic grade resulting froman unusually shallow erosion level. This belthas been divided into 11 volcanic complexes(Sangster & Scott 1976) (Fig. 15.10) which arecommonly many kilometers thick and con-tain one or more mafic to felsic cycles. VMSores occur in just five of these complexes andare exclusively associated with calc-alkalinevolcanic rocks, especially the dacitic andrhyolitic members which form 5.5–10% of thetotal volume of volcanic rocks in this belt andwhich belong to three different affinities whoseimportance have been stressed by Barrie et al.(1993). Production and reserves of VMS oresat 1990 totalled 424 Mt averaging 4.4% Zn,2.1% Cu, 0.1% Pb, 46 g t−1 Ag and >>1.3 g t−1 Au(Spooner & Barrie 1993).

15.3.2 Exploration history

The discovery of Kidd Creek is extremelywell documented and was a triumph for air-borne geophysics and geological understanding(Bleeker & Hester 1999). It is worth quoting theeloquent description of the discovery from aseries of papers (The Ecstall Story in the CIMM

Bulletin May 1974 (Matulich et al. 1974))describing the discovery and establishment ofthe Kidd Creek or Ecstall Mine, as it was origin-ally known.

“The Ecstall orebody does not owe its dis-covery to a fantastic stroke of luck but to theforesight, dedication, tenacity and hard work ofa Texasgulf [then Texas Gulf Sulfur] explora-tion team. The orebody was found as a resultof years of patient reconnaissance with thelatest exploration equipment and using themost advanced technology. It was located only15 miles north of a gold mining center that hadbeen operating seventeen major mines over aperiod of fifty years. For most of those years,prospectors had combed every square foot ofthe area surrounding Timmins looking forgold. In fact, a prospector had his cabin on therock outcrop that is now occupied by the minesurface crushing plant, just 100 yards fromthe orebody’s southern limit. As most of thegold mines had been found by surface outcropsand simple geological interpretation, the pro-spector’s search for the rich bonanza he feltwas nearby was limited to these methods. Thelayers of muskeg and clay over the Kidd Creekorebody protected it from these primitiveefforts.” (Clarke 1974)

The geological thinking behind the explora-tion program stemmed from the successfulapplication of a syngenetic model. At the timeof the development of the model (1954) mostgeologists believed that VMS deposits werereplacement deposits and related to granites,although papers from both Germany andAustralia had proposed an exhalative originfor these deposits (for discussion see Hodgson1989, Stanton 1991). Walter Holyk recognizedthe close association of deposits with the con-tact between sericite-schists (rhyolites) andintervolcanic sediment and applied the modelsuccessfully in conjunction with airborne EMsurveys for Texas Gulf Sulfur in the Bathurstarea of New Brunswick (Hanula 1982).

Holyk then persuaded his management toapply the same techniques and model to theCanadian Shield. He selected the Timminsarea based on the known occurrence of theKam Kotia deposit at the contact betweenrhyolite and sediments and the lack of knownEM coverage. The first stage in the explorationprogram was for the exploration geologist (LeoMiller) to make a compilation of the geology

Page 389: Introduction to mineral exploration

372 A.M. EVANS & C.J. MOON

based on outcrop and airborne magnetics. Afterstarting in Noranda, Miller moved to theTimmins area and investigated a number ofprospects based partly on the previous work ofthe Ontario Department of Mines (Bleeker &Hester 1999). Mapping in 1941 had identifiedtwo outcrops that implied a contact between afragmental rhyolite and basaltic pillow lavas(Berry 1941). Along the edge of the rhyolite out-crop, a narrow zone of sulfide-bearing siliceousrock was mapped (Fig. 15.11). During a visit tothe site on July 24 1958, Miller noted the pres-ence of a small trench dug by an unknown pros-pector. He also noticed traces of chalcopyrite inchert bands, although this was later shown notto be connected with the deposit.

The most promising areas were then flownby an EM system that had been mounted in ahelicopter (Donohoo et al. 1970). Although theKidd Creek anomaly was detected on the firstday of the production flying program (3 March1959), little further work was done as the exactlocation was not known due to lack of up-to-date photo mosaic maps and the general areaof the anomaly was at the boundary of fourprivately owned claims, originally granted toveterans of the Boer War (Fig. 15.12). One of theclaim owners was approached but would notoption the claim and the area was left for thetime being. Other anomalies in the area wereinvestigated and 59 holes were drilled over thenext 4 years with largely negative results. TheKidd Creek anomaly, known as Kidd 55 fromits location on the fifth lot pair and fifth con-cession (Figs 15.11 & 15.13), was further re-corded in routine flying in 1959 and 1960; thistime accurate location of the anomaly wasmade in 1961 at the second attempt. Optioningthe ground from the Hendrie estate began inJanuary 1961 but was not completed until June1963, by which time Miller had been trans-ferred to North Carolina and the prospectpassed to the supervision of Ken Darke. Lack-ing a budget to continue, Darke used moneyfrom his exploration program on Baffin Islandwith the permission of exploration mangerDick Mollinson. Texas Gulf Sulfur undertookground EM surveys to confirm the locationof the conductor and to delineate drill targets.Three conductors were found and drillholessited on the central anomaly (Fig. 15.14).

The first hole cut 8.37% Zn, 1.24% Cu, and3.9 oz t−1 Ag over 177 m in November 1963 as

shown in Table 15.4. It was obvious that amajor deposit had been discovered and the rigwas moved to the east of the property to con-fuse scouts from other companies who mightget wind of the discovery. In addition the drillcrew were kept in the bush until Christmaswhile lawyers optioned claims adjacent to thediscovery site to protect the possible extensionof the deposit. When land had been optioned inApril 1964 further drilling began and the initialphase lasted until October 1965, totalling 157holes and 33,500 m and this delineated thedeposit to the 335 m (1100 ft) level. The thick-ness of the deposit and its subcrop renderedit suitable for open pit mining and the initialdrilling was aimed at determining the orereserves for this operation and the upper part ofthe underground mine. Definition of the con-tinuation of the deposit has been from under-ground and continues to the present.

The primary phase of exploration drillingwas designed to define the ore outlines by drill-ing on levels 120 m (400 ft) apart. This wasfollowed by secondary drilling to define gradeand ore outlines on which the mine layoutwas based. An example (Fig. 15.15) shows thesecondary drilling in the shallow part of theunderground mine. The aim was to have anintersection every 55 m (180 ft) vertically; thiscorresponds to the main levels and every othersublevel. Drill spacing between sections wasnormally 15 m (50 ft). All the core was loggedand zones of mineralisation assayed (Matulichet al. 1974).

Drilling continues on the deposit at depth. In1995 more than 50,000 m were drilled. Of the23,000 m drilled in 1990, underground explora-tion drilling accounted for 7000 m, delineationdrilling 5500 m, and surface exploration drill-ing 10,500 m (Fenwick et al. 1991).

Drilling around the mine proved the exist-ence of a further blind massive sulfide lens,the Southwest Orebody, in 1977. It has sub-sequently been mined from 785 to 975 m and isconsidered to be a distal equivalent of the mainnorth and south orebodies (Brisbin et al. 1991).The area around Kidd Creek has been muchprospected without the discovery of otherdeposits and it is still an open question as towhether Kidd Creek is an isolated deposit orwhether it belongs to a cluster, whose othermembers have been eroded away, or remain tobe found by deep prospecting.

Page 390: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 373

FIG. 15.11 Mapping by Ontario Department of Mines that alerted Miller’s interest. Note the very limitedoutcrops. (After Bleeker & Hester 1999, from Berry 1941.)

N

1km0

75

75

Algoman granite; outcropAlgoman granite; inferredOutcrops of greenstone, pillow lavaOutcrops of rhyoliteStrike and vertical dip of stratumStrike and dip of schistosity

Kidd 55

Kidd Pond

Trench

Kidd

Creek

Trail

Rei

d Tw

pM

acD

iarm

id T

owns

hip

Jam

ieso

n Tw

pCarnegie Township

Jessop Township

Wark Township

Prosser Township

Kidd Township

Murphy Township

Lot pairs

1 2 3 4 5 6

Co

nce

ssio

nLots

123456789101112

I

II

III

IV

V

VI

15.3.3 Exploration geophysics

The discovery of the deposit prompted trials ofa number of techniques to assess their responseover the deposit (Donahoo et al. 1970). BesidesEM with various loop configurations, induced

polarization (IP), gravity, and magnetic groundsurveys were also made. The ground EM sur-veys confirmed the three anomalies detected inthe initial EM survey (Figs 15.14 & 15.16). Thecentral anomaly also gave a very high contrastIP chargeability and resistivity anomaly as well

Page 391: Introduction to mineral exploration

374 A.M. EVANS & C.J. MOON

1S

1aN

2S

3N

4N

5S

Height 33 m

35 m

33 m

47 m

35 m

38 m 0 1 km

FIELD PLOT

E - M

Anomal

y

1a

1 2

3

4

5100

0

In Phase

Quadrature

N

FIG. 15.12 Trace of first airborne anomalies over the Kidd Creek deposit, March 3, 1959. (After Donohoo et al.1970.)

as a 1.6 milligal residual gravity anomaly. Onlymagnetics failed to define the deposit and onlysucceeded in outlining the peridotite.

15.3.4 Exploration geochemistry

Although geochemistry was not used duringthe discovery of Kidd Creek a substantial re-search program was undertaken immediatelyafter the discovery to understand if surficialgeochemistry would have been of use. Detailscan be found in Fortescue and Hornbrook(1969) and there is an excellent summary inmore accessible form in Hornbrook (1975).

Kidd Creek is typical of the problems of sur-ficial geochemical exploration in the AbitibiBelt, an area known by soils scientists as the“Clay Belt.” The area was covered, before min-ing and logging, by dense forest with swampypeats, known as muskeg. Underlying this is athick (up to 65 m) sequence of glacial material:consisting of from top to base of clay till,

varved clay, and lower till. These glacial depos-its have a blanketing effect on the geochem-ical signature and vertical dispersion from thesubcrop of the deposit is limited to the lowertill (Fig. 15.17) as the varved clays have alacustrine origin and are not of local derivation.If the lower till is sampled by drilling a distinctdispersion fan can be mapped down ice fromthe deposit subcrop (Fig. 15.18). The moremobile zinc disperses further than copper.

The blanketing effect of the glacial over-burden limits the surface response to zincanomalies in near-surface organic soils. Thesezinc anomalies are thought to form as a resultof the decay of deciduous trembling aspen treeswhich tap the anomaly at relatively shallowdepths (Fig. 15.19).

Recognition of the lack of surface geochem-ical expression of Kidd Creek and other depositsin the Abitibi Clay Belt led to the widespreaduse of deep overburden sampling in geochem-ical exploration. The basal, locally derived till

Page 392: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 375

N

400m0

Conductor

<1000 γ1000-1100 γ>1100 γ

T.S. Woolings

Abitibi Powerand Paper Co.

T.S. Woolings

T.S. Woolings

T.S. Woolings

T.S. Woolings

T.S. Woolings A.R. McDonald

T.S. Woolings

J.H. RobertsEstate

J.F. Elliot Royal Trust Co.(Hendrie Estate)

Lot 5 Lot 4 Lot 3 Lot 2 Lot 1

VV

I

Kidd 55

DDH K55-1

Outcrop

Sulfides

Pit Outcrop

Trail

Bas

elin

e “A

0 S

20 S

24 S

30 S

40 S

50 S

FIG. 15.13 Map of the Kidd 55 locality showing the main EM conductor and the land holdings, the prospector’strench, and initial drillhole. (From Bleeker & Hester 1999 after Miller’s sketch of 1959.)

et al. (1974), Walker et al. (1975), Coad (1984,1985), and Brisbin et al. (1991).

Host rocks

VMS deposits in the Kidd Creek area exemplifythe classic case of exhalative deposits devel-oped at, or close to, the top of an Archaeanvolcanic cycle where the rhyolitic–dacitic top

is sampled by drilling and any dispersion fanscan be mapped.

15.3.5 Mine geology

The geology of the mine is discussed in detail inEconomic Geology Monograph 10 (Hannington& Barrie 1999) and there is a summary in Barrieet al. (1999). Earlier accounts include Matulich

Page 393: Introduction to mineral exploration

376 A.M. EVANS & C.J. MOON

of the lower cycle is succeeded by a dramaticchange to ultramafic and mafic volcanics atthe base of the succeeding cycle. Kidd Creekis no exception. This deposit occurs in an over-turned volcano–sedimentary sequence (Figs15.20 & 15.21).

The base of the sequence consists ofultramafic flows that structurally overlie grey-wackes of the Porcupine Group. The top of thesequence underlying the deposit consists ofa number of massive rhyolite sills intrudedinto a felsic series of pyroclastics containingintercalated rhyolite flows and tuffs. Theuppermost massive rhyolites, which have a U-Pb zircon age of 2717 ± 4 Ma (Nunes & Pyke1981), have pervasive crackle brecciation inter-preted as resulting from hydraulic fracturing.A stockwork chalcopyrite zone is present inthis crackle-brecciated rhyolite with the mas-sive ore lying immediately above. The orebodyis stratigraphically overlain by a minor devel-opment of rhyolitic rocks, dated at 2711 ±1.2 Ma by Bleeker and Parrish (1996), whichare succeeded by pillowed metabasalts. Therocks of the mine area are intruded by threelarge masses of high Fe tholeiite, gabbroicsills, referred to as andesite–diorite in mineterminology.

The rhyolites in the immmediate areaaround the deposit reach 300 m in thickness,in contrast to 100 m at a distance from theFIG. 15.14 Composite geophysical map of initial

surveys over the Kidd Creek deposit. (AfterDonohoo et al. 1970.)

Grid for groundgeophysics

BA

SE

LIN

E A

Residual gravity anomaly (0.2 milligals)

IP and resistivity anomaly

Ground EM anomaly

Massive sulfides

2N

24S

52S

0 300 m

N

Footage (ft) Copper Zinc Silver Lead(wt%) (wt%) (g t −−−−−1) (wt%)

0–26 Clay overburden26–50 24 1.05 Trace 10 –50–132 82 7.10 9.7 82 –

132–152 20 0.19 11.1 10 –152–196 44 0.11 4.7 17 –196–232 36 0.79 13.0 360 –232–248 16 0.18 3.81 82 –248–348 100 0.33 14.3 144 0.80348–490 142 0.1 18.0 247 –490–530 40 0.24 2.8 113 –530–566 36 0.23 6.1 55 –566–576 10 0.17 3.0 34 –576–628 52 0.20 8.3 62 –628–649 21 1.18 8.1 130 –649–655 6 – – – –

Average: 623 ft grading 1.21% Cu, 8.5% Zn, 138 g t–1 Ag, minor Pb; some Cd,about 0.1% indicated in character samples Texas Gulf Sulphur discovery holeK55-1 was drilled at –60 degrees to the west, in the northern half of lot 3,concession 5, Kidd Township.

TABLE 15.4 Assays for sections fromdiscovery hole K55-1. (From Bleker& Hester 1999. After a press releasein The Northern Miner, April 16,1964, with minor modifications.)

Page 394: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 377

W E

8 - 2 SUB

8 - 1 SUB

800 ft level

Waste

12 - 3 SUB

12 - 2 SUB

12 - 1 SUB

1200 ft level

ProposedFW fringe

ProposedHW fringe

215.

500

EOre l imit

Proposedpit l imit

0 30 m

0

4

2000

0Gam

mas

LIN

E A

Magnetic profile

OverburdenAndesites

Rhyolites

Massive sulfidesGraphitic shale

Peridotite

0 100 200 300 400 500 600Distance (m)

Mill

igal

s Gravity profileBouguer

Residual

+40°

–40°

Electromagnetic profile5000 CPS1000 CPS

BA

SE

FIG. 15.16 Geophysical section on line 24S ofFig. 15.14. (After Donohoo et al. 1970.)

FIG. 15.15 Underground drill patterns used to definereserves. (After Matulich et al. 1974.)

deposit, suggesting control of mineralisationby sub-volcanic rifting as do the geometry ofthe alteration and the shape of the VMS deposit(Fyon 1991).

Mineralisation

The orebody, which consists of three enechelon lenses, has a maximum thickness of168 m, a strike length of 670 m and is knownto extend to a depth of 2990 m. From 1966to 1995 101.5 Mt of ore carrying 2.31% Cu,0.24% Pb, 6.5% Zn, and 94 g t−1 Ag. were mined.At the beginning of 1996 reserves were calcu-lated to be 32.2 Mt grading 2.48% Cu, 0.21%Pb, 6.34% Zn, and 71 g t−1 Ag. For a numberof years tin, which may locally reach 3%, wasrecovered. There are three major ore types.In the stockwork there is stringer ore inwhich chalcopyrite stringers ramify crackle-brecciated or fragmental rhyolite. Pyrite or

pyrrhotite accompanies the chalcopyrite andthis ore averages 2.5% Cu. The massive oreis composed primarily of pyrite, sphalerite,chalcopyrite, galena, and pyrrhotite. Silveroccurs principally as native grains withaccessory acanthite, tetrahedrite–tennantite,stromeyerite, stephanite, pyrargyrite, andpearceite. Tin occurs as cassiterite with onlya trace of stannite. Digenite, chalcocite, andother sulfides occur only in trace amounts.

The massive sulfides are in part associ-ated with a graphitic–carbonaceous stratumcontaining carbonaceous argillite, slate, andpyroclastics. Associated with this horizon arefragmental or breccia ores that are consideredto represent debris flows.

The deposit is divided into the North andCentral plus South Orebodies by the MiddleShear. The orebodies can be divided into a num-ber of zones and, as might be expected, copper-rich areas occur in the base of the massive ore,including a bornite-rich zone at the base of theSouth orebody. In these very high silver gradesoccur (up to 4457 g t−1) with minor (0.35 ppmAu) associated gold values (section 15.2.2).

Page 395: Introduction to mineral exploration

378 A.M. EVANS & C.J. MOON

0 300 m

BA

SE

LIN

E H

Ice direction

BA

SE

LIN

E A

20S

24S

28S

32S

40S

50S08S

40S

1

2

3 13

4

15

5

7 23

Approx. extent of zinc anomaly in lower til lApprox. extent of copper anomaly in lower til l

Outcrop area

Subcrop of ore

Borehole

Area from which soil samples were collected

N

FIG. 15.18 Fan-shaped anomalies in the lowermosttill at Kidd Creek. (After Hornbrook 1975.)

Dep

th (

m)

BOREHOLE No. 23

Peat

Clay til l

Varved clay

Lower til l

Dep

th (

m)

BOREHOLE No. 1

Peat

Clay til l

Varved clay

Lower clay

0

401 5 10 50 100 500 50001000

Zinc (ppm)

1 5 10 50 100 500 50001000

Zinc (ppm)

0

20

FIG. 15.17 Section through glacial overburden overthe deposit and to one side of it; locations are inFig. 15.20. Note the response is limited to lowesttill and in the near surface in borehole 1. (AfterHornbrook 1975.)

Wall rock alteration

At both Kidd Creek and the HemingwayProperty (a drilled area 1.3 km north of the KiddCreek orebody containing minor base metalmineralisation) early serpentinization affectedthe ultramafic flows and silicification therhyolites. Later carbonate alteration was super-posed on both by pervasive CO2 metasomatism(Schandl & Wicks 1993). Strong chloritization,sericitisation and silicification occur in thefootwall stockwork zone along the length ofthe orebody (Barrie et al. 1999).

Genesis

The Kidd Creek Orebody is interpreted as beingof exhalative origin and, apart from its largesize, is similar to many other deposits ofPrimitive-type. Beaty et al. (1988) have shownthat the hydrothermally altered rhyolites areall markedly enriched in 18O compared withmost other VMS deposits. Altered rocks associ-ated with the large Iberian Pyrite Belt depositsand Crandon, Wisconsin (62 Mt of ore) showsimilar 18O enrichment (Munha et al. 1986) andfor Kidd Creek Beaty et al. have suggested that

Page 396: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 379

300

200

100

Stations

AB C D E F G H I J K L M N

NMLKJIHGFEDCBA

ppm

tota

l met

alin

–80

mes

h Organic soil

Mineral soil

Greenstone

Deciduous forest cover(trembling aspen)

Cut over area(alder cover)

Organic soil

Surficial material

Ore zone

Mineral soil

FIG. 15.19 Surface anomaly along line 24S of Fig. 15.18. The surface response in the organic soil is thought tobe the result of the tapping of the deep anomaly by trembling aspen trees. (After Hornbrook 1975.)

evaporation and enrichment of normal seawater produced a high salinity fluid enrichedin 18O that was capable of transporting largeamounts of metals, so accounting for the largesize of the deposit. Taylor and Huston (1998)argued that regional scale low temperaturefluid flow may be related to regional mafic-ultramafic heat sources such as the komatiiteflows.

Strauss (1989) has shown that the sulfurisotopic values are similar to those of manyother Archaean VMS deposits. Maas et al.(1986) using Nd isotopic data suggested thatsome of the base metals in the ore were leachedfrom the rhyolitic host rocks.

15.3.6 Mining operations

The original operators were Ecstall Mining Ltd,a wholly owned subsidiary of Texas Gulf SulfurInc., of the USA. Following a 1981 takeoveroffer by Elf Aquitaine (SNEA) Kidd CreekMines Ltd became a wholly owned subsidiary

of the government-owned Canada Develop-ment Corp. In 1986 Falconbridge Ltd ofCanada, which has subsequently become partof Noranda Inc., acquired Kidd Creek.

When the decision was made in March 1965to build the open pit–railroad–concentratorcomplex two of the major problems were soilsand transportation. The whole vicinity of theorebody was covered with thick deposits ofweak materials: muskeg, till, and clay. Onlytwo small outcrops occurred by the pit site andone had to be reserved for the primary crusher.There was no good ground near the pit largeenough for the concentrator and no roads orrailroads to the pit area at that time. The con-centrator was therefore built 27 km east ofTimmins near a highway, a railroad, an elec-tricity transmission line, and a gas pipeline. Itwas linked to the mine with a new railroad.

The open pit was designed to have an overallwaste-to-ore ratio of 2.58:1 but started with a4:1 ratio (see section 11.2.1). It was 789 × 549 mat the surface, ore was available from the pit on

Page 397: Introduction to mineral exploration

380 A.M. EVANS & C.J. MOON

Mid

dle

shea r

Andesite–diorite Massive ore

Basalt Stockwork ore

RhyoliteQuartz porphyry

0 100 m

N

FIG. 15.20 Bedrock geological map, Kidd Creek.(After Beaty et al. 1988.)

FIG. 15.21 West–east section through the NorthOrebody. (After Beaty et al. 1988.)

bench 1 by October 1965, and large scale pro-duction to supply the concentrator with 8000–9000 t day−1 was achieved by October 1967.

The pit had a projected production life ofabout 10 years, however it was decided thatthere were advantages in developing an under-ground mine fairly quickly. These advantagesincluded: the availability of underground orefor grade control, which would become moredifficult as the pit deepened; the chance todevelop new systems, equipment, and skills,which was possible as a result of revolutionarychanges in underground mining technology inthe 1960s; and the acquisition of knowledge oflocal rock behavior to aid in mine planning.

Shaft sinking commenced in March 1970 andwhat became known as the Upper Mine with a930-m-deep shaft was developed. The 1550-m-deep Lower Mine shaft, 100 m SW of the No. 1shaft was completed in 1978 (Fig. 15.22). Theopen pit was abandoned in 1979. Blastholeand sublevel caving stopes followed by pillarrecovery using blasthole methods have beenemployed in the Upper Mine. In the LowerMine a modified blasthole technique withbackfill of a weak concrete mix is used so thatno pillars are left. Waste is mixed with cementto form backfill at the 790-m level and thentrucked to the stopes. The ore crusher station ison the 1400-m level. A subvertical shaft, the

Page 398: Introduction to mineral exploration

15: KIDD CREEK MASSIVE SULFIDE, ONTARIO 381

W E

2000 level

4000 level

6000 level

7000 level

Crusher

Orebody

Ramp

Open pitShafts

Ultramaficrocks

RhyoliteBasaltic

rocks

Crusher

Ramp

Ore pass

SH

AF

T N

o 2

SH

AF

T N

o 1

SH

AF

T N

o 3

FIG. 15.22 Schematic east–west section through theKidd Creek Mine showing location of shafts andramp. (After Brisbin et al. 1991.)

someters, strain cells, and blast vibrationmonitors to detect and predict ground instab-ility leading to safer operations, the design ofmore stable underground workings, and thusgreater productivity.

15.3.8 The concentrator and smelter

The highly automated concentrator (seesection 10.2.2) has undergone a number ofdevelopments and improvements and by 1981its capacity was up to 12,250 t day−1 (about4.5 Mt yr−1) and it was producing seven differ-ent concentrates: copper concentrates withhigh and low silver contents, two similar zincconcentrates, lead, tin and pyrite concentrates.After about 9 years of planning and the testingof various smelting processes, a smelter andrefinery complex was built near Timmins andwent into production in June 1981. This com-plex treats two thirds of the copper concen-trates and produces 90,000 t of cathode copper.Most of the rest of the copper concentrate iscustom smelted and refined by Noranda Inc.Much of the zinc concentrate is treated inthe electrolytic zinc plant which can produce134,000 t zinc metal per year. The remainingzinc concentrates and all the lead concentratesare sold to custom smelters.

15.4 CONCLUSIONS

VMS deposits can constitute very reward-ing targets as they can be large, high grade,polymetallic orebodies like Kidd Creek. Theyshould be sought for using an integrated com-bination of geological (sections 15.2 & 15.3.1),geochemical (sections 15.2.9 & 15.3.4), geophy-sical (sections 7.4, 7.9, 7.10, 15.2.10 & 15.3.3),and drilling (section 15.2.11) methods. In theimmediate World War II years, when geochem-ical prospecting was in its infancy, the use ofairborne EM led to many successful discoveriesin the Canadian Shield. Experience and re-search since then has shown that sophisticatedgeochemical procedures (section 15.3.4) andother geophysical methods can play an import-ant part in the search for more VMS orebodies.

No. 3 shaft, is being sunk to access the lowerpart of the deposit to a depth of 3100 m. Overallproduction was approximately 2 Mt from theNo. 1 and No. 2 mines in 2001.

15.3.7 Rock mechanics

Considerable use has been made of rockmechanics instrumentation such as exten-

Page 399: Introduction to mineral exploration

382 M.K.G WHATELEY, T. BELL & C.J. MOON

16Disseminated PreciousMetals – Trinity Mine,

NevadaMichael K.G. Whateley, Timothy Bell

and Charles J. Moon

Disseminated precious metal deposits, particu-larly of gold, have been the most popular targetfor metallic exploration through the 1980sand 1990s. Their popularity was based on thehigh price of gold during this period (see sec-tion 1.2.3), the low cost of mining near-surfacedeposits in open pits, and improvements ingold recovery techniques. The key metallur-gical innovation was the development of thecheap heap leach method to recover preciousmetals from low grade oxidized ores. This casehistory deals with the evaluation of the Trinitysilver mine.

16.1 BACKGROUND

16.1.1 Overview of deposit types

A variety of deposit types have been mined inthis way but the most important in Nevada(discussed in detail by Bonham 1989) are:1 Sediment hosted deposits of the Carlin type.2 Epithermal deposits, such as Round Moun-tain.3 Porphyry-related systems rich in gold.

Carlin-type deposits

These deposits are responsible for the majorityof gold production in Nevada, 261 t in 2001.Typically the gold occurs as micronmeter-sizedgrains which are invisible to the naked eye(“noseeum” gold), within impure limestones orcalcareous silstones. The host rocks appear to

be little different from the surrounding rocksand the margins of the deposit are only distin-guishable by assay. An excellent summary isgiven in Bagby and Berger (1985), updated inBerger and Bagby (1991) and Hofstra and Cline(2000).

Individual deposits occur mainly withinmajor linear trends, up to 34 km in length, themost important of which is the Carlin Trendin Northeast Nevada. This trend and typeof deposit are named after the Carlin Mine,which was the first significant deposit of thistype to be recognized, although others hadpreviously been mined. The deposits vary frombroadly tabular to highly irregular withinfavorable beds and adjacent to structures suchas faults that have acted as fluid pathways.The faulting and fluid flow has often resultedin brecciation and several deposits are hostedin breccias.

Silicification is the commonest form of al-teration and the occurrence of jasperoid, whichreplaces carbonate, is widespread in many de-posits. Gold is thought to occur predominantlyas native metal although its grain size is verysmall, usually less than 1 µm, and the metal’sexact mineralogical form remains unknown insome mines. Gold forms films on sulfide andamorphous carbon and is particularly associ-ated with arsenian pyrite. Pyrite is the mostcommon sulfide and may be accompanied bymarcasite and arsenic, antimony and mercurysulfides. The arsenic sulfides realgar andorpiment, which are readily distinguishable bytheir bright color, occur in many deposits as

Page 400: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 383

et al. 1987, Bonham 1989, Henley 1991,Hedenquist et al. 2000).

The high sulfidation or acid sulfate typegenerally contains more sulfide and is char-acterized by the occurrence of the coppersulfide, enargite. Other base metal sulfidesare frequently present as is pyrite. Advancedargillic alteration is common as is alunite(K2Al6(OH)12(SO4)4) and kaolinite. The lowsulfidation or adularia sercite type generallyhas less sulfide and sercitic to intermediateargillic alteration. Alunite may be present butit is only of supergene origin. The structuralsetting of both types may be similar (Fig. 16.2)but the differences between the two types prob-ably reflect their distance from the heat source(Heald et al. 1987). There are obvious parallelsbetween the formation of these deposits andmodern geothermal systems. Evidence frommodern day geothermal systems, such as thedeposition of gold on pipes used to tap thesystems for geothermal power, suggests thatthe gold may have been deposited rapidly.Several major gold deposits occur in alkalicvolcanics that may be spatially associated withalkalic porphyries. The deposits are character-ized by the occurrence of the gold as tellur-ides with quartz–carbonate–fluorite–roscoelite(vanadium mica)–adularia alteration and re-gional propylitic alteration (Bonham 1989).

The Hot Spring deposits of Bonham (1989)appear to be a variant of the epithermal typeand are characterized by the occurrence of arecognizable paleosurface that is usually asso-ciated with a siliceous sinter zone interbeddedwith hydrothermal breccia. The sinter passesdownward into a zone of silicification andstockwork with a zone of acid leaching. Theprecious metals are generally restricted towithin 300 m of the paleosurface and probablyreflect rapid deposition. There is an overallspatial association with major centers ofandesitic to rhyolitic volcanism

Gold-rich porphyries

Gold is of major economic importance inseveral porphyry deposits that produce copperas their main product, such as Ok Tedi (seep. 273). More recent research (Sillitoe 1991)suggested that there may be related copper-deficient systems that contain economic gold

SURFACEFumeroles and hot springs

Fine-grainedsediments

clastic

Jasperoid

Limestone

Fluid flow

Thin bedded,carbonate cemented,carbonaceous,clastic sediments

500 m

Orebody

Limestonesand dolomites

Felsic orintermediate

intrusive

FIG. 16.1 Diagram for the formation of Carlin styledeposits. (After Sawkins 1984.)

does stibnite and cinnabar. The general geo-chemical association is of As, Sb, Tl, Ba, W andHg, in addition to Au and Ag.

The deposits are spatially associated withgranites and many authors favor intrusiondriven hydrothermal systems (Fig. 16.1). Bergerand Bagby (1991) favor a mixing model inwhich magmatic fluids have been mixed withmeteoric water causing precipitation in favor-able host rocks.

Epithermal systems

Much of the bonanza-style precious metalmineralisation in the western USA is hosted involcanic rocks, e.g. the Comstock Lode whichproduced 5890 t Ag and 256 t Au at an averagegrade of about 300 g t−1 Ag and 13 g t−1 Au in thelate nineteenth century (Vikre 1989). Morerecent exploration has resulted in definition ofbulk mineable targets such as Round Moun-tain with reserves of 196 Mt of 1.2 g t−1 Au.Volcanic hosted deposits can be divided intothree general types based on mineralogy andwall rock alteration: acid sulfate (or alunite–kaolinite), adularia sericite, and alkalic (Heald

Page 401: Introduction to mineral exploration

384 M.K.G WHATELEY, T. BELL & C.J. MOON

contents. The deposits are associated withgranitic intrusives that vary in compositionbut with quartz monzonite and diorite as themost important hosts. Alteration shows a

characteristic concentric zoning from innerpotassic through phyllic to propylitic zones.Gold generally correlates with copper gradeand is associated with K alteration (Sillitoe

Pervasive acid leachingand silicificationIntermediate argillicalteration

Adv. argillic alterationOrebody

(a)

(b)

Pa

PaPa

Intrusion

Fumaroles andhot springs

Silica sinterOrebody

Basement

Intrusion

Young volcanicsand sediments

1 km

200°

250°

300°

200°

Basement

250°

300°

250°20

FIG. 16.2 Diagrams showing theformation of two types ofepithermal precious metal depositin volcanic terranes. (a) Acidsulfate type. Pa, propyliticalteration. Note that themineralisation occurs within theheat source. (b) Adularia–sericitetype. The upwelling plume ofhydrothermal fluid is outlined bythe 200°C isotherm. Themushroom-shaped top reflectsfluid flow in the plane of majorfracture systems, a much narrowerthermal anomaly would be presentperpendicular to such structures.The heat source responsible for thebuoyancy of the plume is shown asan intrusion several kilometersbelow the mineralised zone. Inboth diagrams the arrows indicatecirculation of meteoric water. (a) isdrawn to the same scale as (b).(Based on Heald et al. 1987 withmodifications from Henley 1991.)

Page 402: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 385

1991). In some cases high sulfidation goldsystems are mined at higher levels than theporphyry, although a direct connection is oftenhard to prove. For further discussion of por-phyry copper deposits see Evans (1993) andRobb (2004).

16.1.2 Geochemistry

One of the major features in the discovery ofdisseminated gold deposits has been the furtherdevelopments in gold geochemistry. The recog-nition of “noseeum” gold deposits depends onthe ability to detect fine-grained gold. As thisis usually in the micron size class it does notoften form grains large enough to be panned.Much gold remains encased in fine-grainedsilica and therefore sampling and analyticaltechniques must be devised that accuratelydetermine all the gold present. This analyticalproblem has largely been overcome, firstly bythe improvement of graphite furnace AAS andmore latterly by the application of neutronactivation analysis and ICP–MS (see section8.2.3). Graphite furnace AAS methods haverelied on the separation of gold from the rockor soil matrix using an organic solvent. Thisallows analysis with detection limits of lowerthan 5 ppb.

As gold is present as discrete grains withinthe sample either encased in the matrix or asfree gold, sampling methods have to be devisedthat will allow the analytical aliquot to repres-ent accurately the original sample. Investiga-tions by Clifton et al. (1969), well summarizedby Nichol et al. (1989), have shown that it isnecessary to have 20 gold grains in each sampleto achieve adequate precision. In real termsthis means taking stream sediment samples ofaround 10 kg and carefully subsampling. Thegeologist who takes a 1-kg sample can not ex-pect to make an accurate interpretation of ana-lytical data. An alternative strategy, which hasbeen used where rocks have been oxidized, is totake a large sample and leach it with cyanide;this technique is known as Bulk Leach Extract-able Gold (BLEG). This technique has foundwide acceptance in Australia and the PacificRim countries.

Most but not all gold deposits have con-centrations of pathfinder elements associatedwith the gold. In Nevada the most widely used

elements are As, Sb, Hg, Tl, and W for sedimen-tary hosted deposits, with the addition of basemetals for most deposits. Although pathfinderelements can be extremely informative as tothe nature of the gold anomaly, they should notbe used without gold as pathfinder elementsmay be displaced from gold. In addition somegold deposits, particularly adularia–sericite-type deposits, have no pathfinder signatureand the exploration geologist must rely largelyon gold geochemistry. An example is theOvacik deposit in Turkey (see section 4.2.1)that was discovered by the follow-up of BLEGgeochemistry.

In Nevada large stream sediment samplesand extensive soil sampling have proved effec-tive where overburden is residual. Howeverin some areas the overburden has been trans-ported and surface sampling is ineffective.Chip sampling of apparently mineralized rocks,particularly jasperoids, has been successful al-though jasperoidal anomalies are often slightlydisplaced from the associated deposits.

There are a large number of case historieson gold exploration and the reader is advised toread Zeegers and Leduc (1991) and case his-tories on Nevada in Lovering and McCarthy(1977).

16.1.3 Geophysics

Disseminated gold deposits are difficult geo-physical targets. Electrical and EM methodshave been used to map structure and identifyhigh grade veins, e.g. at Hishikari, Japan(Johnson & Fujita 1985), but are not usuallyeffective in detecting mineralisation exceptindirectly. On occasion unoxidized sulfides arepresent allowing the use of induced polariza-tion (IP). One of the more successful uses ofgeophysics in Nevada has been the detectionof silicified bedrock under Tertiary cover(Fig. 16.3). The section shows the results of aControlled Source Audio Magneto TelluricSurvey that clearly detected a zone of silicifica-tion under 50 m of cover. Further case studiescan be found in Paterson and Hallof (1991).

16.1.4 Mining and metallurgy

The recognition that disseminated preciousmetal deposits could be successfully bulk

Page 403: Introduction to mineral exploration

386 M.K.G WHATELEY, T. BELL & C.J. MOON

grade material has previously unsuspected highgrade material with different metallurgicalcharacteristics beneath it. For example, at thePost deposit on the Carlin trend high gradematerial in which gold is difficult to extract, orrefractory, underlies the main surface deposit.The grade of this material is so much higherthat it is worthwhile investing in undergroundmining and pressure leach extraction to re-cover the gold.

16.2 TRINITY MINE, NEVADA

The Trinity Silver Mine was an open pit, heapleach, silver mining operation extractingrhyolite-hosted, disseminated, hydrothermal,silver oxide mineralisation. The mine wasbrought into production on September 3, 1987and mining ceased on August 29, 1988. Theheap leach operation continued for just overone more year. Our choice of a silver minerather than a gold producer has been governedentirely by the availability of data. Most ofour discussion is of course applicable to anydisseminated precious metal deposit. USmeasures are used throughout this chapter asthe mining industry in the USA has yet tometricate. Thus oz t−1 are troy ounces per shortton of 2000 avoirdupois pounds (lb).

This case study focuses upon the statisticalassessment of blasthole assay data, for thepurpose of determining the distribution andquality of mineralisation. It is essential for reli-able ore reserve calculations to derive the bestgrade estimate for an orebody. The actual minedreserves and estimated reserves calculatedby US Borax and Chemical Corporation (USBorax) from exploration borehole data are alsodescribed. The Surpac Mining System Softwarewas used for data management and statisticalevaluation. This work was undertaken as anMSc dissertation at Leicester University (Bell1989) under the auspices of US Borax.

The Trinity Mine is in Pershing County,Northwest Nevada (Fig. 16.4), on the north-west flanks of the Trinity Range, 25 km north-northwest of Lovelock. The area is covered bythe souhtwest part of the USGS Natchez Spring7.5′ topographic map and Pershing CountyGeological Map. Elevations vary from 1200 mto 2100 m (3900–7000 ft). The mineralisation

400200

300

100

1010

Carlin VolcanicsCarlin TertiarySediments

Surface

Queen FormationRuss Formation

DrillholeSilicification in the

Palaeozoic bedrockformations, gold

values are present

1248

163264

128256512

1024204840968192

Fre

quen

cy (

Hz)

0W400W800 E400 E800Station location (ft)

0

100

200Dep

th (

m)

50

150

250

FIG. 16.3 Section showing the application of aCSAMT survey in detecting areas of silificationunder Tertiary cover.Top: contoured data; bottom:drill section. (After Paterson & Halof 1991.)

mined and processed, sometimes with specta-cular returns, has been responsible for the allo-cation of the large exploration budgets spent inNevada. Oxidized deposits are generally minedin open pits, usually with low stripping ratios.One of the keys to a successful operation iseffective grade control (section 16.5) as differ-ent grade ores are treated differently in mostoperations. The higher grade material is usu-ally milled and extracted with cyanide in a con-ventional manner whereas the lower grade oreis crudely crushed and placed on leach pads.These leach pads are built on plastic liners,sprinkled with cyanide, using a system similarto a garden sprinkler, and the cyanide allowedto percolate through the crushed ore. Thecyanide is then pumped to a central treatmentplant where the gold is extracted by passing thecyanide through activated carbon. Marginal oreand waste are stockpiled separately to allow forthe later treatment of marginal material if thegold price increases.

One of the interesting developments inNevada is the recognition that some of the low

Page 404: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 387

ThrustsOrdovician sedimentsCretaceous Granodiorite+ +JurassicQuaternary

Cambrian sedimentsPermian sedimentsTriassic sediments(ALSG)

Tertiary sedimentsand volcanicsRecent

0 20 km

Black RockDesert

+

+

+

+

++

++

+

+

+

++

+

+

++

+

++

++ +

+

+

+

++

++

++

+

+

+

+

+

+

+

R30E R40E

T25N

T30N

Buena VistaValley

BuffalloValley

TobinRange

HumboldtRange

Hum

bold

t R

iver

Lovelock

WillowCanyon

TrinityRange

Granite SpringValley

SevenTrough Range

SeleniteRange

STUDYAREA

CentralRange

STUDY AREA

Lovelock

Las Vegas

NEVADA

N

FIG. 16.4 Sketch map of Pershing County, Nevada showing the outline geology and location of the Trinitysilver deposit. ALSG, Auld Lang Syne Group. (After Ashleman 1988.)

at Trinity principally lies between 1615 and1675 m (5300 and 5500 ft).

16.3 EXPLORATION

16.3.1 Previous work

Mineralisation was first discovered in the Trin-ity Range by G. Lovelock in 1859 and limited,unrecorded production from Pb-Ag and Ag-Au-Cu-Zn veins occurred between 1864 and 1942(Ashleman 1983). Evidence of minor prospect-ing during the 1950s was found within WillowCanyon (Fig. 16.5) (Ashleman 1988). Geophys-ical exploration and trenching during the1960s was carried out by Phelps Dodge withinTriassic sediments to the north of WillowCanyon. Later a geochemical exploration pro-gram aimed at locating gold mineralisation wasundertaken by Knox–Kaufman Inc. on behalf

of US Borax, but they were unable to locatepreviously reported gold anomalies. Howeverin 1982 a significant silver show was foundwithin altered rhyolites. As a result, agree-ments were reached between five existingclaim holders, Southern Pacific Land Company(SPLC), holders of the surface rights and USBorax, after which the Seka claims were staked.The Trinity joint venture between Sante FePacific Mines Inc., the land holder and a sub-sidiary of SPLC, and Pacific Coast Mines Inc.,a subsidiary of US Borax, was operated by USBorax and mining was contracted to LostDutchman Construction.

16.3.2 US Borax exploration

Various stages of exploration were completedby US Borax during the period 1982–86. In 1982and 1983 mapping, geochemical and geophys-ical techniques were used to located payable

Page 405: Introduction to mineral exploration

388 M.K.G WHATELEY, T. BELL & C.J. MOON

FIG. 16.5 Sketch map of the geology in the vicinity of the Trinity silver deposit, Nevada. (After Ashleman1988.)

OXIDE

SULFID

E

ZONE

ZONE

South - westernLobe

North - easternLobe

WILLOW

CANYON

Area>1.3 oz t–1Ag (45 g t–1)

1615

1615

1615

1645

1675

1645

1600

Approximateoutline of

the open pit(South-western

extension)

Tertiary

Early Jurassic-Triassic

Epiclastic + Pyroclastic Tuff

Auld Lang Syne Group

Unconformity

Fault

QuaternaryLate Tertiary

Tertiary

Surficial Deposits

Sugary Rhyolite

Rhyolite Porphyry

Aphanitic Rhyolite

N

0 100 m

Page 406: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 389

sulfide mineralisation. In 1983 drilling com-menced within an area known to host sulfidemineralisation. Exploration continued from1985 to 1986 in an attempt to find additionalreserves. Although assessment of the sulfideore continued with metallurgical testing andfeasibility studies, the possibility of a highgrade oxidized zone to the southwest of themain target area became the focus of attention.Efforts were thereafter concentrated upon theevaluation of this southwestern extension(Fig. 16.5).

16.3.3 Mapping

Initial mapping of the claim area at 1:500 wasfollowed by detailed mapping of the sulfidezone at a scale of 1:100. Mapping of structureand lithology helped define the surface extentof mineralisation and alteration, and the con-trols of the distribution of mineralisation.When the potential of the silver mineralisationwithin the oxidized zone was established inmid 1986, detailed mapping of the southwest-ern extension commenced. Mapping showedthe host rock, of both oxide and sulfide miner-alisation, to be fractured rhyolite porphyryand defined an area of low grade surface miner-alisation intruded by barren rhyolitic dykes(Fig. 16.5). These barren dykes divide the south-west extension into northeast and southwestlobes. The rhyolite dykes were seen to beintruded along east and north-northeast faultsand to disrupt the stratigraphical sequencewithin the southwest lobe.

16.3.4 Geochemistry and geophysics

A geochemical soil survey was completed overthe oxide and sulfide zones and samples wereanalyzed for Pb, Zn, and Ag. The mobility of Agand Zn in the soil horizons placed constraintsupon the significance of anomalies reflectedby these elements. Lead is more stable andwas therefore used as a pathfinder element forAg mineralisation. Significant Pb anomalieswere located over the sulfide zone. Anomaliesof greater than 100 ppm defined potential min-eralized areas and higher levels (>1000 ppm)coincided with the surface intersection ofAg mineralisation. Surface rock samples weretaken as part of a reconnaissance survey, and

contour maps of the results helped define theextent of surface mineralisation.

Between 1985 and 1986 rock geochemistrywas applied to the oxide zone in the southwestextension. Rock chips from drilling, trenchingand surface exposure were analyzed for Pb, Zn,Ag, and Au. High grade samples containing30–50 ppm Ag were reported and localizedsamples containing 40–90 ppm Ag were re-corded as the zone of oxidization was tracedsouthward along strike. Anomalies from therock and soil geochemical surveys were used inthe planning and layout of the drilling programwithin the sulfide zone.

Geophysical surveys, including gradientarray IP, magnetic and gamma ray spectro-metry, were carried out concurrently over thesulfide mineralisation. The gradient inducedpolarization (IP) survey required three arrange-ments of transmitting electrodes with a polespacing of 12,000 ft (3650 m). Readings weretaken every 500 ft (150 m) along lines spaced1000 ft (300 m) apart (Ashleman 1983). A timedomain chargeability anomaly coincided withthe area of sulfide mineralisation. A NNWtrending belt of high resistivity was locatedadjacent to the eastern margin of an area ofknown subsurface mineralisation. The mag-netic survey proved to be of little value due tothe poor contrast in magnetic values. Gammaray spectrometry using K, Th, and U channelswas tested, but only subtle differences wereseen between altered and unaltered rocksand the suvey was discontinued (Ashleman1983).

16.3.5 Drilling

Successive stages of drilling on a 100 ft (33 m)grid orientated 11 degrees west of true north(Fig. 16.6) totalled 29,880 m (98,031 ft) drilled(Table 16.1). Most of the drilling was withinareas of principal sulfide mineralisation, focus-ing on the surface exposure of silver mineral-isation, lead, and IP anomalies and areas offavorable geology. Drilling defined a body ofsilver mineralisation dipping to the west.

The drilling grid was extended to explorefor additional reserves and high grade silveroxide mineralisation was intercepted to thesouthwest (Fig. 16.5). In 1985 drilling wasconcentrated within the southwest extension.

Page 407: Introduction to mineral exploration

390 M.K.G WHATELEY, T. BELL & C.J. MOON

● ●

●●

●●

●●

●●

●●

●●

●● ●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

E 8000 E 8200 E 8400 E 8600 E 8800 E 9000 E 9200 E 9400 E 9600 E 9800

N 10200

N 10000

N 9800

N 9600

N 9400

N 9200

N 10400

N 9000

Conventional percussion drill ing●

Reverse circulation drill ing

Diamond drill ing

Outline of final pit

0 50 m

FIG. 16.6 Map showing the location of the exploration boreholes in relation to the final outline of the Trinitymine open pit from which the oxide mineralisation was extracted.

Ninety-two of the holes were drilled to definethe oxide zone. Although mineralisationwas located south of the southwest extensionalong strike and at depth, it was noted to bediscontinous.

Diamond drill holes were sampled every 1–3 ft (0.3–0.9 m) for metallurgical testing, while

percussion holes were sampled as 5 ft (1.5 m)composites (Ashleman 1986). Samples weresent to the laboratory and analyzed for Ag, Au,Pb, Zn, and As. Analyses on the samples fromthe exploration holes were carried out usingatomic absorption spectrometry (AAS), whichgave the total silver content. Samples recording

TABLE 16.1 Summary of drilling exploration programs.

Drilling methods Number of holes Total meterage

Percussion 199 22,857 (74,990 ft)Reverse circulation 39 6257 (20,530 ft)Cored 10 765 (2511 ft)

Total 258 29,879 (98,031 ft)

Page 408: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 391

Ag in excess of 1000 ppm were reanalyzedusing fire assay.

16.4 THE GEOLOGY OF PERSHING COUNTYAND THE TRINITY DISTRICT

16.4.1 Regional setting – Pershing County

The Trinity Mining District of PershingCounty is part of the Pacific Rim MetallogenicBelt. The mountain ranges of Pershing Countyreflect the surface expression of major NE–SWstructural lineaments sequentially repeated(basin and range) across Nevada. Much of thepresent geological setting is the response toregional compression subsequently modifiedby extension. The regional geology is outlinedin Fig. 16.4.

Cambrian to Permian shallow marinecarbonates and clastics are exposed to thenortheast and east of Pershing County. TheHarmony Formation (U. Cambrian) is com-posed of quartzite, sandstone, conglomerate,and limestone which pass laterally into clasticand detrital sediments of the Valmy Forma-tion (Ordovician). Carboniferous to Permiansediments (limestone, sandstone, and chert)and volcanics (andestic flows and pyroclastics)form most of the Tobin Range to the east.During the Mesozoic fault-bounded, deepwater, sedimentary basins dominated the area,with associated widespread volcanic and silicicintrusive activity.

Triassic cover is widespread in centralPershing County. The lower units to the eastconsist of interbedded andesites, rhyolites, andsediments, locally intruded by leucogranitesand rhyolite porphyries. To the southeast theMiddle Trias is exposed as calcareous detritaland clastic sediments, limestone, and dolo-mite. The upper unit which forms the majorsedimentary cover within the Trinity, SouthHumbolt, and Severn Troughs Range, definescyclic units of sandstone and mudstone withlimestone and dolomite. Adjacent to Cretace-ous granodiorite intrusions, the Triassic toJurassic sediments are locally metamorphosedto slates, phyllites, hornfels, and quartzite.

The Auld Lang Syne Group (ALSG) forms themajority of the Triassic sedimentary sequenceand reflects the shallow marine conditions of

a westerly prograding delta (Johnson 1977).Tertiary basaltic and andestic volcanics passinto a thick, laterally variable, silicic, volcano-clastic sequence consisting of a complex pileof rhyolite domes, plugs, flows, and tuffs. TheTertiary rhyolitic volcanics are overlain andinterdigitate with lacustrine and shallow lakesediments, later intruded by silicic dykes, sills,and stocks. The Tertiary sequence is coveredby Quaternary to Recent gravel and alluvialdeposits. The Miocene to Pliocene depositsform a thick stratified, well-bedded sequence oftuffs, shales, sandstone, clay and gravel, locallyinterbedded with basaltic and andestic flows.

Various structural phases are apparentwithin Pershing County (Johnson 1977). Pre-Cenozoic large scale compressional folding andthrusting during the Sonoma Orogeny (midJurassic) resulted in the south and southeastmovement of basinal sediments and volcanicsover shelf sediments. These folded, imbricate,thrust slices were later offset by dextral strike–slip faulting that has northwest arcuate anddiscontinuous trends (Stewart 1980). Suchthrusting is apparent to the east and southeastwithin the Central and Tobin Ranges (Fig. 16.4).Crustal extension during the Cenozoic resultedin northerly trending basin and range faulting.Associated with the basin and range faultingare effusive phases of bimodal volcanism.Stewart (1980) records both pre- and post-basinand range N–E and E–W fault sets, and impliespossible strike–slip components to the normalfaults.

Tertiary silver mineralisation is reviewed bySmith (1988b), who classifies the silver-bearingdeposits into three categories; disseminatedtype, vein type, and carbonate hosted. Theformer two types are outlined and comparedwith the Trinity deposit in Table 16.2. A fur-ther classification into four subtypes (Bonham1988) includes volcanic-hosted epithermal,silver-base metal limestone replacement de-posits, sedimentary-hosted stockworks, andvolcanic-hosted stockworks. Pershing Countyis dominated by Pb-Zn-Ag-Au veins hostedwithin and adjacent to Cretaceous granodi-orites, with minor occurrences in Triassicsediments. Mineralisation tends to be con-centrated in shear zones and along hornfelsdykes forming pinch and swell structures alongoblique joint sets (Bonham 1988).

Page 409: Introduction to mineral exploration

392 M.K.G WHATELEY, T. BELL & C.J. MOON

TABLE 16.2 Outline of silver deposits of Nevada, including a comparison with the Trinity deposit. (Adaptedfrom Smith 1988.)

Host rock

Type ofmineralisation

Morphology

Mean grade

Geologicalcontrol

Zoning

Mineralogy

Alteration

Process offormation

Hydrothermalsolutions

Disseminated

Tertiary felsic volcanics andclastic sediments, Devonianlimestone

Discrete grains and veinstockworks

Tabular, podiform bodies,defined by assay cut-off of<300 g t−1

270 g t−1

Structural, rockpermeability and proximityto conduit

Increased mineral contentwith depth, high Pb–Zn–Mn, low Au, Ag:Au>100

Acanthite, tetrahedrite,native Ag, pyrite andPb–Zn base metals

Silica–sericite (carbonate,clay, chlorite, potassicfeldspar)

Hypogene

165–200°C, 5 wt% NaCl,low salinity groundwatermixing with oilfield brinesand wall rock alteration

Vein

Tertiary andesites andrhyolites

Banding in quartz veins

Veins defined by faultcontacts with assay cut-offbetween 200 and 300 g t−1.Change in cut-off has littleeffect on tonnage

500–800 g t−1

Structural and fracturedensity, intersection offault and fractures

Sharp vertical zoning,sulfides increase withdepth, Ag:Au>50

Acanthite, low pyrite andsulfide content

Extensive and variable,hematite–magnetite,propylitic, adularia andillite

Hypogene, hydrofracturing,and brecciation

200–300°C, meteoricsolutions, precipitationinduced by mixing, boiling

Trinity

Teritary porphyry rhyolite,minor occurrence in tuff,and argillite

Disseminations,microfractures, veinlets,and breccia infill (veinstockwork)

Lenticular and stackedtabular, sharp assayboundaries of <30 g t−1

160–250 g t−1

Structural, rockpermeability, faultintersections, percentagefracture density

Sulfide-oxide zoning, highAg:Au

Sulfides: friebergite–pyragyrite, chalcopyrite,pyrite, stannite. Oxides

Silicification, adularia–quartz–sericite, minorpropylitic, illite andkaolinite

Epithermal andhydrotectonic fracturing

Not available

(Ashleman 1988) have been subjected to bothlow grade regional metamorphism and contactmetamorphism. The latter is associated withthe intrusion of Cretaceous granodiorite dykesand stocks located to the northeast of the mine(Fig. 16.4). Outcrops of the ALSG to the southare dominated by phyllites and slates whichgrade into coarse-grained facies (siltstones,sandstones, and quartzites) northwards. To theeast the unit is represented by carbonaceous

16.4.2 The geology of the Trinity Mine

Stratigraphy

The local stratigraphy of the Trinity Mine andsurrounding area is outlined in Fig. 16.5 andTable 16.3. The ALSG is described by Johnson(1977) as a fine-grained clastic shelf and basinfacies with interbedded turbidites. The localsilicic and calcareous units of the ALSG

Page 410: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 393

TABLE 16.3 Stratigraphy of the Trinity area. (From Bell 1989.)

Age

Quaternary

Pleistocene

Tertiary(early Pliocene)

Cretaceous

Triassic–earlyJurassic

Unit name

Surficial deposits

Volcanics, fluvialdeposits, Upper Tuff

Rhyolite porphyry

Lower tuff

Argillite breccia

Auld Lang Syne Group

Lithologies

Gravel, alluvium, and colluvium

Basalt, fanglomerates and channelsands, phreatic-clastic welded andunwelded, tuff

Rhyolite flows, agglomerates, andbreccias

Epiclastic and pyroclastic airfall,reworked tuff

Subangular to angular argillite clasts(1–10 cm) in fine-grained matrix.Breccia is gradational, tectonic inorigin, and fault associated

Shelf and basin facies. Shale,siltstone, slate, phyllites, andargillites. Siliceous and carbonaceoussandstones and limestones.Quartzites

Intrusions

Latitic and rhyoliticdykes

Rhyolitic, exogenousdomes, porphyritic toaphanitic

Granodiorite, medium-grained, hypidiomorphic(90 Ma)

siltstones and impure limestones (Ashleman1988). The transition between the Tertiaryrhyolite volcanics and the ALSG is marked byan argillite breccia. The breccia is fine-grainedand matrix supported, it contains semiangularto angular argillite clasts and is closely associ-ated with faulting.

Tertiary volcanism resulted in a complexsequence of rhyolitic flows superseded by epi-clastic and pyroclastic deposits. The Tertiaryrhyolites dominate the geology of the Trinityarea, with extensive exposure to the southand flanking the Trinity Range to the north.Latitic and “sugary” rhyolitic dykes disrupt thesequence, with effusive phases and brecciasspatially associated with rhyolite domes. Thevolcanics are capped by Pliocene to Pleistocenefluvial sediments and locally by basaltic flows.Quaternary alluvium infills valleys and flanksthe Trinity Range forming channel fill and allu-vial fan deposits.

Structural geology

Polyphase deformation involving compres-sional tectonics and granodiorite intrusions

resulted in two phases of metamorphism ofthe Mesozoic sediments: (i) low grade regionaland (ii) contact metamorphism. Pre-Tertiaryfolding and faulting of the ALSG resultedin isoclinal folds (Nevadian Orogeny). Tertiarydeformation produced large scale, north-trending open folds and low to high anglefaulting.

Mid Miocene tectonic activity marked theonset of extensional tectonics, resulting inNNE and NE trending, high angle, basin, andrange faulting. This was preceded by N, NNW,and WNW fault sets and localized thrustzones (Ashleman 1988). Faulting appears topostdate the Tertiary rhyolites and predatealteration and mineralisation; however agerelationships may not be so easily defined.Fault displacement and reactivation in con-junction with localized NE thrusts and obliqueENE shear zones (Johnson 1977, Stewart1980) confuses the structural history of thearea. The structural history of the Trinityarea appears to be one of initial NE trendingblock faults and associated thrusts and shearzones, later offset by NW trending normalfaults.

Page 411: Introduction to mineral exploration

394 M.K.G WHATELEY, T. BELL & C.J. MOON

Alteration

The ALSG was not generally receptive toTertiary hydrothermal alteration, except loc-ally along faults and breccia zones. Tertiaryvolcanics were receptive to hydrothermalfluids which resulted in varying degrees ofalteration and hydrofracturing. Rhyolitic tuffsand porphyritic flows have been extensivelyaltered, defining an alteration halo that extends2.5 km beyond the main mineralized zone(Fig. 16.7). Sericitization, silicification andquartz–adularia–sericite (QAS) alteration tendsto be most intense along faults, within per-meable lithologies and breccia zones. Minorpropylitic alteration and kaolinite–illite claysare also reported (Ashleman 1988). QAS altera-

tion is most extensive within the rhyoliticporphyry and tuffs. Silicification is restricted todiscrete W dipping lenticular zones within theupper zone of the deposit, with minor exten-sions along dykes and veins. Silicification isconcentrated within the NE lobe extendingfrom and parallel to NE and ENE faults. PatchyFe staining and limonite is spatially associatedwith high grade Ag mineralisation.

Mineralisation

Silver mineralisation in Pershing County isreported to be hosted within breccias periph-eral to rhyolite domes with mineralisationconcentrated in microfractures (Johnson 1977).

FIG. 16.7 Aerial photograph of Trinity Mine. The light areas in the center represent waste and low grade orestockpiles. The medium tone areas in the foreground show the extent of the alteration in the area. Therectangular area is the heap on which the ore is placed for cyanide leaching. The ponds in which the pregnantsolutions collect are seen next to the mine buildings. (Photograph supplied by US Borax.)

Page 412: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 395

defined by sharp but irregular boundaries, anddips steeply (55–70 degrees) to the W and NW.The SW lobe is disrupted by postmineralisationfaulting and dykes. The oxide orebody defines alinear feature 75 × 610 m (250 × 2000 ft) ex-tending to depths of 90–135 m (300–450 ft)forming lenticular to stacked tabular bodies.

Mineralisation appears to be confined to thenorthwest side of a NE trending fault zone withthe southern boundary defined by a steep, pos-sibly listric, normal fault which brought tuffinto contact with the rhyolitic porphyry. NWtrending offshoots cross the footwall contactinto the rhyolitic tuffs and parallel NW faults.The hanging wall, or northern contact, is notmarked by any significant lithological changeother than a brecciated form of the rhyoliticporphyry at a high stratigraphical level. Thisboundary has been inferred to reflect a changein host rock permeability or a high angle re-verse fault consistent with basin and rangefaulting.

NE trending basin and range faulting offsetby a cogenetic or later, minor, NW fault systemprobably acted as feeders for the invasion ofearly Pliocene hydrothermal fluids. Hydro-fracturing and brittle failure resulted in theformation of a conjugate fracture zone. Hydro-thermal fluids exploited the fracture zone and,in conjunction with alteration, made the hostrock amenable to mineralisation. Mineralisa-tion is concentrated at fault intersections(northwest offshoots) and in a zone parallel butdisplaced from the fault planes. The sharpfootwall contact is either the function of min-eralisation being dissipated by a highly poroustuffaceous rhyolite or ponded by an imperme-able fault gouge. Alternatively, both the sharpnorth and south boundaries may reflect theeffects of fault reactivation, i.e. dextral strikeslip displacing mineralisation along strike andproducing further uplift and erosion. This how-ever is complicated by NW–SE offshoots thatshow no displacement.

16.5 DEVELOPMENT AND MINING

The exploration drilling program defined theore zone on which the block model for the re-serve estimation and the open pit mine planwere based. Drilling continued during mining

However, the Trinity silver orebody is describedas a hydrothermal, volcanic hosted, silver-basemetal (Cu-Pb-Zn-As) deposit (Ashleman 1988).Bonham (1988) describes it as occurring instockwork breccias formed in and adjacent toa rhyolite dome of Miocene age. Both sulfideand oxide ore is present, but high grade zonesamenable to cyanide leaching are restrictedto the oxide zone. The principal sulfide phasesidentified in pan concentrates from rotarydrill cuttings consist of pyrite–marcasite–arsenopyrite–sphalerite–galena with lesseramounts associated with chalcopyrite–pyrrhotite–stannite. They were deposited withsilica along stockwork fractures. The silver-bearing sulfide phases consist of freibergite(silver-rich tetrahedrite) and pyrargyrite.Acanthite and native silver are also presentin trace amounts. Silver in this zone is fine-grained (0.01–0.5 mm), locked within sulfidephases as inclusions and intergrowths, im-posing significant metallurgical problems(section 16.6).

Only the southwest oxide zone, on whichthis chapter concentrates, has been exploited.The Ag mineralogy is essentially bromine-richcerargyrite (Ag(Cl,Br)) with minor amountsof acanthite, often closely associated withlimonite and Fe staining. There are somechloride and sulfide minerals present too.Payable mineralisation is mainly restricted toporphyritic rhyolite flow units (75% of total)as disseminations, fracture infills, veinlets, andreplacements of potassic feldspars. Weakmineralisation occurs within rhyolitic tuffs,breccia, and argillites (25% of total) associatedwith QAS alteration. High grade zones(>10 oz t−1 (>343 g t−1)) exhibit a spatial rela-tionship with areas of strong jointing, sericiticalteration, and limonite (Ashleman 1988).

Base metals (Pb-Zn-Cu-As) exhibit a spatialassociation with the distribution of silver min-eralisation, i.e. a gradual increase to the south.Lead shows the strongest association withsilver in the oxide zone while zinc values tendto be low.

A sharp redox boundary within the oxidezone marks the transition from upper oxide oreinto lower sulfide ore. The oxide orebody issubdivided into a continuous NE lobe andfragmented SW lobe, the former containing thehighest grades. The NE lobe trends NE to ENE

Page 413: Introduction to mineral exploration

396 M.K.G WHATELEY, T. BELL & C.J. MOON

FIG. 16.8 Photograph showing thecolor-coded flags used to delineateore from waste in the pit. The darkflags denote ore blocks and the paleflags denote waste blocks.

to confirm ore reserves. Pit modification wasrequired during mining to improve ore control(section 16.8.4), slope stability, ore shoot ex-cursion, and operational access. The final pitdimensions were 1400 × 500 × 246 ft (427 × 152× 75 m) (Fig. 16.6), with 15-ft (4.5-m) benchesattaining slopes of 60–72 degrees. Pit slopeswere 52 degrees to the west and 43 degrees tothe east, and safety benches were cut every45 ft (14 m). Production haulage ramps were55 ft (16.5 m) wide with 4 ft (1.2 m) berms andhad a maximum gradient of 10%.

Standard mining methods of drilling andblasting, front end loading, and haulage bytruck were carried out by the contractor. A totalof 16,500 t of ore and waste were removed dailyin a 10-hour shift, 5 days a week. Drilling forblasting, using a 55/8 in (143 mm) down-the-hole hammer, was carried out at 15 ft (4.5 m)centers, and at 12 ft (3.7 m) centers in areas ofblocky silicification.

Grade control was based upon the blastholesamples and pit geology. The orebody wasmined principally as a vein type deposit (Perry1989). The eastern boundary (hanging wallcontact) was taken to be a steeply dippingrange front fault and the western boundary wasan assay boundary. Blastholes were sampled,by sample pan cut or manual cone cuts, eachrepresenting a 15 ft (4.5 m) composite. Incontrast to the AAS analytical method usedon the exploration borehole samples, theblasthole samples were analyzed by a cyanide

(CN) leach method, giving a more represent-ative extractable silver content. Statisticaland financial analyses had shown that thesilver grades could be subdivided into fourgroups: (i) ore at >1.3 oz t−1 (45 g t−1); (ii) lowgrade ore at 0.9–1.3 oz t−1 (31–45 g t−1); (iii) leanore at 0.5–0.9 oz t−1 (17–31 g t−1); and (iv) wasteat <0.5 oz t−1 (<17 g t−1). All conversions ofounces per short ton to grammes per tonnehave been made using a factor of 34.285(Berkman 1989). Each bench was divided intoblocks, each block centered on a blasthole. Theassayed grade for each blasthole was assignedto the block. The grades were color coded andeach block was marked on the bench in thepit by the surveyors with the appropriate color-coded flags (Fig. 16.8). This ensured that wastewas correctly identified and sent to the wastepile, while the different grade materials weresent either to the cyanide leach heap or to oneof the two low grade stockpiles.

Sulfide ore was not amenable to heap leach-ing and where encountered was selectivelymined and hauled to low and high grade sulfideore stockpiles (Fig. 16.8) (Perry 1989). Sulfideore was differentiated from oxide ore on thebasis of color (yellow oxide ore and grey sulfideore) and a percentage comparison of total AASassays (or fire assay) and cyanide leachableassays i.e.:

AAS CNAAS

− × 100

Page 414: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 397

a percentage difference of 6–17% defined theoxide ore and the sulfide ore was marked by a24–40% difference.

16.6 MINERAL PROCESSING

Metallurgical testing of core samples deter-mined that cyanide leaching was the most cost-effective method of treating the oxide ore. Anexcellent overview of heap leach technology isgiven by Dorey et al. (1988). Sulfide ore resultedin high cyanide consumption and required finegrinding to give a 78–84% recovery. The oxideore was amenable to direct leaching resultingin a 94–97% recovery. Flotation tests on thesulfide ore liberated 90–95% of the silver and90% of the Pb and Zn providing the pH valuesof the collectors, which suppressed Fe and As,were high. Oxide ore recoveries by flotationwere low (50–60%). Laboratory tests indicatedthat it would be possible to heap leach the ox-ide ore, but the sulfide ore had to be stockpiledwith a view to possible flotation extraction at alater date. A flow sheet of the process is shownin Fig. 16.9.

Mined ore was hauled and dumped on to anore surge pile and fed in to a primary jawcrusher (Fig. 16.9). The −3/4 in (−19 mm) prod-uct passed over a two-deck vibrating screen onto an in-line agglomerator. Secondary crushingof the oversized ore was completed by conecrushers in parallel circuit. The crushed orewas automatically sampled and weighed.

Ten pounds (4.5 kg) of cement was added to1 t of ore prior to agglomeration. The agglomer-ated ore was screened to remove the −3/8 in(−10 mm) fines, and was then spread on theleach pad with a slough stacker. The700 × 1100 ft (210 × 335 m) leach pad was linedwith 60-mm heavy duty polyurethane (HDPE)and divided into six cells. Each cell had a 183 tcapacity when stacked to 41 ft (12.5 m). Pri-mary and secondary leaching utilized drip feedemitters providing a flow rate of 0.005–0.008gallons per minute per square foot (200–350 ml min−1 m−2). Sprinklers were used duringthe final stage of leaching and rinsing.

Pregnant solutions passed into the pregnantpond and were then processed by a Merrill–Crowe Zn precipitator plant with a1000 gallons min−1 (3785 L min−1) capacity.

After initial filtering the precipitate was dried,analyzed, fluxed, and smelted (Fig. 16.9), toproduce silver doré bullion (99% silver). It wasestimated that 75% of the silver in the oxideore was recovered by cyanide leaching.

16.7 ENVIRONMENTAL CONSIDERATIONS

Any mining project which will result in thealteration of the surface of a property, eitherthrough the addition of mine buildings orbecause of mining, is required to submit aPlanning Application (in the UK) or a Plan ofOperation (in the USA) to the authority respon-sible for administering the area. This applica-tion should describe the proposed miningoperation and the infrastructure requirementsand include a schedule of the proposed activ-ities (Thatcher et al. 1988). Often the operatorwill include a plan of the procedures he pro-poses to implement to protect the environmentduring and after mining (reclamation).

In the case of a precious metal heap leachoperation, additional permits and approvals arerequired before a heap leach operation can becommissioned. Thatcher et al. (1988) discussthe various regulatory aspects and permittingrequirements for precious metal heap leachoperations. These requirements include airquality, and surface and groundwater qualitypermits as well as cyanide neutralization re-quirements. In the latter case, Nevada specifiesdetailed heap rinsing procedures and analyticalmethods for cyanide detection as operatingconditions in their water quality or dischargepermits (Thatcher et al. 1988). In Nevada opera-tors are expected to obtain cyanide neutraliza-tion levels of 0.2 mg per litre free CN− as atarget concentration, but most operators areunable to meet these standards. Similarly,there should be only 1 ppm cyanide in the re-sidual weak acid digest (WAD) or <10 ppm CNin the soil. More realistically, they expect thecompany to attempt to meet these criteria andthen show that the cyanide remaining behindwill not affect or put at risk the groundwater.

The Trinity operators had several altern-atives to consider with regard to cyanidedegradation and heap detoxification. These aredescribed in some detail by Smith (1988a),and the various methods are outlined here. The

Page 415: Introduction to mineral exploration

398 M.K.G WHATELEY, T. BELL & C.J. MOON

FIG. 16.9 A diagrammatic flow sheet of the mining and processing of the ore at the Trinity Mine, Nevada.

BLAST HOLE DRILLINGSampled and analyzed

ORE CONTROLBroken ore

defined and flagged

SELECTIVE MINING

SURGE PILEAg oxide ore

STOCK PILEWaste

sulfide oreAg bearing waste

PRIMARY CRUSHER 2 DECK VIBRATINGSCREEN

2 CONE SECONDARYCRUSHER

Cyanide solutionemitters and wobblers

AUTOMATIC SAMPLINGAND WEIGHING

SLOUGH STACKER

AGGLOMERATION2.4 × 8.5 m belt lined

agglomerator45 kg cement ton−1 ore

@ 10% moisture

LEACH PAD22 mm ore pad cover on

60 mm HDPE liner6 cell pad 210 × 335 × 12.5 m

200,000 t capacity

PREGNANT POND

CLARIFIERS

INTERMEDIATE POND BARREN POND

Cyanide

DEAERATOR TOWER FILTER PRESSES

DRYING MERCURYRETORT

FURNACES

Fresh water

Barren solutions

Zinc dustFlux

Filtered silverprecipitate

Waste slag toheap

High grade slagrecycled

Doré to refinery

−19 mm FEED

+19 mm FEED

PRIMARY LEACH SECONDARY LEACH

Page 416: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 399

methods fall into two categories, namely nat-ural processes or chemical treatment.

16.7.1 Natural degradation anddetoxification

1 Passive abandonment. High UV levels andstrong winds for 9 months of the year wouldmake this a possible alternative. The naturaldegradation of the cyanide would be fast in thenear-surface layers, but would slow rapidlywith depth.2 Rinsing with barren solutions. Rinsing toclean the entire heap has been tested empir-ically in the laboratory. A column in the labor-atory was rinsed with fresh water and thecyanide levels were reduced from 700 to50 ppm. Unfortunately this took between 12and 21 months to achieve and extremely largequatities of water (of the order of 4500 L t−1)were required.

16.7.2 Chemical treatment

Five types of chemical treatment are currentlyavailable and were considered by US Borax.1 Acid leach. The leachate thus producedwould be collected in the ponds, neutralizedwith lime, and the metals which precipitatedwould be removed mechanically.2 Sulfur dioxide and air oxidation method.(Patented by INCO). These are added in con-junction with lime and a copper catalyst, butspecial equipment is required and this methodis expensive.3 Alkaline chlorination. Either chlorine gas orcalcium hypochlorite are used in the oxidationprocess but these methods have very highreagent costs, and careful pH control with thesecond method is required to prevent the for-mation of toxic cyanogen chloride.4 Hydrogen peroxide oxidation. This methodis clean and nontoxic, but the H2O2 is expen-sive. This method is used by OK Tedi in PNGwhere H2O2 is added to the recirculation pondand the cyanide-free water is then recirculatedthrough the heap using the original sprinklersystem.5 Ferrous sulfate. Chemical destruction ofthe cyanide is used whereby 1 mole of ferroussulfate ties up 6 moles of cyanide to producePrussian Blue. This reaction takes place in

the pond so that the recirculated water is freefrom cyanide. Both the last two methods arequick acting, especially with the initial flush.Thereafter there is a longer term diffusionmechanism which completes the flushingprocedure.

16.7.3 Summary

There are advantages and disadvantages witheach process as outlined above. The State ofNevada discourages the use of hypochloriteand peroxide neutralization methods (Thatcheret al. 1988), but the INCO SO2–air and the per-oxide processes are thought to be more techni-cally efficient (Smith 1988a). However the finalconclusion may well be that each of the majordetoxification processes in use has its applica-tion in a particular set of circumstances whicheach operator must establish for each property.

16.8 GRADE ESTIMATION

It is essential for reliable ore reserve calcula-tions to derive the best Ag grade estimate for anorebody. The 5220 bench of the Trinity silverdeposit was chosen to illustrate the pro-blems and procedures of grade estimation. The5220 level exhaustive dataset contains 1390blasthole samples. No spatial bias or clusteringeffects are evident due to the regular samplepattern (15 ft (4.5 m) centers).

For ore reserve calculations a global estimateof mean silver is required. The global estimatemust, however, be controlled by local esti-mates because of the spatial variation of silvervalues. Measures of variability are importantin evaluating the accuracy of a grade estimate,especially within a spatial context, and havestrong implications for mine planning.

In many geological environments the prob-lem of grade estimation is intrinsically linkedto the type of statistical distribution. Manyproblems can be solved from an assessment ofthe univariate statistics, histograms, and prob-ability plots (Davis 1986), but this assessmentcan not be divorced from the spatial contextof the data (Isaaks & Srivastava 1989). Thefollowing sections explore the problems ofgrade estimation and derive simple solutionsfor modeling skewed distributions.

Page 417: Introduction to mineral exploration

400 M.K.G WHATELEY, T. BELL & C.J. MOON

16.8.1 Univariate statistics

For the purpose of grade estimation the dis-tribution of assay values within the populationis a prime requirement, especially regardingfrequencies above a specified lower limit, e.g.cut-off grades. Most statistical parameters areapplied to the Gaussian distribution, yetwithin a geological environment a normal dis-tribution is not always immediately evident.It is common to find a large number of smallvalues and a few large values, the lognormaldistribution is then a good alternative (Isaaks &Srivastava 1989). The histogram (Fig. 16.10a)describes the character of the distribution. Theclustering of points at low values and the tailextending toward the high values indicatesthat the assays on the 5220 bench do notconform to a normal distribution. Most of thecumulative frequencies for the lower valuesplot in a relatively straight line, but departuresoccur toward the higher values (Fig. 16.10b).This is further evident from the curved natureof the normal probability plot (Fig. 16.10c).

The departure from a straight line at lowervalues on the log normal probability plot(Fig 16.10d) indicates that the data also do not

fit a log-normal distribution. Although it ispossible to model a “best fit” distribution tothe overall population, errors will occur duringgrade estimation. This is evident from the sum-mary statistics (Table 16.4). It is clear from thefrequency distributions that a global estimatebased upon a normal distribution will be biasedtoward high values (overestimated). Similarlythe log transformed data will be biased by lowvalues (underestimated) (Fig. 16.10d). Due tothe mathematical complications that arisefrom a log transformation, such a transforma-tion is best avoided. It is a preferred alternativeto try to establish Gaussian distributionswithin the exhaustive dataset in order to im-prove grade estimation.

The important features of the distributionare captured by the univariate statistics (Table16.4) and illustrate immediately the problem ofgrade estimation for the 5220 bench. The sum-mary statistics provide measures of location,spread, and shape. The mean Ag content of thebench is recorded as 1.43 oz t−1 (49 g t−1), yetother estimates of central tendency, themedian (0.33) and the trimmed mean (0.48), in-voke caution in using the mean as an accurateglobal estimate. The high variance and stand-ard deviation describe the strong variability ofthe data values. The implication is that insuffi-cient confidence can be placed upon the meanas an accurate global estimate. A strong posit-ive skew with a long tail of high values to theright is evident from the high positive skew-ness (+10.62). The degree of asymmetry is alsosupported by the coefficient of variance (2.58).Knudsen (1988) implies that if the coefficientof variance is greater than 1.2 a log-normaldistribution could be modeled. The Sichel-Testimator can be used to estimate the mean ofa log normal distribution, and is calculated tobe 1.28.

The frequency histogram (Fig. 16.10a) illus-trates how the data are proportioned. It isimportant to note that even though the datarange is 0–85 oz t−1 Ag, only 5% is > 7.2, 2%> 12.5, and 0.5% > 20 oz t−1 Ag. The upper quar-tile indicates that 75% of the data lie between arestricted range of 0–1.1 oz t−1 Ag, and the meanwhich is higher (1.43) does not reflect themajority of the data. It is obvious therefore thatthe magnitude of the positive skew will have adisproportionate effect upon the mean.

40

30

20

10

0 4 8 12 160

20

40

60

80

100

0 4 8 12 16

99

90

50

10

1

0 4 8 12 16 0.1 1 10 100

1

10

50

90

99

Cum

ulat

ive

freq

uenc

y %

Cum

ulat

ive

freq

uenc

y %

(a)

(c) (d)

(b)

Fre

quen

cy %

Cum

ulat

ive

freq

uenc

y %

FIG. 16.10 (a) Histogram of (cyanide leach) silvervalues (oz t−1) obtained from blasthole sampleson the 5220 bench, Trinity Mine, Nevada. (b)Cumulative frequency plot. (c) Normal probabilityplot. (d) Log-normal probability plot.

Page 418: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 401

Total Excluding 1.3–20population >>>>>20 oz t−−−−−1 Ag oz t−−−−−1 Ag

Number of assays 1390 1385 309Mean 1.43 (49) 1.30 (45) 4.65 (159)Median 0.33 (11) 0.33 (11) 2.30 ( 79)Trimean 0.48 (16) 0.48 (16) 3.45 (118)Sichel-T 1.28 1.22 4.52Variance 13.59 6.93 16.36Standard deviation 3.69 2.63 4.04Standard error 0.10 0.07 0.36Skewness 10.62 3.89 1.84Coefficient of variance 2.58 2.02 0.87Lower quartile 0.16 ( 5) 0.16 ( 5) 1.90 ( 65)Upper quartile 1.11 (38) 1.09 (37) 5.84 (200)Interquartile range 0.95 (33) 0.93 (32) 3.93 (135)

Silver values are in oz t−1 (g t−1).

TABLE 16.4 Univariate statisticsof the silver values derived fromcyanide leach analysis of the blasthole samples on the 5220 level,Trinity Mine, Nevada.

The initial statistical evaluation illustratesa potential error in the global estimate result-ing in possible overestimation and bias towarderratic high values. More detail is requiredgiven that the mean is only a robust estimateif applied to a Gaussian distribution and a logtransformation is not adopted. It is importantat this stage to investigate the distributionfurther. Having ascertained that the data donot conform to a normal or log-normal dis-tribution, the next step is to evaluate potentialmodality and departures from a continuous distribution.

The probability plot and histogram are veryuseful in checking for multiple populations.Although breaks in the graphs do not alwaysimply multiple populations, they representchanges in the character of the distributionover different class intervals. The initial step isto explode the lower end of the histogram bymaking the class interval smaller (Fig. 16.11)and searching for potential breaks in thepopulation. Breaks in the distribution arechosen on the basis of significant changes inthe frequency between class limits or repeatedpatterns reflecting polymodality. If subpopula-tions are found, explanations must be soughtand may depend upon sample support, geolo-gical control, or population mixing. The pos-itively skewed distribution of the data may bea function of the overprinting of multiplepopulations. Defining subgroups which may

0

Freq

uenc

y %

oz

0

10

20

1 2 3 4

FIG. 16.11 Histogram of (cyanide leach) silver values(oz t−1) obtained from blasthole samples on the 5220bench, Trinity Mine, Nevada, with class intervalsreduced to 0.05 oz t−1. The higher values (>5 oz t−1)are omitted for clarity.

approximate normal distributions couldimprove the global estimate.

Such an evaluation reveals a number ofbreaks within the population of silver gradeon the 5220 level. The following class breakshave been interpreted: 0.4, 0.8, 1.3, 5, 10, and20 oz t−1 Ag (14, 27, 45, 171, 343, and 686 g t−1).Those values greater than 20 oz t−1 are

Page 419: Introduction to mineral exploration

402 M.K.G WHATELEY, T. BELL & C.J. MOON

TABLE 16.5 Univariate statistics of the subpopulations of the silver values derived from cyanide leach analysisof the blasthole samples on the 5220 level, Trinity Mine, Nevada.

Silver values in oz t −−−−−1 (g t −−−−−1)

0.0–0.4 0.4–0.8 0.8–1.3 1.3–5 5–10 10–20

Number of assays 762 212 1032 215 731Mean 0.18 (6) 0.57 (20) 1.03 (35) 2.56 (88) 7.46 (256) 14.37 (493)Median 0.17 (6) 0.55 (19) 1.01 (35) 2.25 (77) 7.34 (252) 13.60 (466)Trimean 0.17 (6) 0.56 (19) 1.02 (35) 2.35 (81) 7.44 (255) 13.93 (478)Sichel-T 0.20 0.57 1.03 2.56 7.46 14.37Variance 0.01 0.01 0.02 1.04 2.48 10.79Standard deviation 0.10 0.11 0.15 1.02 1.57 3.28Standard error 0.00 0.01 0.02 0.07 0.21 0.58Skewness 0.19 0.28 0.19 0.74 0.08 0.25Coefficient of variance 0.55 0.19 0.15 0.40 0.21 0.23Lower quartile 0.10 (3) 0.47 (16) 0.89 (31) 1.71 (59) 5.98 (205) 11.18 (383)Upper quartile 0.26 (9) 0.67 (23) 1.16 (40) 3.20 (110) 9.09 (312) 16.70 (573)Interquartile range 0.16 (6) 0.20 (7) 0.27 (9) 1.49 (51) 3.11 (107) 5.52 (189)

considered outliers (5 out of 1390 samples) andare not evaluated. To assess the statistical sig-nificance of the subgroups univariate statisticsare calculated for each group (Table 16.5). Inall instances normal distributions are appro-ximated with significant reductions in thevariance, coefficient of variance, and skewness.Estimates of central tendency lie within statis-tically acceptable limits indicating improvedconfidence in the mean of each group. The 1.3–5 oz t−1 group is slightly skewed and a furthersplit possible. However, improved detail maynot result in enhanced accuracy or confidencein the mean.

The exhaustive dataset can be split into sub-groups according to grade classes and the gradeestimate for each group is improved. The statis-tical and financial evaluation undertaken byUS Borax resulted in a similar grouping (section16.5). US Borax applied a 1.3 oz t−1 cut-off to thedata such that everything above 1.3 oz t−1 wasmined as ore. It is important to evaluate theeffect of cut-off on the global estimate and theimplications in relation to the subgroups.

16.8.2 Outlier and cut-off grade evaluation

Those values in excess of 20 oz t−1 (686 g t−1) Aghave been classed as anomalously high values,and if rejected have a significant effect upon the

univariate statistics and grade estimate (Table16.4) despite only representing 0.5% of thedata. The mean is reduced by 10% to 1.3 oz t−1

(45 g t−1) Ag while the variance and skewnessare reduced by up to half. However the distribu-tion is still strongly skewed and the problemsof estimation still remain. If the 1.3 oz t−1 cut-off is then also applied the distribution be-comes flatter, the global estimate more precise,but accuracy is reduced because of increasedvariability (Table 16.4). The 1.3 oz t−1 cut-offalso reflects a major break in the populationand improved estimates may be a function of aseparate population approximating to a normaldistribution. With a reduced skewness the vari-ability of values become more symmetricalabout the mean.

A further investigation of outlier removal(Fig. 16.12) shows that as the tail effect of thedistribution is reduced the mean grade andstandard deviation decreases in a linear fashion.The decline is more gradual for the total popu-lation than when the 1.3 g t−1 cut-off is applied.It is obvious that as values are included orexcluded the magnitude of the mean varies.This must be viewed in relation to the numberof samples, the effect of which is evident withvalues above or below 20 oz t−1 Ag. Althoughthe magnitude of the grade estimate can beincreased by including outliers the rate of

Page 420: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 403

Mean Ag (>1.3 oz t−1 / 45 g t−1)

SD (>1.3 oz t−1 / 45 g t−1)

Mean (Total)

SD (Total)

100 200

Ag

oz t

−1

0

1

2

3

4

5

6

7

0

Outlier cut-off grade (Ag oz t−1)

FIG. 16.12 Graph showing changes in mean grade andstandard deviation when different outlying (high)grades are removed from the database of the silvervalues from the 5220 bench, Trinity Mine, Nevada.

improvement in the ore reserve estimation isreduced. A grade estimate must also evaluatehow the values fluctuate around the mean,therefore the relationship of the standarddeviation to the mean is important.

Within a mining environment the con-fidence of the grade estimate is derived fromthe standard deviation. The value of the stand-ard deviation, however, is subject to the groupof values from which the estimate wasobtained and the spatial context. The standarddeviation expresses confidence in symmetricalintervals, therefore when the standard devia-tion exceeds the mean the magnitude of under-estimates is significantly different from themagnitude of overestimates. This is illustratedin Fig. 16.12 for the total population such thatthe global estimate would exceed the localestimate in the majority of cases. The degreeof variability and error increase in magnitudeas the tail effect of the distribution is moreprominent, therefore the above effect becomesmore pronounced.

The relationship of the standard deviation tothe mean changes when the same evaluation isapplied to ore grade material. The graph (Fig.16.12) indicates that the estimate for the oregrade material is more reliable and subject toless fluctuation, although as the distribution

(a)

(b)

Mean Ag

SD

Mean Ag

SD

Cut-off grade (Ag oz t − 1)

Ag

oz t

−1

1

2

3

4

5

6

7

8

9

10

0 0.5 1 1.5 2 2.5 3 3.5 4

Ag

oz t

−1

1

2

3

4

5

6

7

8

9

FIG. 16.13 Graphs showing the average silver gradewhen varying cut-offs are applied to the databaseof the silver values for the 5220 bench, TrinityMine, Nevada: (a) total population; (b) with values>20 oz t−1 removed.

expands, variability increases. The grade cut-off curves reinforce the relationship of themean to the standard deviation for the totalpopulation (Fig. 16.13a) and the effect of remov-ing the 20 oz t−1 outliers (Fig. 16.13b). Asextreme values of the distribution are alteredthe degree of variability declines. A point isachieved where the number of local underesti-mates and overestimates as compared with theglobal estimate balance each other. As low orhigh values are included or removed the meanchanges accordingly and when an approximatenormal distribution is modeled the variabilityreaches a constant (Fig. 16.13b). At this pointthere is no improvement in the confidence ofthe grade estimate and, in terms of reserve

Page 421: Introduction to mineral exploration

404 M.K.G WHATELEY, T. BELL & C.J. MOON

FIG. 16.14 A contour map of the silver grades (oz t−1) on the 5220 bench, Trinity Mine. Contours are codedaccording to grade classes.

estimation, it is the spatial distribution ofsilver that becomes important.

16.8.3 Spatial distribution

Although it is possible to improve grade esti-mates by evaluating the statistical distribu-tion, the spatial features of the dataset areimportant. A spatial evaluation will focus uponthe location of high values, zoning, trends, andcontinuity. A contour map of the 5220 level(Fig. 16.14) shows an E–W trend to the datawhich fragments and changes to a NW–SEtrend to the west. Zoning occurs and discretepods of ore grade material are evident. Lowgrade values less than 0.8 oz t−1 (27 g t−1) covera greater area. The closeness of the contoursespecially at the 1.3 oz t−1 Ag (45 g t−1) contourindicates a steep gradient from high to low val-ues. The abrupt change from waste to ore gradematerial marked by a steep gradient will bereflected by a high local variance.

It is obvious that the subgroups previouslydefined are not randomly distributed but con-form to a zonal arrangement. The indicatormaps (Fig. 16.15) support the spatial integrityof the subgroups, illustrating the concentriczoning and isolated groupings of the data. Thestatistics of each subgroup can then be appliedto specific areas of the 5220 level and local

estimates can be applied to the contouredzones. A further assessment of local estimatescan be achieved by using moving windows(Isaaks & Srivastava 1989, Hatton 1994a).Statistics were calculated for 100 ft (30 m)square windows overlapping by 50 ft (15 m)(Fig. 16.16). The contour plots of the mean andstandard deviation for each window (Fig. 16.17)show that the average silver values and vari-ability change locally across the area. Zonesof erratic ore grades can be located and flaggedfor the purpose of mine planning and gradecontrol.

In general the change in variability reflectsthe change in the mean, although at the edge ofthe high grade zone to the east, the variabilityincreases at a greater rate. High variability isa function of the mixing of two populations orthe transition from one population or zone toanother. This is seen in the profile taken alongN1300 (Fig. 16.18a). A profile passing throughthe high grade zone (N1200, Fig. 16.18b) showsa reduced change in variability compared witha pronounced increase in the magnitude of themean. The strong relationship between thelocal mean and variability (Fig. 16.18) is re-ferred to as a proportional effect which impliesthat the variability is predictable although thedata do not reflect a normal distribution. Thecontour plots (Fig. 16.17) show that in certain

1500

1400

1300

1200

1100

0 250 500 750 1000

1.3

0.80.4

5.0

0.8

1.3

0.4

0.4

200.8

0.8

5.0

0.81.3

1.3

1.3

1.3

0.40.8

0.40.8

1.30.4

20105.0 10

2010

1.8

0.4

0.4

0.4

0.8

1.3

5.0

0.8

0.4

1.3

20

0.4

0.4

Pit outline

Page 422: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 405

(a)

(b)

(d)

(e)

(c)

11

12

13

14

15

16

×10

2

0 250 500 750 1000

11

12

13

14

15

16

×10

2

0 250 500 750 1000

11

12

13

14

15

16

×10

2

0 250 500 750 1000

11

12

13

14

15

×10

2

0 250 500 750 1000

16

11

12

13

14

15

16

×10

2

0 250 500 750 1000

FIG. 16.15 Indicator maps of the Ag grades (oz t−1) of the 5220 bench demonstrating the spatial integrity of thedata. The grade boundaries are at (a) 0.4, (b) 0.8, (c) 1.3, (d) 5.0 and (e) 10.0 oz t−1 (14, 27, 45, 171, and 343 g t−1).

13 to 8 oz Ag2. This again demonstrates theeffect that the few high values have on thevariability.

Directional semi-variograms show differentranges in different directions (Fig. 16.19a–d).Continuity was greater in the E–W direction(300 ft) (Fig. 16.19c) compared with all otherdirections. The minimum continuity was 90 ftin a N–S direction (Fig. 16.19a). An E–W elipsemeasuring 300 × 90 ft (Fig. 16.19f) would be theoptimal moving window shape and size to givethe lowest variability for a local estimate.

The NW–SE directional semi-variogram(Fig. 16.19d) shows a “hole effect” (Journel& Huijbregts 1978, Clark 1979) which resultsfrom alternate comparison of high and lowzones in the NW half of the deposit.

locations the change in local variance mirrorsor is greater than the change in the mean (theproportional effect). However elsewhere thevariability remains roughly constant whilstthe local mean fluctuates.

16.8.4 Semi-variograms

Semi-variograms are used to quantify thespatial continuity of the data (see Chapter 9).The range (a) of the semi-variogram definesa radius around which the local estimate hasthe least variability. The omnidirectionalsemi-variogram of the 5220 bench (Fig. 16.19a)shows a range of 120 ft (35 m). When sampleslarger than 20 oz t−1 (686 g t−1) were excluded(Fig. 16.19b) the sill was reduced by a third from

Page 423: Introduction to mineral exploration

406 M.K.G WHATELEY, T. BELL & C.J. MOON

(a)

(b)

0.1 0.1 0.2 0.4 0.3 0.1 0.1 0.5

3.6 2.3 1.8 1.1 0.3 0.2 0.4 0.5 0.5

8.1 7.1 7.3 5.2 3.6 2.8 1.5 1.9 1.9

3.4 5.7 6.6 6.3 7.7 0.1 3.0 3.3 2.2

1.92.22.03.13.82.11.31.21.3

0.2 0.2 0.3 0.3 0.2 0.2 0.6 1.0 2.8

0.1 0.1

0.2 0.2 0.1 0.1

0.2 0.3 0.9 2.2

0.8 0.6 1.4 4.7

1.4 0.9 1.0 3.3

1.10.60.50.6

0.1 0.1 0.2 0.2

0.3 0.4 0.3 0.1 0.0 0.0

0.2 0.8 0.9 0.2 0.1 0.0

0.4 1.4 1.5 0.6 0.2 0.1 0.10.30.3

0.5 1.2 1.5 1.0 0.3 0.3 0.30.40.4 0.2

0.5 0.6 1.0 0.9 0.6 1.0 1.4 0.8

0.6 0.9 1.0 0.5 0.7 1.4 1.6 0.9

0.5 0.6 0.5 0.9

0.4 0.4 0.3 1.0

0.40.3

3.3 0.1

0.1 0.4 0.2 0.1 0.0 0.0

0.1 2.6 2.6 0.3 0.0 0.0

0.2 0.2 0.6 2.5 2.5 0.6 0.2 0.1 0.2

0.5 0.5 0.5 0.9 0.9 0.9 0.3 0.4 0.4 0.2 0.1 0.0

0.3 0.4 0.9 0.9 0.5 0.9 2.1 2.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.7 0.6 0.1 0.1 0.3

0.3 0.8 0.9 0.5 1.0 1.9 1.8 0.1 0.3 1.1 4.1 6.6 4.3 3.8 3.1 0.5 0.2 0.3 0.3 0.3

0.5 0.6 0.4 1.2 1.2 0.5 1.1 4.4 6.7 5.4 5.6 5.2 5.5 4.8 1.8 2.6 2.8

0.3 0.3 0.2 1.4 1.4 0.8 0.8 0.8 4.7 5.0 5.1 4.2 5.5 5.5 2.0 2.5 2.8

0.2 1.00.8 0.7 0.5 1.4 1.9 1.9 1.7 3.1 5.2 4.6 1.7 1.6 4.2

16.4 0.1 0.0 0.1 0.2 0.1 0.3 0.1 0.1 0.2 0.2 0.2 0.5 0.4 6.6

0.7

1100

1200

1300

1400

1500

1600

0 250 500 750 1000

1100

1200

1300

1400

1500

1600

0 250 500 750 1000

FIG. 16.16 Window statistics performed on the silver grades of the 5220 bench, based on a transformed gridoriented E–W with 30 m square, moving window overlapping by 15 m: (a) mean Ag oz t−1; (b) standarddeviation.

Page 424: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 407

1400

1300

1200

1100

Fig 16.18a

Fig 16.18b

1.0

0.80.

60.4

0.60.81.0

1.0

2.0

1.0 0.4

0.6

0.8

0.6

0.4

1.02.03.04.0

0.8

6.0

6.0

3.0

3.04.0

1.0

2.0

Pit outline(a)

1400

1300

1200

11000 250 500 750 1000

Fig 16.18b

Pit outline(b)

Fig 16.18a

2.0

1.00.80.60.4

1.00.6

0.41.02.0

8.0

1.0

0.8

0.60.4

2.0

6.0

4.03.02.0

1.00.8

2.0

2.0 3.0

4.01.0

0.40.61.0

FIG. 16.17 Window statistics performed on the silver grades of the 5220 bench, based on a transformed gridoriented E–W with 30 m square, moving window overlapping by 15 m: (a) contours of mean Ag oz t−1,(b) contours of standard deviation. (From Bell & Whateley 1994.)

(a)Mean Ag

Standard deviation

100

Ag

oz t

−1

Distance (m)

0

3

4

5

6

300 500 700 900 1100

1

2

FIG. 16.18 Profiles 1 and 2 (a and b respectively) of the window mean and standard deviation showing theproportional effect and highlighting areas of expected high error. (From Bell & Whateley 1994.)

(b)

100

Ag

oz t

−1

Distance (m)

0

3

4

5

6

300 500 700 900 1100

1

2

7 Mean Ag

Standard deviation

Page 425: Introduction to mineral exploration

408 M.K.G WHATELEY, T. BELL & C.J. MOON

(a)

0 100 200 300Distance (m)

400 500

C0 = 0.5C = 12.0a = 110

600

131110

98654310

Gam

ma

(h)

(d)

0 100 200 300Distance (m)

400 500

C0 = 1.0C = 10.3a = 80

600

1110

987653210

Gam

ma

(h)

(b)

0 100 200 300Distance (m)

400 500

C0 = 1.3C = 11.8a = 310

600

131211

98754310

Gam

ma

(h)

(c) (f )

0 100 200 300Distance (m)

400 500

C0 = 1.0C = 8.1a = 130

600

11987654310

Gam

ma

(h)

(e)

0 100

Anisotropy factors1.4 3.9

110 m310 m

130 m

80 m

Ellipse of anisotropy

1.6

1.0

200 300Distance (m)

400 500

C0 = 1.5C = 8.2a = 110

600

1110

876543210

Gam

ma

(h)

FIG. 16.19 Semi-variograms (spherical scheme models) for the silver assays of the 5220 bench: (a) N–Sdirection; (b) NE–SW direction; (c) E–W direction; (d) SE–NW direction; (e) omni-directional; (f) ellipse ofanisotropy and anisotropy factors. (From Bell & Whateley 1994.)

understand what constitutes the population,in order that sensible subdivisions into sub-populations can be made. It is incorrect toevaluate an exhaustive dataset if the data arederived from different episodes of mineralisa-tion, when their physical and chemical charac-teristics differ. If the global estimate is used tomake local estimates of grade, then the grade

16.8.5 Conclusions

Attempting a global estimate of the grade forthe entire bench is irrelevant in a mining situa-tion because geologists and mining engineersare only interested in ore grade material. Theskewed population means that ore grade ma-terial will be underestimated. It is essential to

Page 426: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 409

will be overestimated, because no account ismade of the spatial relationship of the samplesor of the high variability.

To improve the grade estimation it is neces-sary to try to establish normal distributionswithin the whole skewed population, usinggeology, structure, mineralisation, alteration,etc., rather than transforming the whole popu-lation. The variability within each subpopu-lation is thus reduced, which increases theconfidence in the subpopulation mean.

The high values cause extreme variabilitywhich influences the mean. By excluding themthe mean and variance are reduced and theestimation becomes more reliable.

By applying these techniques to the dataabove cut-off grade (omitting outliers), a popu-lation which approximates a normal distribu-tion is achieved. This improves the confidencein the estimation of the mean. By subdividingthe data above cut-off into subpopulations,variability is further reduced.

Indicator maps show that subgroups, asdefined by statistical parameters, are spatiallyarranged. High local variance is a function ofboundaries between these populations. Appli-cation of the moving window statistics shouldtherefore be viewed in terms of the spatiallydistributed groups, trying to avoid crossingsubpopulation boundaries.

From the univariate statistics and spatialdistribution of the data the conclusion is thata global estimate must be derived from localestimates rather than the whole population.

In hindsight, the best ore reserve estimationcould be derived using the average of the esti-mates from discrete zones or the average of theestimates derived from the areas between con-tours of the grades. Bell and Whateley (1994)undertook global estimation of the grade byapplying a number of estimation techniquesto the original exploration borehole data of the5220 level. They compared the results withthe blasthole dataset and concluded that linearinterpolation provided grade estimates withthe lowest variance.

16.9 ORE RESERVE ESTIMATION

During the life of the project, from the explora-tion stages to the feasibility and mine designstudies, different ore reserve calculations were

carried out, each with a different confidencelevel. The confidence in the reserve estimategrew as the amount of borehole data increased.

16.9.1 Evaluation of initial exploration data

Immediately after exploration was initiated inearly 1982, it was realized that a significanttonnage of disseminated silver existed. Duringthe 1982–83 exploration program the oxidemineralisation had not been fully evaluated,and the initial reserve estimates only took thesulfide mineralisation into account.

16.9.2 Evaluation of additional explorationdata

Additional exploration drilling took place in1986 with the intention of identifying addi-tional sulfide mineralisation, and to collectsamples for metallurgical and engineeringstudies. In mid 1986 it became apparent thata small area of high grade silver oxide miner-alisation could possibly support a heap leachoperation. The reserve potential of the oxidezone was calculated by several members ofthe joint venture, using different methods andcut-off grades. Table 16.6 summarizes theresults.

The tonnage calculations vary from 0.932 to1.304 Mt where a Ag cut-off grade of 1.5 oz t−1

(50 g t−1) was used. Where the cut-off grade wasraised to 2 and 3 oz t−1 (69 and 103 g t−1), as onewould expect, the tonnage was reduced but theaverage grade increased. These differences intonnage and average grade were accounted forwhen thin, lower grade intersections used inearlier calculations were omitted, and modifiedpolygons were used. No reference was made tothe geology of the deposit in any of the abovecalculations.

This demonstrates the difficulty in calculat-ing reserve estimates in the exploration phasesof a program (Bell & Whateley 1994). Differentmethods, by different people using differingcut-off grades, produces a variety of estimates.It is not possible to quantify the error of estima-tion using these 2D manual methods of reserveestimation. It is not even practical to comparethe early reserve estimates with the final ton-nage and grade extracted, because in the finalestimate a cut-off grade of 1.3 oz t−1 (45 g t−1)was used.

Page 427: Introduction to mineral exploration

410 M.K.G WHATELEY, T. BELL & C.J. MOON

TABLE 16.6 Summary of the ore reserve calculation for the silver oxide zone, Trinity Mine, Nevada.

Method Tonnage Composite Ag cut-off Tonnage Average Total Agfactor length grade ××××× 106 Ag grade oz (g) ××××× 106

ft3 t−−−−−1 (SG) ft (m) oz t−−−−−1 (g t−−−−−1) oz t−−−−−1 (g t−−−−−1)

Polygons–USB 13.3 (2.41) 20 (6.1) 1.5 (50) 0.967 6.95 (238) 6.22 (213)N–S cross-sections, USB 13.3 (2.41) 10 (3.0) 1.5 (50) 1.304 6.16 (211) 8.03 (275)E–W cross-sections, USB 13.3 (2.41) 10 (3.0) 1.5 (50) 1.293 5.90 (200) 7.63 (262)N–S cross-sections, USB 13.3 (2.41) 10 (3.0) 1.5 (50) 0.932 7.69 (264) 7.17 (246)Polygons, Santa Fe 13.3 (2.41) 20 (6.1) 3.0 (100) 0.669 9.10 (310) 6.09 (209)N–S cross sections, 13.3 (2.41) 10 (3.0) 2.0 (69) 0.870 8.00 (274) 6.96 (240)Santa Fe and USB

TABLE 16.7 Reserve estimates on the silver oxide zone, using the 1/D2 block model, Trinity Mine, Nevada.

Tonnage Tonnage Composite Ag cut-off Average××××× 106 factor length ft (m) grade oz t−−−−−1 Ag grade

ft3 t−−−−−1 (SG) (g t−−−−−1) oz t−−−−−1 (g t−−−−−1)

2.65 13.7 (2.34) 15 (4.5) 0.5 (17) 3.05 (105)2.05 13.7 (2.34) 15 (4.5) 1.0 (34) 3.76 (129)1.65 13.7 (2.34) 15 (4.5) 1.5 (50) 4.46 (153)1.26 13.7 (2.34) 15 (4.5) 2.0 (69) 5.19 (178)0.88 13.7 (2.34) 15 (4.5) 3.0 (100) 6.33 (217)0.63 13.7 (2.34) 15 (4.5) 4.0 (137) 7.52 (258)

16.9.3 Reserve estimate for mine planning

In 1987 a computerized 3D block modelingtechnique was used to estimate reserves andthe data thus generated were used for mineplanning. A conventional 2D inverse distancesquared (1/D2, see section 9.5.1) technique wasused to interpolate grades from borehole infor-mation to the block centers, using a 150 ft(46 m) search radius (Baele 1987). The blockswere given a 25 × 25 ft (7.6 × 7.6 m) area andwere 15 ft (4.6 m) deep. Borehole assays werecollected at 5 ft (1.5 m) intervals, but werecomposited by a weighted average techniqueover 15 ft (4.6 m) to conform to the blockmodel. Geology and alteration were incorpor-ated into the block model by assigning a rockand alteration code to each block. Tonnage andgrade was estimated at various cut-off grades(Table 16.7).

Despite the geological control, this methodof reserve estimation appears to have overesti-mated the reserve potential (in comparison

with the early ore reserve estimates). This maybe partially a result of an edge effect where themajority of any given block at the edge of thedeposit may lie outside the mineralized area,but because of the shape of the boundary theyhave been included in the reserve estimate. Anadditional factor is the use of 13.7 cu ft t−1

(2.34 t m−3) as the ore and waste specific grav-ity, to calculate in situ tonnage comparedwith only 13.3 cu ft t−1 (2.41 t−1 m−3) used in theexploration evaluation.

16.9.4 Updated tonnage and grade estimate

The grade and tonnage of the ore mined fromthe first three benches was significantly differ-ent from that predicted by the 1/D2 blockmodel. An updated reserve estimate washand-calculated using cross-sections to evalu-ate the discrepancy. The geology and structurefrom the mined-out area were plotted on thecross-sections and used to project the orebody

Page 428: Introduction to mineral exploration

16: TRINITY SILVER MINE, NEVADA 411

TABLE 16.8 Updated reserve estimate on the silver oxide zone, Trinity Mine, Nevada.

Method Tonnage Composite Ag cut-off Tonnage Average Totalfactor length ft (m) grade ××××× 106 Ag grade Ag oz (g)ft3 t−−−−−1 (SG) oz t−−−−−1 (g t−−−−−1) oz t −−−−−1 (g t−−−−−1) ××××× 106

N–S cross-sections 13.7 (2.34) 15 (4.5) 1.5 (50) 0.899 6.94 (238) 6.24 (214)Bench plans* 13.7 (2.34) 15 (4.5) 1.6 (55) 0.859 7.33 (251) 6.29 (216)Blocks 13.7 (2.54) 15 (4.5) 1.6 (55) 0.963 6.55 (225) 6.31 (217)15 × 15 × 15 ft(4.5 × 4.5 × 4.5 m)

* These estimates exclude the first three benches which had already been mined out.

geometry downwards (Reim 1988). Reim feltthat the unsatisfactory grade control given bythe 1/D2 method could be accounted for bythe inadequate structural control in the 1/D2

model, variability in grade, and paucity ofsampling data within the southwest portionof the oxide ore body. Additional drillingincreased the sample density in poorly repres-ented areas and this helped define the orebodygeometry. The reserve estimate calculatedusing the updated cross-sections is shown inTable 16.8.

For mine planning purposes a mineable re-serve was calculated on a bench by bench basis.Mineralized zones were transferred from thecross-sections on to 15 ft (4.5 m) bench plans.Polygons were constructed around eachborehole and a reserve calculated (Table 16.8).

16.9.5 Grade control and ore reserveestimation using blastholes

Mine grade control was exercised by samplingand assaying the blasthole cuttings. Minedblocks were 15 × 15 × 15 ft (4.6 × 4.6 × 4.6 m)and only blocks >1.3 oz t−1 (45 g t−1) were minedas ore. A final reserve was calculated usingthese blocks and a slightly higher reserve witha lower grade was estimated (Table 16.8).

16.9.6 Factors affecting ore reservecalculations

A number of factors influenced the validity ofore reserve estimates, namely geological con-trol, sampling method, cut-off grade, samplemean, analytical technique, specific gravityand dilution factors:

1 Geological control. It is most important toestablish the geological and structural controlof mineralisation as early in a project as pos-sible. In the example used here, the 1/D2 blockmodel had good geological control, but thestructural control was lacking. This resulted inoverestimation of reserves.2 Sampling methods. It is important to controlthe sampling method to ensure consistentand reliable sampling, e.g. down-the-variationin silver values derived from either reversecirculation or percussion drilling may showdisparity.3 Cut-off grade. This varied between estima-tion methods from 3.0 oz t−1 (103 g t−1) in theexploration assessment to 1.3 oz t−1 (45 g t−1)during mining (Tables 16.6–16.8). Compari-sons of reserve estimates using different cut-offgrades can be achieved if grade–tonnage graphsare constructed. These are difficult to con-struct during the exploration phase as data aresparse. Normally economic criteria are usedto establish the cut-off grade (Lane 1988). USBorax assumed a commodity price of $6.50 peroz and expected silver recoveries between 65and 79% (amongst other criteria) to establishthe 1.3 oz t−1 (45 g t−1) cut-off grade used duringmining.4 Sample mean. The mean grade is dependentupon the method used to estimate it (section16.8.2). The arithmetic mean probably over-estimated the mean in this study but thiscould have been improved by establishingsubpopulations with normal distibutions andaveraging the estimates from these discretezones.5 Analytical technique. In order to comparesamples used to estimate reserves using

Page 429: Introduction to mineral exploration

412 M.K.G WHATELEY, T. BELL & C.J. MOON

exploration boreholes and blastholes, a simplelinear regression of cyanide-extractable silveron total recoverable silver was performed. Acorrelation coefficient of 0.94 indicated astrong positive relationship. Above 10 oz t−1

(343 g t−1) Ag, the conversion of total silver tocyanide-eachable silver was less accurate.Changing analytical procedure during an ex-ploration or feasibility exercise, without under-taking an overlapping series of analyses whichcan be used for correlative purposes, introducesan unnecessary risk of error.6 Specific gravity. Initially, densities of 13.0and 13.3 ft3 t−1 (2.34 and 2.41 g cm−3) were usedto calculate tonnages. Density was determinedon a number of different rock types andan average of 13.7 ft3 t−1 (2.34 g t−1) was calcu-lated, which was used for the final tonnagecalculations.7 Mining dilution. Dilution of ore by wasteduring mining is inevitable but careful gradecontrol will minimize the risk of dilution.Dilution usually occurs in ore deposits wheregrade is highly variable, as at the Trinity Mine.The assumption that a blasthole in the centerof a 15 × 15 × 15 ft block (4.5 × 4.5 × 4.5 m) isrepresentative of that block is not necessarilytrue. The arbitrary boundary between a blockshowing a grade less than cut-off and a highgrade block is drawn half way between the twoholes. The low grade ore may extend fartherinto the high grade block than one expects,thus adding to the mining dilution problem. US

Borax initially expected a 20% dilution, butthis was later modified to 7% after furtherassessment during mining.

16.10 SUMMARY AND CONCLUSIONS

The Trinity oxide orebody defined a small buteconomic silver deposit, amenable to cyanideheap leaching. The orebody was fairly continu-ous to the northeast but became fragmentedand fault controlled to the southwest. The orezone was defined by stacked tabular to lenticu-lar bodies dipping steeply to the west andnorthwest and offset by N–S and NW–SEcross-faulting. Mineralisation tended to becontrolled by faulting, alteration, and fracturedensity. The sharp ore grade boundaries alongthe footwall and hanging wall suggest geo-logical control. These boundaries may reflectchange in host rock permeability or sharp faultcontacts.

The Trinity deposit illustrates the estima-tion problems encountered in deposits whichhave highly skewed data. The problem iscompounded when trying to estimate gradeand resource potential from a relatively lownumber of exploratory boreholes. Recognizingthe importance that a few very high valuesmay have upon the estimation of the globalmean is important. If the variability of thedata can be reduced the confidence in the meangoes up.

Page 430: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 413

The discovery of major diamond deposits inthe barren lands of the Canadian Arctic wasone of the great exploration successes of the1990s. Targeting of the remote discovery arearesulted from remarkable persistence by ajunior company, although two major miningcompanies undertook the detailed explorationand development.

17.1 WORLD DIAMOND DEPOSITS ANDEXPLORATION

Diamond is the hardest mineral known and hassignificant industrial uses, e.g. in the cutting ofother materials and diamond drilling, but it is

its use in jewellery that is most important forthe exploration geologist (Fig. 17.1). Although90% of industrial diamonds are produced syn-thetically, it is, at present, uneconomic to pro-duce the much more valuable gem diamondsartificially.

Diamond deposits have many similarities toindustrial mineral deposits in that the valueof the raw material is only a small proportionof the value of the finished product. The worldrough diamond market in 2001 was worth$US7.9 billion whereas the retail diamondjewellery market was worth $US56 billion ofwhich $US13.5 billion was the value ofpolished stones (Diamond Facts 2003). Unlikeother minerals, wholesaling of gem diamonds

FIG. 17.1 The final product:diamonds from the Diavik Mine.(Courtesy Diavik Diamond MinesInc.)

17Diamond Exploration –Ekati and Diavik Mines

Charles J. Moon

Page 431: Introduction to mineral exploration

414 C.J. MOON

1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 20000

1

2

3

4

5

6

7

8

9

US

$ bi

llion

0

1

2

3

4

5

6

7

8

9

US

$ billion

Canada

Australia

Botswana

Russia

Angola

DRC/Zaire

Namibia

Other

South Africa

Africa alluvialsSouth Africa Big pipes

FIG. 17.2 World Production of Gem Diamonds. (De Beers 2003a.)

has been dominated by a single company, theDiamond Trading Company (formerly CentralSelling Organisation) owned by the De BeersGroup, which also controls many of the dia-mond mining companies. The De Beers Grouphas also recently set up a company to mar-ket retail diamonds. Unlike other industrialminerals, however, individual mineral grainshave a value that depends on their particularcharacteristics, especially if the diamonds arelarge.

The value of an individual diamond is deter-mined by the four Cs:

Carat, i.e. size. The term carat derives fromthe bean of the carob tree, which was used asa measure in ancient times. The measurewas metricated as 0.2 g in 1914.Color. Although diamonds are mainlythought of as colorless, colored or “fancy”stones are sought after. Colors range frompink through brown to blue. Even inclusion-rich black stones have their advocates.Clarity. This reflects the number of fracturesand inclusions in the stones and also affectsthe way in which diamonds are cut.Cut. The appearance after cutting deter-mines the return of light (“brilliancy”) andcolor flashes (“dispersion” or “fire”).

17.1.1 World diamond deposits

Diamonds are mined on a large scale in onlya few countries: South Africa, Botswana,Namibia, Angola, Democratic Republic ofCongo, Russia, Australia, and, recently,Canada (Fig. 17.2). Diamond production in2001 is estimated at 117 Mct (23.4 t) of which10% are gems, 55% near gems (stones withmany inclusions but usable in cheap jewellery),and 35% industrial (Diamond Facts 2003). Ofthe gems and near gems approximately 73%were from bedrock deposits and 27% fromalluvial workings (USGS 2003). However, thequantity is perhaps less important than thequality and hence monetary value (Table 17.1).Although Australian production is by far thelargest by mass (26 Mct) the price per carat islow ($US10.60 ct−1) and the value of productionat $US278M is less than the $US417M pro-duced in Namibia from 1.49 Mct at an averageprice of $US280 ct−1.

Bedrock sources

Although diamonds have been found in anumber of different bedrock sources, with theexception of minor (approximately 10% of

Page 432: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 415

TABLE 17.1 World diamond production by country. (From Diamond Facts 2003.)

Country Production Average Revenue(000 ct) ($ ct−1) ($ USM)

Angola 5175 141.0 730Australia 26,152 10.6 278Botswana 26,416 82.0 2160Brazil 650 40.0 26Central African Republic 450 146.0 65Canada 3691 134.0 450Congo (Dem. Rep.) 19,500 21.0 526Ghana 870 21.0 18Guinea 400 200.0 80Namibia 1490 280.0 417Russia 20,000 83.0 1665Sierra Leone 300 230.0 69South Africa 11,158 77.0 958Tanzania 191 120.0 23Venezuela 350 115.0 40Other 530 118.0 63

Total 117,323 65.0 7568

the eruption of volatile-rich magmas (Nixon1995). Typically pipes are small (Fig. 17.3) withsizes of 1–10 ha (120–350 m diameter) com-mon, although the largest pipes such as theMwadui and Orapa kimberlites and the Argylelamproite exceed 1 km2. At depth pipes coa-lesce into dykes (Fig. 17.4) which tend to bethin (often <1 m) but elongated (up to tens ofkilometers), often becoming more complex inshape. Where the magmas reached the paleo-surface they generally formed maars, volcaniccraters filled with lacustrine sediments, knownto reach up to a depth of 300 m. Other typesof eruption are also known and fissure-styleeruptions may generate kimberlitic tuffs thatinterfinger with other sediments, such as atFort a la Corne, Saskatchewan (Nixon 1995).In the classic pipes of Kimberley, South Africathe crater facies sediments are absent and thepipes are carrot shaped with walls dippinginwards at approximately 80 degrees. Thediatreme zone at Kimberley has a depth of500 m from the present surface to its basewith probably 1400 m of overlying sedimentsremoved by erosion (Clement et al. 1986).

In nonglacial areas, kimberlites weather andoxidize from the “blue ground” and “harde-bank” (resistant kimberlite), gradually becom-ing “yellow ground” towards the surface. This

world output) production from lithified placerdeposits, two rare igneous rock types are thehosts for virtually all commercial bedrock dia-mond deposits. The rock types are lamproiteand the predominant host, kimberlite. Agedating shows that the diamonds are much olderthan their host rocks in both types of igneousrock and are therefore present as xenocrysts(Richardson et al. 1984, 1990). Laboratory andtheoretical investigations demonstrate thatdiamonds form at high temperatures and pres-sures, corresponding to depths of at least150 km for a low continental geothermal grad-ient (Boyd & Gurney 1986). Kimberlites andlamproites are two of the few rock types gener-ated at these depths that are capable of bringingdiamonds close to the surface. It must how-ever be noted that only a few lamproites andkimberlites carry economic diamond deposits.De Beers reported that of the 2000 kimberlitestested over a period of 20 years only approxim-ately 2.5% (50) carried grades that might beeconomic (Diamond Geology 2003).

Morphology of kimberlites and lamproites

Most diamond-bearing kimberlite and lam-proites occur in pipe-like diatremes which aregenerally grouped in clusters and result from

Page 433: Introduction to mineral exploration

416 C.J. MOON

Mwadui

146

Catoca

Udachnaya

Koffiefontein

Frank Smith

11

66.2

62

Orapa

Premier Zarnitsa Finsch

Mir Dainya Leningrad PandaEkati

De Beers Dutoitspan

Liqhobong Hololo

Jagersfontein Kimberley

Hanaus I

500 m

Sloan I

110.6

32.2 28 17.9

7

3.7 51110

7

Kao

19.8

Letsang

A154-S andA154-NDiavik

16

FIG. 17.3 Pipe sizes. The size of thePanda pipe is taken from Nowickiet al. (2003) and the sizes of theA154 pipes digitized from anIkonos image. (Modified afterHelmstaedt 1993.)

is colored by the weathering of serpentine tosaponite and smectite and release of iron. In theupper levels of pipes kimberlite often forms an“agglomerate” of xenoliths of country rock andnodules that can be mistaken in exploration fora water-borne conglomerate (Erlich & Hausel2002).

Kimberlite petrology

Kimberlite is ill-defined petrologically andoften divided into two distinct groups.Mitchell (1991) defines Group I kimberlites ascomplex hybrid rocks consisting of mineralsrepresenting:1 Fragments of upper mantle xenoliths(mainly peridotites and ecologies, some includ-ing diamond).2 Minerals of the megacryst (large crystals) ordiscrete nodule suite (rounded anhedral grainsof magnesian ilmenite, Cr-poor titanian pyropegarnet, olivine, Cr-poor clinopyroxene, phlo-gopite, enstatite, and Ti-poor chromite).3 Primary phenocryst and groundmassminerals(olivine, phlogopite, perovskite, spinel, mon-ticellite, apatite, calcite, primary serpentine).

Group II kimberlites have also been termedmicaceous kimberlite or orangeite (Mitchell1995). They consist mainly of rounded olivinemacrocrysts in a matrix of phlogopite anddiopside with rare spinel, perovskite, andcalcite.

Lamproite geology

Lamproites are defined as potash- andmagnesia-rich rocks consisting of one ormore of the following phenocryst and/orgroundmass phases: leucite, Ti-rich phlo-gopite, clinopyroxene, amphibole (typically Ti-rich potassic richterite) olivine, and sanidine.Accessories may include priderite, apatite,nepheline, spinel, perovskite, wadeite, andilmenite. Xenoliths and xenocrysts includingolivine, pyroxene, garnet and spinel may bepresent, as well as diamond.

Peridotite and eclogite xenocryst andxenolith petrology

Diamonds have been found associated withseveral different types of xenolith but the

Page 434: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 417RESEDIMENTED VOLCANICLASTIC ROCKS

PYROCLASTIC ROCKS

Tephraring

Lac de Gras

DIATREME

Kirkland Lake

Somerset Island

Rootzone

BLOW

EROSION LEVEL

Volc

anic

last

icfa

cies

Dia

trem

e fa

cies

Hyp

abys

sal f

acie

s

PRECUSOR

DYKES

SILLS

DIATREME RELATED DYKES

FIG. 17.4 Model of South African kimberlite pipe formation with Canadian deposits for comparison. Theapproximate vertical scale of the diatreme facies is 2000 m. (Taken from McClenaghan & Kjarsgaard 2001,after Mitchell 1986.)

Page 435: Introduction to mineral exploration

418 C.J. MOON

two most important types are eclogite andperidotite:1 Eclogite. These are the most commonxenoliths in kimberlite and consist mainly ofomphacite clinopyroxene and almandine–pyrope garnet. Diamond inclusions in this typeof material are common and can reach thephenomenal grade of 18,000 ct t−1 at Orapa,Botswana (Robinson et al. 1984). This obser-vation, coupled with the observation that dia-monds in the interiors of eclogite xenoliths arepristine compared to the corroded nature ofstones in kimberlite, suggests that diamondscould have been derived by disaggregation ofeclogite during upward transport in kimberlite.The ultimate origin of eclogites is uncertainbut they could result from melting of garnet–peridotite or as fragments of subducted crust(Helmstaedt 1993).2 Peridotite. Although garnet–peridotite xeno-liths are relatively rare in kimberlites they arethought to be the most common xenoliths

based on the frequency of peridotite inclu-sions in diamond and the likelihood that itis the most common rock type in the mantle.The lack of peridotite xenoliths has beenascribed to their preferential disaggregationduring transport in the kimberlite (Boyd &Nixon 1978). The peridotite xenoliths consistof olivine, clinopyroxene, orthopyroxene, andgarnet. They are also enriched in diamond,although much less strongly than eclogite,with grades of up to 650 ct t−1 at the Finschmine (summary of Helmstaedt 1993).

Dating of the diamond with eclogitic inclu-sions has given isotopic ages of 2700–990 Mawhereas diamonds with peridotitic inclu-sions gave dates of approximately 3300 Ma(Richardson et al. 1984). This could mean thatthe two different types of diamond could haveformed by different processes although bothare associated with the roots beneath Archeancratonic areas (Fig. 17.5). The host kimberlitesand lamproite intrusives vary greatly in age,

Diamond stable

Graphite stable

100

50

0

Dep

th (k

m)

150

200

Relatively homogeneous

Rapid emplacementdiamonds

Slow emplacement –diamonds resorbed

Shear zone:heterogeneous and weak

Window for degassingof kimberlite magma

Percolating carbonatite liquid

Kimberlite

Crust

100

50

0

Depth (km

)

150

200

AsthenosphereLithosphere

FIG. 17.5 Section through the crust beneath southern Africa. Note the deep keel beneath the craton that isassociated with diamond generation. (From Vearncombe & Vearncombe 2002.)

Page 436: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 419

Cratons > 1500 Ma old

Nondiamondiferous kimberlites and lamproites Single intrusive body Province

Diamondiferous kimberlites and lamproites Single intrusive body Province (size of circle indicates approximate economic importance)

FIG. 17.6 World distribution of kimberlites and lamproite. (Updated from Evans 1993.)

e.g. pipes in southern Africa range from 60 to1200 Ma, although many economic pipes areapproximately 80–90 Ma in age.

Structural control of lamproites andkimberlites

Kimberlites and lamproites are restricted tocontinental intraplate settings with mostkimberlites intruding Precambrian rocks(Fig. 17.6). Clifford (1966) was one of the firstworkers to recognize that economically viablediamond deposits in kimberlite are restrictedto Precambrian cratons, particularly those ofArchean age (Fig. 17.6). This appears in the lit-erature as Clifford’s rule and has guided mucharea selection. The exact time control has beena matter of much debate and some authorshave divided Precambrian areas into archonsand protons, with archons being older than2.5 Ga (Levinson 1998, Hart 2002) and protonsconsisting of Proterozoic mobile belts adjacent

to archons (Janse & Sheahan 1995). Howeverthe diamond favorability criteria are by nomeans set in stone and many recent discoveriesin southern Africa have been in the LimpopoMobile Belt, which experienced a 2.0-Ga defor-mation event (Barton et al. 2003). Readersshould also note that not all archons hosteconomic kimberlites, e.g. South Americancratons, are almost barren in spite of muchexploration.

The relationship between pipe emplace-ment, particularly kimberlite emplacement,and structure has been much debated but noclear consensus has emerged. Vearncombeand Vearncombe (2002) presented a review ofthe distribution of kimberlites in southernAfrica (Fig. 17.7). They conclude that kim-berlite emplacement is related to corridorsparallel to, but not within prominent shearzones and crustal faults (Fig. 17.7). Jaques andMilligan (2003) report that similar deep struc-tures can be recognized in Australia based

Page 437: Introduction to mineral exploration

420 C.J. MOON

Legend

Kimberlite with

significant diamond

production

Other kimberlite

Alluvial diamonds

Edge of craton

River

International boundary

NAMIBIA BOTSW ANA ZIMBABW E

Namdeb

OrangeRiver

Finsch

7070

Kimberley60 –9 0

Jwaneng80 –90

VaalRiver

Premie r1200

80–9 0

150 –190

Marsfontei n

150

Orapa80 –90

Venetia510

120

60 –70

N

SOUT

LESOTHO

H

AFRIC A

0 100 200 400 600 800kilometers

FIG. 17.7 Kimberlite and alluvial diamond distribution in southern Africa, numbers are ages of kimberlites inMa (De Beers 2003b). (The kimberlite locations are modified from Vearncombe and Vearncombe 2002 andLynn et al. 1998, as well as LANDSAT images.)

has been substantial although of poor quality,has a source in Precambrian sediments as hassome alluvial production in Ivory Coast (Janse& Sheahan 1995). Although diamonds areknown from high grade metamorphic massifsin Kazakhstan, Norway, and China the dia-monds appear small and of poor quality (Nixon1995). They, like the diamonds reported fromophiolites, appear the result of rapid uplift

on geophysical data sets and suggest that theintrusions may be at the edge of differentlithospheric blocks.

Other primary sources

Lithified placer deposits are the major othersource of current diamond production. Forexample, Ghanaian alluvial production, which

Page 438: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 421

from the diamond stability field. Diamonds arealso known to occur as the result of meteoriteimpact, particularly at Popigay in northernRussia where stones up to 2 mm in size arefound in related alluvial deposits (Erlich &Hausel 2002).

Placer diamonds

Placer diamond production has, with notableexceptions, been mainly small scale. Even thissmall-scale alluvial production is howeverimportant in west African countries, not-ably Liberia, Sierra Leone, and Guinea. Majorrecent placer production has been derived fromthe alluvial deposits in Angola as well as thelarge alluvial and beach placers of southernNamibia and northwestern South Africa.These deposits are apparently derived by trans-port of diamonds in the paleo- and presentOrange and Vaal drainage systems from theinterior of South Africa as shown in Fig. 17.7(Moore & Moore 2004). Longshore currentsthen reworked these diamonds, particularlyconcentrating them at the bedrock–coarsesediment interface in a beach environment.The onshore reserves in Namibia are largelyexhausted and future mining will concentrateon offshore deposits. These diamonds are of

FIG. 17.8 Typical indicatorminerals. Minerals from top left,clockwise: picro-ilmenites,ecologitic garnets (G3), chromepyrope garnets (G9/G10),chromites, chrome diopsides,chrome Fe-titanium pyropegarnets (G1/G2), olivines. (Withpermission from SRC Vancouver.)

particular interest as 95% are gems, and rep-resent the largest gem diamond resource in theworld. However the mining environment isdifficult and some recent private ventures havenot been successful.

17.1.2 Geochemistry

Although diamond exploration usually relieson a combination of methods, geochemicaltechniques (in the broad sense) have beenvery successful in detecting diamondiferouspipes. A number of distinctive, dense min-erals (Table 17.2, Fig. 17.8) are associated withkimberlites and lamproites and it is theseminerals that have been used to indicate thelocations of pipes, hence their name indicator(or, in translation from Russian, satellite) min-erals. The chemistry of some of these mineralsalso indicates their sources and transport his-tory and, by comparison with known deposits,can be used to predict whether the source isdiamondiferous as well as some indication ofgrade.

The indicator minerals used since the latenineteenth century and shown in Fig. 17.8 aregarnet (both pyrope and eclogitic), chrome diop-side, chromite, and picro (Mg-rich) ilmenite(Gurney & Zweistra 1995, Muggeridge 1995).

Page 439: Introduction to mineral exploration

422 C.J. MOON

TA

BLE

17.2

Indi

cato

r m

iner

al c

har

acte

rist

ics.

(Sim

plifi

ed fr

om M

cCle

nag

han

an

d K

jars

gaar

d 20

01.)

Min

eral

Pyr

ope

garn

et

Mg-

ilm

enit

e

Cr-

diop

side

Cr-

spin

el

Fors

teri

tic

oliv

ine

Dia

mon

d

Mai

n so

urce

rock

s

Kim

berl

ite,

lam

proi

te,

peri

doti

te, l

ampr

oph

yre

Kim

berl

ite

Kim

berl

ite,

lam

proi

te,

peri

doti

te, o

ther

ult

ram

afics

Lam

proi

te, k

imbe

rlit

e,ca

rbon

atit

e, v

ario

us

basi

can

d u

ltra

mafi

c ro

cks

Per

idot

ite,

lam

proi

te,

kim

berl

ite,

car

bon

atit

e

Kim

berl

ite,

lam

proi

te,

cert

ain

lam

prop

hyr

es,

met

amor

phic

roc

ks

Com

posi

tion

Mg

Al s

ilic

ate

Mg

Fe T

i oxi

de

Ca

Mg

sili

cate

Mg

Fe C

r A

lox

ide

Mg

sili

cate

C

Col

or

Pu

rple

,re

d, o

ran

ge

Bla

ck

Em

eral

dgr

een

Bla

ck,

redd

ish

brow

n

Pal

e ye

llow

to g

reen

Col

orle

ss,

pale

col

ors

Mag

neti

sm

Non

e to

ver

yw

eak

Med

ium

to

stro

ng

Non

e to

ver

yw

eak

Wea

k t

o m

ediu

m

Non

e to

ver

yw

eak

Non

e

SG 3.5

4.5

–5

3.2

–3.6

4.3

–4.6

3.2

–3.3

3.5

Har

dnes

s

7.5

5–6

5–6

5.5

6–7

10

Dia

gnos

tic

feat

ures

An

hed

ral,

kel

yph

ite

rim

, col

or

An

hed

ral,

rou

nde

d,gl

assy

lust

re,

con

choi

dal f

ract

ure

An

hed

ral,

bloc

ky,

col

or

Oct

ahed

ral t

oir

regu

lar

shap

e,re

ddis

h b

row

n o

ned

ges

Col

or, i

rreg

ula

rcr

ysta

l api

ces

Adm

anti

ne

lust

re,

crys

tal f

orm

Page 440: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 423

garnet grains can then be plotted to highlightareas for follow-up (see Fig. 8.16).

Indicator mineral chemistry

The study of the chemistry of indicator min-erals requires the use of a scanning electronmicroscope or electron microprobe, typicallyat a cost of $US10 per analysis. The aim is toidentify the diamond potential of the intrusivefrom which the minerals have been derived byanswering three questions (Fipke et al. 1995):1 How much diamond-bearing peridotite andeclogite are present?2 What was the diamond grade of the sourcerocks?3 How well were the diamonds preservedduring transport to the surface?

The first question can be addressed by exam-ining the chemistry of garnets and chromitesto assess the contribution of disaggregatedharzburgite (olivine and orthopyroxene) mat-erial. Plotting of CaO against Cr2O3 was usedby Gurney (1984) to define a line to the leftof which 85% of known Cr-pyrope diamondinclusions plot (Fig. 17.9). Garnets plotting tothe left of the line are termed G10 garnets andthose to the right G9 after a garnet classifica-tion of Dawson and Stephens (1975). Lowercalcium, higher chromium garnets are associ-ated with higher diamond contents. The line at2% Cr2O3 was used to discriminate betweengarnets of peridotitic and low-Ca garnets whichare probably of eclogitic or megacrystic origin.A plot of MgO against Cr2O3 for chromitedefines a highly restricted Cr-rich field fordiamond-related chromite. The diamond po-tential of eclogitic garnets can be estimatedfrom a plot of Na2O against TiO2. In general theNa2O content of garnets from eclogitic dia-mond inclusions is greater than 0.07% Na2O.

The preservation of diamonds during trans-port from the mantle to the surface is esti-mated by attempting to assess the oxygenfugacity of the magma. It had been observed thatthe diamond content of kimberlite correlatedwith trends in picritic ilmenite, particularlyFe2O3 against MgO, with low Mg representinghighly oxidized kimberlite (Gurney & Zweistra1995). It should be noted that the diamondpotential can only be downgraded by estimatesof the diamond preservation potential.

Typical abundances in the host kimberlite arehard to assess, but Erlich and Hausel (2002)estimate ilmenite as between 1 and 10 wt%of some kimberlites. At the Premier pipe,ilmenite was estimated at 0.5% of kimberlite,and thought to be the highest of any kimberlitemined by De Beers in South Africa (Bartlett1994). The ilmenite–garnet–diopside ratio atthe Premier mine was estimated at 100:1:1.Lamproites have very different indicatormineral assemblages: chromite, Nb-rich rutile,ilmenite, zircon, tourmaline, and garnet havebeen reported as most useful, although garnetsare much less abundant than in kimberlites(Erlich & Hausel 2002).

Heavy mineral samples are generallystrongly diluted with nonkimberlitic mater-ial, and finding the signal of a kimberlite ona regional basis requires detecting very smallnumbers of indicator minerals; sometimes asingle grain may be significant. Correct sampl-ing and processing of samples is therefore veryimportant. Although orientation studies arerequired to optimize sample size, generally10–40 kg is used, with some lamproite searchesrequiring larger samples (Muggeridge 1995).

The size of indicator minerals is generally inthe range 2 mm to 0.3 mm and samples arescreened in the field through a 2-mm screenand then bagged. Many exploration programsin the past have panned or jigged samples inthe field but this approach has been largelyreplaced by laboratory separation of the wholesample. Laboratory processing is complex andexpensive with typical processing costs of sev-eral hundred dollars per sample. The key phasein processing is generation of a heavy mineralconcentrate using either a shaking table ordense media separation. The heavy mineralconcentrate is then split into ferromagnetic,paramagnetic, and nonmagnetic fractions usinga magnetic separator, followed by size frac-tionation if any of the fractions are heavierthan 9 g. Each fraction is then visually exam-ined under a binocular microscope and anyindicator minerals counted and mounted onstubs for possible examination by a scanningelectron microscope or electron microprobe.The binocular examination is time consumingand requires training to recognize the minerals,as well as immunity to boredom once the train-ing is complete. The number of ilmenite or

Page 441: Introduction to mineral exploration

424 C.J. MOON

16

12

8

4

00 2 4 6 10

CaO (wt%)8

G10 G9

80

70

60

50

40

300 5 10 15 20

MgO (wt%)

Diamond Inclusion Field

6

4

3

2

1

00 2 4 18

MgO (wt%)

5

6 8 10 12 14 16

1.2

0.8

0.6

0.4

0.2

00 0.5 0.1 0.15 0.2

Na2O (wt%)

1.0

TiO

2 (w

t%)

Cr 2

O3

(wt%

)C

r 2O

3 (w

t%)

Cr 2

O3

(wt%

)

(a) (b)

(c) (d)

MegacrystsGp.I EclogiteGp.II Eclogite

FIG. 17.9 Electron microprobe data from indicator minerals collected during the Ekati discovery. Fordiscussion see text. (From Fipke et al. 1995.)

Conventional exploration geochemistry

Conventional geochemistry has also beenused where there is a geochemical contrast be-tween the kimberlitic material and their hostrocks. Elements of use include Ba, Ni. Cr, V,Co, Mg, K, partial Sr, Nb, Ta, P, and light rareearth elements (McClenaghan & Kjarsgaard2001).

17.1.3 Geophysics

Geophysical techniques are widely used todetect kimberlite and lamproite pipes, par-ticularly at the reconnaissance stage. Thetechniques used will depend on the climaticand weathering history, the country rock intowhich the pipes have been intruded, and the

age of the pipes. In general the magnetic, elec-trical (conductivity and induced polarisation),density, and seismic velocity properties of apipe are significantly different from the coun-try rock.

Airborne magnetic and electromagnetic(EM) surveys have been widely used for directkimberlite detection, although responses canbe highly variable with some pipes in a clusterdetectable and others not (Macnae 1995, Smith& Fountain 2003). In addition, close flight linespacing is needed as the targets are relativelysmall. For example, in northwest Canada,regional government surveys flown at 800 mline spacing and 300 m terrain clearance de-tected only one pipe, whereas company surveysflown at 125 m line spacing and 30 m terrainclearance were able to detect more than 100

Page 442: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 425

pipes (section 17.2) (St. Pierre 1999). Most pipesare present as discrete bull’s eye anomaliesalthough some may be associated with linearfeatures. The formation of clay minerals duringweathering of kimberlite diatremes or craterfacies kimberlite often reduces the magneticresponse but produces a very conductive coverto pipes, which makes them amenable to detec-tion by EM techniques (Macnae 1995). A goodexample is the Point Lake pipe discussed indetail in section 17.2.1 (Smith et al. 1996).In other areas such as Western Australiakimberlite can be more resistive than conduc-tive soils. Airborne gravity gradiometers, suchas the Falcon system of BHP Billiton, are beingwidely used to search for pipes, especiallywhere the pipes are covered by sand, as in partsof Botswana. This technology will detect large(>200 m diameter) pipes using a 100 m linespacing, and in test flights over the Ekati areaabout 55% of 136 known pipes were detected(Liu et al. 2001, Falcon 2003). In general kim-berlite is less dense than country rock. Radio-metrics have also been used where intrusivescrop out and where there is a contrast withcountry rock.

Ground geophysics is also widely used todefine and refine the source of the airborneanomalies. The individual phases of intrusioncan also often be defined, which considerablyaids in the definition of grade (Macnae 1995).In addition to the methods mentioned above,seismic surveys have been used both inYakutia (Macnae 1995) and at Snap Lake innorthern Canada (Kirkley et al. 2003), althoughtheir expense has restricted their use to defini-tion of pipes and dyke geometry.

17.1.4 Remote sensing

The circular nature of many pipes and theircross-cutting relationship with local strati-graphy make their outcrops relatively easy torecognize on air photography and high resolu-tion satellite imagery. The distinctive min-eralogical composition of pipes can also bedetected using spectral measurements in theshortwave red infrared wavelengths for miner-als such as phlogopite and serpentine. How-ever Kruse and Boardman (2000) reported greatdifficulty in interpreting Hymap airborne scan-ner data over kimberlites in Wyoming due to

the intense weathering and very limited out-crop of the distinctive minerals.

17.1.5 Evaluation

The relative scarcity of diamonds in hardrockdeposits and the great variability of value ofindividual diamonds cause many problemsin evaluation and valuation when a pipe hasbeen found. In some low grade primary, andmany alluvial, deposits the recovery of largestones is crucial to viability. Rombouts (2003),who has provided some of the few publisheddiscussions of the problem, suggests that isnot uncommon for a single stone to repre-sent 30% of the entire value of a parcel ofdiamonds. During commercial mining onlystones larger than 1 mm (macrodiamonds) aregenerally recovered, although some Russianmines recover down to 0.2 mm for use asdiamond powder.

Evaluation begins once a pipe has been dis-covered and consists of a number of stageswhich improve confidence in the grade andvalue of the diamonds. These stages may takeconsiderable time, often years. In addition asthe pipes occur in clusters there are often anumber of targets to test and it is usually notclear which pipe in the cluster is likely to be ofmost economic interest. Within pipes with anumber of differing kimberlite phases, thephases are likely to have different grades thatneed to be identified. Thomas (2003) dividedthe evaluation phases into four:1 Initial delineation drillhole.2 Delineation drilling: determine the size andgeometry of the pipe and whether there aremultiple intrusive phases.3 Mini bulk sampling: large diameter drilling(or pitting) to recover a sample that is processedthrough commercial dense medium plant torecover diamonds larger than 1.0 mm.4 Underground or surface bulk sampling:recover 5–10,000 ct by dense medium separa-tion for sorting and valuation.

Initial confirmation of the kimberlite isnormally by core drilling but analysis of thecore, as well as that of the delineation stage(2 above), can provide information on theindicator assemblage and possible grade. Thecore is dissolved either by caustic fusion orhydrofluoric acid and minerals separated at a

Page 443: Introduction to mineral exploration

426 C.J. MOON

99

90

50

10

1

0

0.00001 0.0001 0.001 0.01 100

Weight (ct)

1

Stone size (mm)

0.1 1 10

100

0.1 10

Cum

ulat

ive

freq

uenc

y

Carats

Numbers

FIG. 17.10 Use of microdiamonds in predicting overall grade using a logarithmic (x axis) probability (y axis)plot. The sample of 52 kg contained 58 stones with a total mass of 0.0169 ct. Extrapolation gives a grade of0.06 ct t−1 above 0.01−1ct or 1 mm. (Taken from Rombouts 2003.)

cost of approximately $US350 per 10 kg ofsample. The grade of the macrodiamonds canbe estimated from the microdiamond (<1 mm)distribution, as the central part of the distribu-tion is log-normal (Fig. 17.10). Poor recoveryof very fine stones and the chance occurrenceof large stones may cause variation from log-normality at the ends of the graph. A log–logplot of stone size against cumulative grade willprovide an estimate of overall grade. Typicallystage 1 will cost about $US10,000. Stage 3will take samples of 200 t or more. In Canadathis will comprise about 10 large diameterdrillholes depending on the grade. The fourthstage in which representative diamonds are re-

quired for commercial valuation is much moreexpensive and may reach the order of $US10M.

17.1.6 History of discovery

Kimberley, South Africa and the discoveryof kimberlite pipes

Diamonds have been known since at least thefourth century BC but it appears that India wasvirtually the world’s sole source of diamondsuntil 1730, when diamonds were discoveredin Brazil (Levinson 1998). All the Indian pro-duction was considered alluvial, although onearea was an unrecognized primary deposit.

Page 444: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 427

Similarly, all the diamonds produced in Brazilfrom 1730 to 1870 were also alluvial mainlyfrom Minas Gerais and Bahia states. Total pro-duction from 1730 to 1891 in Brazil has beenestimated at 13 Mct (Levinson 1998).

The breakthrough in recognition of a bed-rock source for diamonds came in the 1870sin South Africa. Alluvial diamonds were fol-lowed from their discovery on the banks ofthe Orange River, northwest of Hopetown(Fig. 17.7) to a number of sources in the Kim-berley area. These sources were then minedvertically, first in yellow ground and then intoblue ground. Individual claims were amalg-amated into the major operations at theKimberley and De Beers diggings by CecilRhodes, an entrepreneur who founded the DeBeers company, which eventually controlledmining of these two main pipes (Hart 2002).

The nature of the blue ground was firstrecognized by Emil Cohen as a volcanic rockin 1872. The term pipe was first applied to thecylindrical form of the blue ground in 1874and kimberlite was first applied to the diamon-diferous rock type in 1897. Three of the pipes atBultfontein, Dutoitspan, and Wesselton minesare still in underground production, althoughoutput is relatively small (Table 17.3). TheDe Beers and Kimberley mines are now closedalthough the latter now forms part of a well-known museum.

Other African pipes and alluvials

The success of the Kimberley mines led to thesearch for other diamond deposits in SouthAfrica. The most notable discovery was in 1902by a former bricklayer, Thomas Cullinan, whofound the Premier pipe, approximately 100 kmnorth of Johannesburg, that was three timelarger at surface than the largest Kimberleypipe. In 1905 the pipe produced the largestknown diamond, the Cullinan of 3106 ct, themain part of which is now in the British crownjewels. De Beers eventually bought control ofthe Cullinan company in 1914 and productioncontinues to the present in an undergroundmine. Although the grade is low the pipe pro-duces a considerable number of large (>100 ct)stones.

The next major diamond development insouthern Africa was not of a kimberlite pipe

but of alluvial diamonds on the coastline ofthe Atlantic Ocean to the south and north ofthe Orange River. Although the stones arerelatively small they are mainly gems andwere initially discovered in sand dunes in, thethen, German South West Africa by a railwayworker. The deposits were taken over by SouthAfrica after the World War I, the area closed offto outsiders, and sold to a company controlledby Anglo American Corporation. These con-cessions were then merged with the mines ofDe Beers in 1929 and are currently operated byNamdeb, a joint venture between De Beers andthe Namibian government.

As a result of the abundant supply of dia-monds there was relatively little further sys-tematic exploration in Africa until the 1950s.However, in the early 1940s an ex De Beersemployee, John Williamson, discovered amajor pipe at Mwadui in what is now Tanzania.Looking for the source of previously knownalluvial diamonds, he followed ilmenite indic-ator minerals until he eventually panned adiamond (Krajick 2001).

Starting in the mid 1950s De Beers beganlarge scale exploration in Africa, based aroundindicator mineral sampling, backed up by airphotography and geophysics. The most sig-nificant discoveries resulting from this pro-gram were made in central Botswana. De Beersdeveloped techniques to sample the surfacedeflation layer for indicator minerals that areuseful even in areas of thick sand cover. Theheavy minerals are thought to be brought to thesurface (at least partly) by the action of ter-mites (bioturbation) and concentrated by windaction. De Beers geologists collected samples,usually of the –2 mm +1 mm fraction, at 10- to15-m intervals that were then aggregated into8-km-long sections.

Although indicator minerals and alluvialdiamonds were known in Botswana in the early1960s, no kimberlites were known (Baldocket al. 1976). A small occurrence of alluvialdiamonds was traced westwards from the farnortheast corner of Botswana but the disper-sion trail of diamonds was cut off to the west.Working on the premise that the trail hadbeen cut off by Pliocene upwarping, GavinLamont moved the sampling further west. Thissampling confirmed the occurrence of indic-ator minerals and located a high amplitude

Page 445: Introduction to mineral exploration

428 C.J. MOON

TA

BLE

17.3

Pip

e si

zes

and

grad

e. (M

ain

ly a

fter

Jan

se &

Sh

eah

an 1

995,

Jen

nin

gs 1

995,

De

Bee

rs 2

001.

)

Pipe

Cou

ntry

Surf

ace

Yea

rT

otal

pro

duct

ion

Gra

de$

ct−1

$ t o

re−1

area

(ha)

disc

over

ed(M

ct o

r res

erve

)(c

pht)

Kim

berl

ey, K

imbe

rley

Sou

th A

fric

a3.

718

7132

.710

020

020

0D

e B

eers

, Kim

berl

eySo

uth

Afr

ica

5.1

1871

36.4

9075

65B

ult

fon

tein

, Kim

berl

eySo

uth

Afr

ica

9.7

1869

36.2

4075

30D

uto

itsp

an, K

imbe

rley

Sou

th A

fric

a10

.818

6921

.320

7515

Fin

sch

Sou

th A

fric

a17

.919

6093

(to

1995

)80

4032

Ven

etia

Sou

th A

fric

a12

.719

80~3

5 (1

15)

136

100

136

Jwan

eng

Bot

swan

a45

1972

171

(413

)14

711

016

1.7

Ora

paB

otsw

ana

110.

619

6714

6 (3

20)

6854

34M

wad

ui

Tan

zan

ia14

619

406

150

9M

arsf

onte

inSo

uth

Afr

ica

1997

1.83

169

165

278

Arg

yle

Au

stra

lia

4619

7954

154

99

49.4

Mir

Ru

ssia

719

55In

tern

atio

nal

yaR

uss

ia1.

740

012

048

0U

dach

nay

aR

uss

ia52

1955

100

100

100

Lom

onos

ovR

uss

ia11

875

8060

Gri

bR

uss

ia16

1996

R 6

769

7955

Pan

da E

kat

iC

anad

a3

1993

R 1

2.6

(pit

)10

913

014

2K

oala

Ek

ati

Can

ada

1993

R 1

4.6

(pit

)76

122

93Fo

x E

kat

iC

anad

a19

93R

17.

740

125

50M

iser

y E

kat

iC

anad

a19

93R

12.

642

626

111

A15

4-S

Dia

vik

Can

ada

1994

R 1

0.8

467

5927

6A

154-

N D

iavi

kC

anad

a19

94R

2.7

289

3395

A41

8 D

iavi

kC

anad

a19

94R

8.3

388

5320

6A

21 D

iavi

kC

anad

a19

94R

3.7

289

3610

4

cph

t, c

arat

s pe

r h

un

dred

ton

nes

; R, p

re-m

inin

g re

serv

e.

Page 446: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 429

FIG. 17.11 Aerial view of Kimberley Pipe. The view shows the city of Kimberley to the east of the “Big Hole.”The pit in the background is the De Beers pipe.

indicator mineral anomaly associated with acircular structure on air photography. The firstpits in 1967 returned 12 diamonds within 45minutes of running a rotary pan test. This ledto the establishment of the major mine atOrapa, accompanied by the location of thenearby smaller Leklhatkane pipes in 1971. DeBeers then moved to the south under thicker(up to 100 m) sand discovering the first pipein the Jwaneng area in 1972 and the Jwanengpipe (DK2 of Fig. 8.16) in 1973 (Lock 1987).The Jwaneng pipe is, and has been, one of theworld’s most profitable mines.

South Africa has also been the target of muchsystematic exploration. The major discoveriesinclude the Finsch mine, found by a prospectorin the 1950s, and Venetia, discovered by DeBeers in 1980. Smaller pipes and fissures, such asMarsfontein, make attractive targets for smallercompanies (Davies 2000). The Marsfonteinpipe paid back its investment in 1 week!

Russian pipes

Until 1991, little was publicly known of dia-mond exploration or mining in the USSR. It isnow apparent that the major producing area incentral Siberia was discovered as the result of asystematic search organized after World War II(Erlich & Hausel 2002). Although diamondswere known from the Urals placer gold work-ings since 1829, the discovery resulted from theuse of geological mapping and heavy mineralsampling (Fig. 17.12). Follow-up of diamond oc-currences in black sands using the occurrenceof pyrope garnets and mapping eventually ledto the discovery of the Zarnitsa pipe in 1954.Among the pipes discovered shortly after werethe high grade Mir (1955), Udachnaya (1957),and Aikhal (1960) pipes (Table 17.3). In the1970s a further cluster of pipes was discoveredin the area to the east of the White Sea, nearthe city of Akhangl’sk. The pipes include the

Page 447: Introduction to mineral exploration

430 C.J. MOON

Norilsk

YakutskW. Siberian

Platform

Taimyr Fold Belt

Verkhoyansk Fold Belt

Baikal

Fo

ldBelt

AnabarAnabarAnabarShieldShieldShield

Aldan ShieldAldan ShieldAldan Shield

Laptev Sea

Olenek

R.

R. Lena

R. Lena

R. Lena

R. Cha

ra

R. Cha

ra

R. Cha

ra

Tunguska R.Tunguska R.Tunguska R.

1

2

3

R U S S I A

Area of Map

Kimberlite groups

Archean shields

Boundary of EastSiberian Platform

60°N

70°N

80°E 100°E 120°E 140°E 160°E

N

500 km500 km500 km0

FIG. 17.12 Distribution of Kimberlite pipes in Siberia. The major economic pipes are the Mir pipe in cluster 1,Udachnaya in cluster 2. The Leningrad pipe is in cluster 3. (After Pearson et al. 1997.)

Lomonosov deposit, consisting of several pipes,and the Grib pipe, which was discovered by anon-Russian junior company in 1996.

Argyle

Australia was an obvious target to search fordiamonds with its large Archean cratons, par-ticularly the Yilgarn and Pilbara shield areain Western Australia as well as the Gawlercraton in South Australia. However until 1970most of the limited production of diamonds(200,000 ct) had been recovered from alluvialdeposits in the Paleozoic fold belt area ofWestern Australia (Smith et al. 1990). The

north of western Australia, including thecentral older Kimberley block, was targetedby junior exploration companies based on thesouthern African model, the known occur-rence of ultra-potassic intrusives, which werethought analogous to the barren alkalic rocksoff craton in southern Africa, as well as a gooddrainage network. The rivers allowed the use ofsediment sampling by helicopter for indicatorminerals to define follow-up areas (Fig. 17.13).Follow-up of indicator minerals in the northKimberley area showed barren kimberlitedykes with pyrope and picro–ilmenite indic-ator assemblage but in the west Kimberleyarea, a breccia pipe of lamproitic affinity was

Page 448: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 431

lllllllllllll

llll

llllllll

lll

llllllllllllllllllllllllllll

KimberleyCraton

Argyle

AKIFitzroy

TroughKimberliteLamproiteDiamondoccurrence

N

200 km200 km200 km0

FIG. 17.13 Lamproites andKimberlite distribution: Kimberleyarea of Western Australia. Note theoccurrence of the Argyle mine offthe craton. (From Evans 1993, afterAtkinson et al. 1984.)

found to contain diamonds associated withchromite as an indicator mineral. This was thefirst time that diamonds had been reportedfrom lamproite. This discovery led to thediscovery of the (then) uneconomic Ellendalecluster of pipes. Further regional sampling byCRA, who had farmed into the project, basedon 40-kg samples returned two diamonds in asample. This was followed 20 km upstream tolocate the Argyle AK1 pipe and some associ-ated alluvial deposits. The grade of the depositis very high with 75 Mt proven and probablereserves at approximately 6700 carats perhundred tonnes (cpht). However the quality ofdiamonds is low: only 5% gem, 40% near gem,and 55% industrial stones with an averageprice in 1982 of $US11 ct−1.

The Argyle discovery was important in thatit was the first significant discovery of a dia-mondiferous nonkimberlitic pipe and this pipewas in a mobile zone, if Precambrian in age.

17.2 CANADIAN DIAMOND DISCOVERIES:EKATI AND DIAVIK

Although Canada has the largest area ofArchean shield in the world and is thus afavorable area, it was not until 1991 that eco-nomic diamond deposits were discovered. Asummary of early work by Brummer (1978)

shows that diamonds were known from easternCanada as early as 1920. A diamond was foundtogether with indicator minerals in till andthin kimberlitic dykes discovered in an up-icedirection in the Kirkland Lake gold mining areaof Ontario, but further sampling was discourag-ing. De Beers also has a long history of explora-tion in Canada and investigated kimberlite inArctic Canada on Somerset Island in 1974–76.In a joint venture with Cominco, at least 19pipes were discovered and sampled, althoughgrades were low.

17.2.1 Ekati discovery

Although major companies had been involvedin the diamond search in Canada as detailedby Brummer, a junior company, led by two geo-logists, Chuck Fipke and Stewart Blusson,located the remote Lac de Gras target area in1988. The history of the Ekati discovery is welldocumented in the very readable book byKrajick (2001) and the career of Fipke describedin a book by Frolick (1999).

Fipke set up a consulting mineralogical com-pany after working as an exploration geologistfor Kennecott, Cominco and JCI in southeastAsia, Brazil, and southern Africa for 7 years.During his stay in South Africa he had visitedKimberley and the Finsch mine and became in-terested in using large heavy mineral samples

Page 449: Introduction to mineral exploration

432 C.J. MOON

FIG

. 17.

14R

egio

nal

set

tin

g, L

ac d

e G

ras

expl

orat

ion

. (A

fter

Kra

jick

200

1.)

BO

RE

AL

FO

RE

ST

BO

RE

AL

FO

RE

ST

BO

RE

AL

FO

RE

STBBB

AAARRR

RRREEENNNLLLAAANNNDDDSSS

TTTRRR

EEEEEE

LLLIIINNN

EEE

TTTRRR

EEEEEE

LLLIIINNN

EEE

MMMAAACCCKKKEEENNNZZZ IIIEEE MMM OOOUUU NNN TTTAAA IIINNNSSS

NNNOOO

RRRTTT

HHHWWW

EEESSS

TTTTTT

EEERRR

RRRIII

TTTOOO

RRRIII

EEESSS

AAARRR

CCCTTT

IIICCC

CCCIII

RRRCCC

LLLEEE

Gre

at B

ear

Lake

Gre

at S

lave

Lake

Bla

ckw

ater

Bla

ckw

ater

Bla

ckw

ater

Lake

Lake

Lake

Mai

n E

sker

Mai

n E

sker

Mai

n E

sker

Mai

n E

sker

Mai

n E

sker

Mai

n E

sker

Hud

son

Bay

Con

twoy

to L

ake

Con

twoy

to L

ake

Con

twoy

to L

ake

Yam

baY

amba

Yam

baLa

keLa

keLa

ke

Des

teffa

nyD

este

ffany

Des

teffa

nyLa

keLa

keLa

ke

Bat

hurs

t Inl

etB

athu

rst I

nlet

Bat

hurs

t Inl

et

MMM

aaaccckkkeeennnzzziiieee RRRiiivvveeerrr

Mackenzie River

Mackenzie River

Mackenzie River

CCCoooppp

pppeeerrrmmmiiinnneee RRRiiivvveeerrr

Lac

duLa

c du

Lac

duS

auva

geS

auva

geS

auva

ge

Mis

ery

Poi

nt/

Mis

ery

Poi

nt/

Mis

ery

Poi

nt/

The

Nar

row

sT

he N

arro

ws

The

Nar

row

s

Gla

cial

Lak

eG

laci

al L

ake

Gla

cial

Lak

eM

cCon

nell

McC

onne

llM

cCon

nell

Hor

nH

orn

Hor

nP

late

auP

late

auP

late

au

Mou

ntai

nM

ount

ain

Mou

ntai

nD

iatr

eme

Dia

trem

eD

iatr

eme

ToToToE

dmon

ton

Edm

onto

nE

dmon

ton

Nor

th

Am

eric

anCoa

st

NorthAm

eric

an

Co

ast

Nor

thA

mer

ican

Coast

Nor

man

Wel

lsN

orm

an W

ells

Nor

man

Wel

ls

De

Bee

rs C

amp

De

Bee

rs C

amp

De

Bee

rs C

amp

For

t Sim

pson

For

t Sim

pson

For

t Sim

pson

Lac

la M

artr

eLa

c la

Mar

tre

Lac

la M

artr

e

Aca

sta

Aca

sta

Aca

sta

Gne

isse

sG

neis

ses

Gne

isse

s

Nor

ms

Nor

ms

Nor

ms

Cam

pC

amp

Cam

p

Cop

perm

ine

Cop

perm

ine

Cop

perm

ine

(Kug

lukt

uk)

(Kug

lukt

uk)

(Kug

lukt

uk)

Blo

ody

Fal

lB

lood

y F

all

Blo

ody

Fal

l

Yello

wkn

ifeYe

llow

knife

Yello

wkn

ife

Prin

ce o

f Wal

es’s

For

tP

rince

of W

ales

’s F

ort

Prin

ce o

f Wal

es’s

For

t17

21–1

782

1721

–178

217

21–1

782

Lac

deLa

c de

Lac

de G

ras

Gra

s G

ras

(Eka

ti)(E

kati)

(Eka

ti)

200

km20

0 km

200

km000

NNN

The

Bar

ren

Land

s an

dS

urro

undi

ng R

egio

ns

Page 450: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 433

to target areas for follow-up. This was the tech-nology he developed for his own business thathe set up in 1977 in his home town in southernBritish Columbia. One of his clients wasSuperior Oil, which was engaged in a diamondsearch in the USA headed by exploration man-ager, Hugo Dummett, a former De Beers geo-logist. Blusson’s career was initially moreacademic. He had worked for the GeologicalSurvey of Canada but then became involvedin exploration as a consultant and on his ownaccount. He had the added advantage of beinga qualified helicopter pilot.

Fipke began his Canadian search as a juniorpartner with Superior Oil, in the Rocky Moun-tains area near Kimberley British Columbia ina company called C.F. Minerals. Cominco andother companies had investigated a numberof small nondiamondiferous kimberlite pipesdiscovered in 1976 during logging operations(Fig. 17.14). The Superior-funded search locateda number of pipes, including the Jack pipe,north of Kimberley, based on indicator mineralanalysis, particularly chrome diopsides. TheJack Pipe samples however contained only onediamond and few garnets. In spite of this lackof success, the joint venture moved north andexamined a pipe named Mountain Diatremethat was discovered by geology summer stu-dents exploring for gold (Fig. 17.14). When theywere examining this pipe they learned from ahelicopter pilot that De Beers (exploring underthe name Diapros in Canada) had a large ex-ploration program 250 km to the east in theBlackwater River area. They then (1982) stakedan area adjacent to Diapros’s claims after find-ing favorable indicator minerals, particularlyG10 garnets. However Diapros failed to findany kimberlites and withdrew, as the De Beersboard was unwilling to pursue the dispersiontrail to the more remote areas to the east. Atthis stage (1981) Superior withdrew from dia-mond exploration, due to changes in seniormanagement, and terminated the joint ventureleaving the claims to C.F. Minerals.

In order to finance further exploration, Fipkeand Blusson floated a company, Dia Met Min-erals, raising funds from individuals in Fipke’shome town and from corporate investors inVancouver. Convinced that the source of theindicator minerals was to the east, they con-firmed the eastern extension of the glacial train

associated with a major esker by samplingaround Lac de Marte. In 1984 they eventuallyraised enough finance for further regional-scaleheavy mineral sampling to establish the sourceof the dispersion train, which eventually tookplace in 1985. By mid 1988 processing of thesampling was complete and a pattern becameclear. The area east of Lac de Gras (Fig. 17.15)was essentially barren and one sample G71from an esker beach contained a huge num-ber of indicator minerals: 100 ilmenites, 2376chrome diopsides, and 7205 pyrope garnetgrains relative to the usual 10 or 20 of the re-gional dispersion train. The area to the north ofLac de Gras was obviously the area of interest(Fig. 17.15), although the complex glacial his-tory made it difficult to tie down the location ofany other pipes. This lack of definition meantthat a large land position was needed so thatany pipes could be located and tested. Themajor problem was raising the finance to dothis, each claim of 2582.5 acres would cost $C1per acre to register and have to be physicallystaked. Staking required putting a woodenpost in the ground every 458 m (1500 ft) andmarking it, an intensive operation requiringfloat plane or helicopter support. By the end ofthe summer season they had staked 385,000acres in the name of Norm’s ManufacturingCompany to avoid raising suspicion at theYellowknife Mining Recorder’s office. In April1990, as soon as conditions permitted, Fipkestaked further ground to the south and recog-nized that a lake surrounded by steep cliffs at aplace called Misery Point could be the locationof a pipe. While collecting a till sample of abeach at this location, Fipke’s son Mark founda pea-sized chrome diopside that suggestedthat a kimberlite pipe was very close, probablyunder the lake.

The next stage was to find a joint venturepartner to undertake detailed exploration,test any pipes, and most of all maintain theexpenditure requirements on the large landposition ($C2 per acre per year, a total of$C770,000). They approached Dummett, bythen North American exploration manager forBHP Minerals and agreed a joint venture. Hewas presented with data on indicator mineralabundances and results of an analysis byconsultant John Gurney. They showed thatthe garnets and ilmenites suggested very high

Page 451: Introduction to mineral exploration

434 C.J. MOON

No. pyrope/10 kg

Regional ice flow patterns

250–455 grains50–249 grains20–49 grains10–19 grains5–9 grains1–4 grains0 grains

kimberlite

3 (youngest)

2

1 (oldest)

Ranch Lakedispersal train

kilometers

FIG. 17.15 Abundance of Cr pyrope in 0.25–0.5 mm fraction of 10 kg till samples: Lac de Gras area. The dataare from the Geological Survey of Canada but are similar to that Diamet obtained from their regionalsampling. The broad halo around Lac de Gras is clear. The narrow dispersion train from a solitary pipe atRanch Lake to the north of the main esker is not detected. Some licence blocks are shown for comparison.DB, De Beers Canada Inc; Diavik, Diavik Diamond Mines and joint ventures. (Garnet data from McClenaghan& Kjarsgaard 2001.)

Page 452: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 435

FIG. 17.16 Point Lake Geophysics.Individual flight line across thePoint Lake pipe. (Courtesy Fugro.)

ALT

TAU

MAG

VG

EM20

EM25

EM06

FLIGHT DIRECTION

BA

diamond potential (Fig. 17.9). A minimumexpenditure of $US2M was established withBHP agreeing to fund exploration and develop-ment up to $C500M in exchange for 51% of anymine, Dia Met retained 29% and Fipke andBlusson each were to personally get 10% of anymine. The new joint venture engaged in stak-ing, in an attempt to protect its position andbuilt a buffer zone of 408,000 acres around theoriginal claim block (Fig. 17.15). Geophysicalmethods were then used to locate any pipes.Initially magnetic and EM methods were usedat the Point Lake prospect and confirmed thecircular anomaly (Fig. 17.16). This was testedin September 1991 and the first hole cutkimberlite from 455–920 ft. Caustic dissolu-tion of parts of the initial intersection returned81 small diamonds from 59 kg. The stock priceof Dia Met Minerals began to rise from $C0.68to $C1, and in order to forestall any accusa-tion of insider trading on November 7, 1991the following press release was made: “Thefollowing information has been released tothe Vancouver Stock Exchange, Canada: TheBHP–Dia Met diamond exploration jointventure, in Canada’s Northwest Territories,announces that corehole PL 91-1 at Point Lakeintersected kimberlite from 455 feet to the

end of the hole at 920 feet. A 59-kg sample ofthe kimberlite yielded 81 small diamonds, allmeasuring less than 2 mm in diameter. Someof the diamonds are gem quality.”

Although it was not clear whether there wasa deposit or that it was economic, the releasewas enough to trigger the greatest staking rushin Canadian history. By 1994, more than 40Macres had been staked. Initial large positionswere taken by De Beers (now exploring asMonopros) which took a land immediately tothe west of the BHP joint venture and a syndic-ate part financed by Aber Resources whichtook a position to the south (Fig. 17.17).

The BHP–Dia Met joint venture (eventuallyrenamed Ekati from the aboriginal name forLac de Gras) systematically explored the claimblocks using airborne geophysics and heavymineral sampling in an attempt to define fur-ther targets for drilling. At Point Lake theinitial diamond drilling was followed up bylarge diameter drilling in early 1992 to take a160-t sample for processing. The initial sampleconfirmed that the pipe was diamondiferous,101 ct diamonds were recovered of which25% were gems to give an average grade of56 cpht. An airborne geophysical survey inthe spring of 1992 located a number of pipes,

Page 453: Introduction to mineral exploration

436 C.J. MOON

One of the major problems that had dis-couraged potential explorers from the start wasthe pristine Arctic environment. The issue ofdevelopment therefore needed to be addressedand BHP started a $C14M environmental base-line study. Although the population of theNorthwest Territories is very low, the aborig-inal and Inuit inhabitants are highly dependenton the caribou and fishing. The claim blockcontains over 8000 lakes so water quality isimportant. Climatically the area is extreme;the average temperature is −11°C and wintertemperatures can go down to −70°C includ-ing wind chill (Fig. 17.18). After application fora mining review in 1995, a panel was estab-lished by the federal government and made a

including the higher grade Koala pipe. Thispipe, along with the Fox and Panda pipes, wasbulk sampled by drilling and underground de-velopment between 1993 and 1995 (Fig. 17.18).The underground sampling collected bulksamples of 3500–8200 t. The Panda pipe re-turned 3244 ct from 3402 t with an averageprice of $US130 ct−1 to give a in-ore value of$US124 t−1. Similarly the Koala pipe returned agrade of 95 cpht at a price of $US122 ct−1

($US116 t−1) from a 1550-t sample and the largeFox pipe 2199 ct from 8223 t at price of 125 ct−1

for $US34 t−1. By comparison with other de-posits worldwide (Table 17.3) it is clear thatthe Ekati pipes are significant, especially ifthey are mined as a group.

FIG. 17.17 Landsat image: Lac de Gras Area. The licence blocks and kimberlite locations are shown. (Ekatikimberlites from Nowicki et al. 2003, licence blocks and other kimberlites as at 2003.)

Page 454: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 437

series of recommendations that led to thegranting of conditional approval of the projectin August 1996. One of the major conditionswas the involvement of the three differentAboriginal groups and one Inuit group. Anothermajor recommendation was that commit-ments made by BHP were to be formalizedand that an independent monitoring agencyshould be established to assess the mine per-formance (Independent Environmental Mon-itoring Agency 1998). A significant issue wasthe impact of mining and processing on waterquality. The major problem was that tailingsfrom the Fox pipe would have a pH of 9.5and might impact on fish. By contrast, schistbedrock from the Misery pipe will be acidproducing and will need to be neutralized. Theonly major water quality problem was theimpact of sewage waste on the lake in which itwas disposed. Another concern was the impactof mining on caribou migration although the

dumps have been tailored to allow their freemovement.

Project construction was difficult as accessfor bulk items is restricted to shipment on a475-km-long ice road usable for 10 weeks inwinter. Any other items and personnel have tobe flown in from Yellowknife. The mine has aworkforce of 760 employees with accommoda-tion for a maximum of 600 people. Commercialproduction began in October 1998, only 1.5years after the beginning of construction. Theoverall mine plan included six pipes (Table17.3) with a further two at the permitting stage.The initial mining was an open pit on the highvalue Panda pipe. Subsequently mining startedat the Koala pit, the very small Koala North pit(Fig. 17.19), and the larger, but more distant,Misery open pit. Underground production isplanned from the Panda, Koala North, andKoala pipes, with the small Koala North de-posit as a test for underground mining. The Fox

FIG. 17.18 Winter drilling. (Courtesy Diavik Diamond Mines Inc.)

Page 455: Introduction to mineral exploration

438 C.J. MOON

17.2.2 Diavik

The most significant other project in north-ern Canada is at Diavik, immediately south ofthe Ekati buffer zone (Fig. 17.20). Fipke andBlusson were not alone in their interest in thearea (see Hart 2002). Another geologist, ChrisJennings, had been active in the Lac de Grasarea for a year before the discovery announce-ment at Ekati. This activity was based on hisdiamond exploration experiences in Botswanaand Canada with Superior and its then Cana-dian associate Falconbridge, and subsequentlyhe had also explored the indicator train east ofthe Blackwater Lake area with Selco. A geo-logist working on his instructions had alreadyfound indicators and 0.1-ct diamonds on theshore of Lac de Gras near, but outside, the DiaMet–BHP claim blocks. However, the companyfor whom he was working at the time was fo-cused on gold and not interested in diamonds.

FIG. 17.19 View of the Koala, Koala North (center), and Panda (nearest) pipes, Ekati in 2002. (Courtesy BHPBilliton Diamonds.)

and Beartooth pipes will be mined next fol-lowed by the Sable and Pigeon deposits. Totalresources at the end of the feasibility studywere estimated at 85 Mt with a grade of1.09 ct t−1 and a price of $US84 ct−1. In 2001BHP (by then BHP Billiton) bought Dia MetMinerals for $US436M and currently (2004)holds 80% of the Ekati joint venture.

By 2003 BHP had discovered 150 pipes on theEkati property (Nowicki et al. 2003). All pipeshave been core drilled, usually with a singlehole, and the core samples used for indicator ormicro-diamond analysis. These analyses arethen used to prioritize pipes for initial bulksampling by reverse circulation drilling. Typ-ically this produces samples of 50–200 t thatare processed on site in a dense media separa-tion plant. Diamonds were originally marketedpartly (35%) through the Diamond TradingCompany but are now sold totally in theAntwerp market.

Page 456: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 439

FIG. 17.20 Ekati Layout of the Ekati and Diavik Mines. (Courtesy Diavik Diamond Mines Inc.)

Beartooth Pit

Koala North PitKoala Pit

Fox Pit

Misery Pit

A154NA154S

A418

A21

BHP/Dia MetEkati Diamond Mine

Pigeon Pit

Panda Pit

DIAVIKDIAMONDS PROJECT

Road toSabe Pit

Dyke B

Dyke A

Dyke CDyke D

Outlet Dam

Airstrip

MiseryCamp

Diavik Campand Process Plant

Dyke

DykeAirstrip

OldCamp

Lac de Gras

Lac de Gras

Lac deGras

Lac duSauvage

Waste KimberliteMined RockCamp and Infrastructure AreaRoadDiavik Claim BoundaryOpen Pit

0 5km

N

100km0

Ice Road

Lac de GrasLac de Gras

Page 457: Introduction to mineral exploration

440 C.J. MOON

He was approached by a group of explorers ledby longtime Yellowknife prospector, GrenvilleThomas and Robert Gannicott to stake groundimmediately after the Dia Met–BHP press re-lease of November 1991. The syndicate (knownas Aber Resources) staked the claims in thewinter season. As Monopros had immediatelystaked the area adjoining Aber Resources’ firstchoice northeast corner of the Dia Met–BHPblock, the Aber syndicate started staking at itssecond choice area at the southeast corner.Aber Resources was, like Dia Met, dependenton larger companies to undertake the system-atic expensive exploration and agreed a jointventure with Kennecott Canada, a subsidiaryof Rio Tinto plc. Kennecott took an option onthe property with the agreement that it wouldearn 60% of the property by spending $C10million over 4 years with Jennings receiving a1% royalty for his work on the property. Thejoint venture flew the area with airborne geo-physics, magnetics, and EM, and began drill-ing the most prominent anomalies in 1992–93.Although kimberlite was intersected the microdiamond counts were not encouraging. Be-tween April 1992 and January 1994 20,500 linekilometers of airborne geophysics and 1700heavy mineral samples were taken. A total of25 kimberlites were discovered based on 65diamond drillholes totalling 6630 m.

Kennecott was also involved in otherprojects, notably the Tli Kwi Cho project, also ajoint venture with Aber Resources but withother junior company partners. At this loca-tion airborne geophysics located a complexpipe, which yielded 39 microdiamonds and 15macrodiamonds from 86 kg of core. Spurred onby the need to compete for exploration fundswith this DHK joint venture, the AberResources–Kenecott joint venture drilled loweramplitude magnetic targets under Lac de Graswhich had distinct indicator trails with G10garnets. Drilling of target A21 from the Lac deGras ice in April 1994 intersected kimberliteswith many indicator minerals. Initial resultsshowed the pipe was diamondiferous with 154stones from a 155-kg sample taken from theproperty. As the ice was beginning to melt onLac de Gras the rig was moved quickly northto test a further target A154 which appearedto be a twin pipe. The first hole intersectedkimberlite including a 3-ct diamond visible in

the core in May 1994. The initial core sample of750 kg contained 402 macro diamonds and 894micro diamonds. Of the macro diamonds sevenexceeded 0.2 ct in mass. Drilling from a bargein November 1994 Kennecott intersected theA154 north pipe and the A418 pipe, approxi-mately 800 m to the southwest, in the springof 1995. Delineation drilling of the A154 pipesbegan in 1995 using inclined 15.2-cm diamondholes from the shore of East Island and othersmall islands in 1995. As an example theA154S pipe was delineated by 39 contacts from6792 m in 26 holes (Thomas 2003). Mini bulksamples from large diameter core drilling pro-vided 56.5 t of kimberlite from which 255.6 ctwere recovered and valued at $US56.70 percarat or a phenomenal $US255 per tonne. Un-derground bulk sampling of the A154S andA418 pipes in 1996 and 1997 was next under-taken to generate large parcels of diamonds forsorting and valuing. The bulk sample of theA154S pipe returned 2585 dry tonnes ofkimberlite containing 12,763 ct and that of theA418 3350 t containing 8323 ct. These stoneswere valued both by Rio Tinto’s valuers inPerth, Australia and independently in Ant-werp. The cost of underground sampling anddeep delineation drilling as well as preliminaryenvironmental studies was $C23.5M. Testingof the A418 North and A21 pipes followed in1996 and 1997.

The very high revenues projected from thetwo high grade pipes led to commissioning of apre-feasibility study, which was completed inSeptember 1997. A project description was sub-mitted to the Federal government in March1998 triggering a formal environmental assess-ment. The feasibility study recommended min-ing the A154S and N pipes in one open pit to285 m depth for 10 years. Open pits were alsoplanned for the A418 and A21 pipes (Table17.4). Underground mining was consideredeconomic for the A154S and A418 pipes. Oneof the most important considerations was thatunlike Ekati, the pipes were underneath amajor albeit shallow lake. The lake would haveto be partially drained and dykes constructed(Fig. 17.21). The environmental review processwas similar to that undertaken for the Ekatidevelopment with a major commitment towin approval from the Aboriginal and Inuitinhabitants of the area. One hiccup was that

Page 458: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 441

TABLE 17.4 Ore reserves Diavik.(From Ellis Consulting 2000.) Pipe A-418 A-154S A-154N A-21 Total

Open pitTonnes (Mt) 4.2 9.3 2.7 3.7 19.9Grade (ct t−1) 3.79 4.69 2.89 2.89 3.92Carats (Mct) 17.0 43.6 7.8 10.8 78.2

UndergroundTonnes (Mt) 4.1 1.5 5.7Grade (ct t−1) 3.97 4.55 4.12Carats (Mct) 17.4 7 23.3

Total reserveTonnes (Mt) 8.3 10.8 2.7 3.7 25.6Grade (ct t−1) 3.88 4.67 2.89 2.89 3.96Carats (Mct) 32.4 50.6 7.8 10.8 101.5

ValuationPrice ($US ct−1) 53 59 33 36 53Value (US$M) 1715 2984 256 389 5344

FIG. 17.21 Aerial view of pre-stripping of the A154 pipes at Diavik in 2002. (Courtesy Diavik Diamond MinesInc.)

Page 459: Introduction to mineral exploration

442 C.J. MOON

TA

BLE

17.5

Pro

ject

ed c

ash

flow

, Dia

vik

in 1

998

$CM

. (Fr

om E

llis

Con

sult

ing

2000

.)

Yea

rC

apit

alR

esou

rce

Priv

ate

Ope

rati

ngC

losu

reC

apit

alR

esou

rce

Aft

er-t

axex

pend

itur

ein

com

ero

yalt

ies

cost

sco

sts

reco

very

profi

tsin

com

e

1995

−199

920

620

0012

320

0143

020

0243

020

0324

619

94

750

120

00

2004

039

88

125

026

60

020

050

569

1114

30

415

00

2006

7356

911

143

041

50

020

0773

569

1114

30

363

5228

2008

1556

911

143

015

400

215

2009

1556

911

143

015

400

215

2010

8955

111

143

089

308

165

2011

8953

911

143

089

296

159

2012

1552

110

143

015

353

189

2013

1541

68

143

915

241

129

2014

1540

18

143

915

226

121

2015

6032

97

143

960

110

5920

1615

323

616

89

1512

467

2017

1531

26

140

915

141

7620

1815

422

815

29

1523

612

720

1915

417

815

29

1523

212

520

2015

353

713

19

1519

110

320

210

233

589

90

130

7020

220

197

475

90

108

58

Tot

al19

6884

5616

926

7594

1968

3550

1905

Page 460: Introduction to mineral exploration

17: EKATI AND DIAVIK DIAMOND MINES 443

construction in 2000 required the use of the iceroad to move materials from Yellowknife.However the agreement was only signed inMarch and the key land use permit issued,causing some delay in construction. The totalcost of the exploration and development pro-cess, including feasibility studies and publicreview process, was $C206M.

As part of the review Diavik Diamond MinesInc. has released an independent assessment ofthe resource revenue shown in Table 17.5 (EllisConsulting 2000). Reserves are estimated to bea total of 25.6 Mt at 396 cpht with an averageprice of $US53 ct−1 (Table 17.4). The high gradeA154-S pipe is being mined first optimizingcash flow with underground mining scheduledfor 2008. The overall cost of construction wasestimated at $C1286M and production begansuccessfully in January 2003, $C50 millionunder budget and 6 months ahead of schedule.

Since the Ekati and Diavik discoveries, threeother projects have been discovered that appearto have potential to become mines at SnapLake, Gahcho Kue, and Jericho in the SlaveProvince. There have also been a number ofless successful ventures reminding investors

of the difficult process of proving a viabledeposit. At Tli Kwi Cho, Kennecott elected tomove directly from the delineation stage tobulk sampling and spent $C15M extracting anunderground bulk sample in 1994, at the sametime as the Diavik pipes were discovered.Although the pipe returned reasonable gradesat 35.9 cpht the quality was poor with less than30% gems and the deposit was clearly not ofeconomic interest.

17.3 SUMMARY

The discovery of the Ekati and Diavik was theresult of perseverance by a small companyaided eventually by major company invest-ment. The technology of regional sampling forindicator minerals was well known, althoughthe detailed chemical investigation of heavyminerals grains required refinement and in-vestment in technology. Although the targetswere small, many were easily detectable by air-borne geophysics. The remoteness of the areadid not in the end prove to be an obstacle as thedeposits were high grade with a ready market.

Page 461: Introduction to mineral exploration

444 INTERNET LINKS

Internet links

The following are some of the more useful internet links by chapter and current at 2005.

CHAPTER 1:

Mining Companies

Index of companies with links http://www.goldsheetlinks.com

Mining Information

Infomine http://www.infomine.comMining Journal http://www.mining-journal.comMineweb http://www.mineweb.netNorthern Miner http://www.northernminer.comReflections http://www.reflections.com.au/miningandexplorationBreaking New Ground: Mines And Minerals Sustainable Development Project. http://www.iied.org/mmsd/finalreport/index.html

CHAPTER 2: MINERALOGY

Athena mineralogy database – includes search by elements http://un2sg4.unige.ch/athena/mineral/mineral.htmlBarthelmy’s Mineralogy database http://webmineral.com/Mineralogy Database http://mindat.orgRob Ixer’s Virtual Atlas of Opaque and Ore Minerals http://www.smenet.org/opaque-ore/

CHAPTER 3: DEPOSIT MODELS (ON LINE)

British Columbia http://www.em.gov.bc.ca/Mining/Geolsurv/MetallicMineralsMineralDepositProfiles/USGS http://minerals.cr.usgs.gov/team/depmod.html

Page 462: Introduction to mineral exploration

INTERNET LINKS 445

CHAPTER 6: REMOTE SENSING

Data: Landsat, Global Land Cover (free) http://glcf.umiacs.umd.edu/data/Landsat Mr. Sid format coverage (free) https://zulu.ssc.nasa.gov/mrsid/ASTER http://asterweb.jpl.nasa.gov/Topography (90 m resolution) http://www2.jpl.nasa.gov/srtm/

CHAPTER 7: GEOPHYSICS

Fugro http://www.fugro.nl/geoscience/airborne/minexpl.asp

CHAPTER 8: GEOCHEMISTRY

Association of Applied Geochemists http://www.appliedgeochemists.org/FOREGS Geochemistry http://www.gsf.fi/foregs/geochem/

CHAPTER 9: DATA MANAGEMENT-SOME SOFTWARE PROVIDERS

Acquire http://www.acquire.com.au/Arcview/ArcGIS http://www.esri.com/Datamine http://www.datamine.co.uk/Geosoft http://www.geosoft.com/Interdex/Surpac http://www.surpac.com/Mapinfo http://www.mapinfo.com/Micromine http://www.micromine.com/Vulcan http://www.vulcan3d.com/

CHAPTER 10: SAMPLING AND EVALUATION

Boart drilling http://www.boartlongyear.com/

Web page on Mine Sampling

http://pws.prserv.net/Goodsampling.com/sampling.html

CHAPTER 12: CLIFFE HILL

Midland Quarry Products http://www.mqp.co.uk/

CHAPTER 14: WITWATERSRAND

Western Areas http://www.westernareas.co.za/Placer Dome http://www.placerdome.com/operations/southdeep/southdeep.html

Page 463: Introduction to mineral exploration

446 INTERNET LINKS

CHAPTER 15: KIDD CREEK

Kidd Creek Mine http://www.falconbridge.com/our_business/copper_kidd_creek.htmlKidd Creek Monograph http://www.nrcan.gc.ca/gsc/mrd/projects/kiddcreek monograph10_e.html

CHAPTER 17: DIAMONDS

De Beers http://www.debeersgroup.com/Ekati/BHP Diamonds http://ekati.bhpbilliton.com/Diavik http://www.diavik.ca/NWT Diamond site http://www.gov.nt.ca/RWED/diamond/

Page 464: Introduction to mineral exploration

REFERENCES 447

References

Abrams M.J., Conel J.E. & Lang H.R. (1984) JointNASA/GEOSAT test case report. Am. Assoc.Petroleum Geol., 3 vols.

Abrams M. & Hook S. (2002) Aster UsersHandbook, Version 2. Jet Propulsion Laboratory,Pasadena, USA. http://asterweb.jpl.nasa.gov/documents/documents.htm.

AcQuire (2004) Acquire Corporate Website. http://www.acquire.com.au.

Adams S.S. (1985) Using geological informationto develop exploration strategies for epithermaldeposits. In Berger B.R. & Bethke P.M. (eds),Geology and Geochemistry of Epithermal Sys-tems, 273–298. Economic Geology Publishing Co.,Littleton, Colorado.

AGSO (Australian Geological Survey Organisation)(1998) Geology and mineral potential of majorAustralian mineral provinces. AGSO J Aust. Geol.Geophys., 17(3), 260; (4), 320.

Akçay M., Özkan H.M., Moon C.J. & Scott B.C.(1996) Secondary dispersion from gold depositsin West Turkey. J. Geochem. Explor., 56, 197–218.

Ali J.W. & Hill I.A. (1991) Reflection seismics forshallow geological investigations: a case studyfrom central England. J. Geol. Soc. London, 148,219–222.

Allen M.E.T. & Nichol I. (1984) Heavy-mineral con-centrates from rocks in exploration for massivesulphide deposits. J. Geochem. Explor., 21, 149–165.

Allum J.A.E. (1966) Photogeology and RegionalMapping. Pergamon Press, Oxford.

Andrew R.L. (1978) Gossan Evaluation. ShortCourse, Rhodes University, Grahamstown.

Annels A.E. (1991) Mineral Deposit Evaluation: apractical approach. Chapman and Hall, London.

Anon. (1977) Consolidated Rambler finds zinc bear-ing mercury. World Mining, March, 79.

Anthony M. (1988) Technical change – the impend-ing revolution. In Australasian Institute of Miningand Metallurgy Sydney Branch, Minerals and Ex-ploration at the Crossroads, 19–25, Sydney.

Arcatlas (1996) Arcatlas CD-ROM. RussianAcademy of Sciences & ESRI, ESRI, Redlands,California, USA.

Arc-SDM (2001) Spatial Data Modeller for Arcview.Natural Resources Canada. http://cgpd.cgkn.net/sdm/default_e.htm.

Ashleman J.C. (1983) Trinity Silver Project 1982–1983, Progress Summary Report. Unpublished Re-port EX84–10, US Borax.

Ashleman J.C. (1986) Trinity Venture Project, 1986Summary Report. Unpublished Report, EX 87–3,US Borax.

Ashleman J.C. (1988) The Trinity Silver Deposit,Pershing County, Nevada. Unpublished FieldGuide.

Atkinson W.J., Hughes F.E. & Smith C.B. (1984)A review of the kimberlitic rocks of WesternAustralia. In Kornprobst J. (ed.) Kimberlites I:kimberlites and related rocks, 195–224. Elsevier,Amsterdam.

Australian Drilling Industry Training Committee(1997) Drilling: the manual of methods, applica-tions, and management. CRC Press, Boca Raton.

Austr. Inst. Geoscientists (1987) MeaningfulSampling in Gold Exploration, bull. 7. Duncanand Associates, Leederville.

Austr. Inst. Geoscientists (1988) Sample Preparationand Analyses for Gold and Platinum Group Ele-ments, bull. 8, Duncan and Assoc., Leederville.

Australasian Joint Ore Reserves Committee (2003)The JORC Code. http://www.jorc.org/main.php.Australasian Institute of Mining and Metallurgy,Carlton, Victoria.

Baele S.M. (1987) Silver Ore Reserve Estimation,Trinity Heap Leach Project, Pershing County,

Page 465: Introduction to mineral exploration

448 REFERENCES

Nevada. Unpublished Report, MD 87–3, USBorax.

Bagby W.C. & Berger B.R. (1985) Geologic character-istics of sediment-hosted, disseminated precious-metal deposits in the western United States. InBerger B.R. & Bethke P.M. (eds), Geology andGeochemistry of Epithermal Systems, 169–202.Reviews in Economic Geology, vol. 2. PublishingCo., Littleton, Colorado.

Bailey D.C. & Hodson D.I. (1991) The Neves-CorvoProject – the successes and challenges. Trans.Instn Min. Metall., Sect. A, 100, A1–A10.

Baldock J.W., Hepworth J.V. & Marengwa B.S. (1976)Gold, basemetals, and diamonds in Botswana.Econ. Geol., 71, 139–156.

Ball Aerospace & Technologies Corp. (2002)Quickbird. Boulder, Colorado. http://www.ball.com/aerospace/quickbird.html.

Barnes E. (1997) MINDEV 97, The Third Interna-tional Conference on Mine Project Development.Australasian Institute of Mining and Metallurgy,Carlton, Victoria, 280 pp.

Barnes J.W. & Lisle R.J. (2003) Basic Geological Map-ping, 4th edn. Wiley, New York.

Barnicoat A.C., Henderson I.H.C., Knipe R.J., et al.(1997) Hydrothermal gold mineralization in theWitwatersrand basin. Nature, 386, 820–824.

Barrett W.L. (1987) Auger drilling. Trans. Instn Min.Metall., Sect. B, 96, B165–169.

Barrett W.L. (1992) A case history of pre-extractionsite investigation and quarry design, Cliffe HillQuarry, Leicestershire. In Annels A.E. (ed.) CaseHistories and Methods in Mineral ResourceEvaluation, 69–76, spec. publ. 63. GeologicalSociety, London.

Barrie C.T. & Hannington M.T. (1999) Classificationof volcanic-associated massive sulfide depositsbased on host-rock composition In Barrie C.T.& Hannington M.T. (eds) Volcanic AssociatedMassive Sulfide Deposits and Examples in Mod-ern And Ancient Settings, 1–12. Reviews inEconomic Geology, vol. 8. Economic Geology Pub-lishing Co., Littleton, Colorado.

Barrie C.T., Ludden J.N. & Green T.H. (1993)Geochemistry of volcanic rocks associated withCu-Zn and Ni-Cu deposits in the Abitibi Sub-province. Economic Geology, 88, 1341–1358.

Barrie C.T., Hannington M.T. & Bleeker W. (1999)The giant Kidd Creek volcanic-associated mas-sive sulphide deposit, Abitibi subprovince,Canada. In Barrie C.T. & Hannington M.T. (eds)Volcanic Associated Massive Sulfide Depositsand Examples in Modern And Ancient Settings,247–260. Reviews in Economic Geology, vol. 8.

Economic Geology Publishing Co., Littleton,Colorado.

Bartlett P.J. (1994) Geology of the Premier Mine. InAnnhaeuser C.R. (ed.) Proceedings of the XVthCMMI Congress, 201–214. South African I.M.M.,Johannesburg.

Barton J.M., Barnett W.P., Barton E.S., et al. (2003)The geology of the area surrounding the Venetiakimberlite pipes, Limpopo Belt, South Africa:a complex interplay of nappe tectonics andgranitoid magmatism. S. Afr. J. Geol., 106, 109–128.

Barton P.B. (1986) Commodity/Geochemical Index.In Cox D.P. & Singer D. (1986) Mineral DepositModels. US Geological Survey Bulletin 1693,Appendix C, 15 pp. Reston, VA.

Bateman A.M. (1950) Economic Mineral Deposits.Wiley, New York.

BCGS (British Columbia Geological Survey) (2004)British Columbia Mineral Deposit ProfilesVersion 2.1. http://www.em.gov.bc.ca/Mining/Geolsurv/MetallicMinerals/MineralDeposit-Profiles/default.htm.

Beaty D.W., Taylor H.P. Jr & Coad P.R. (1988) Anoxygen isotope study of the Kidd Creek, Ontario,volcanogenic massive sulfide deposit: evidence fora high 18O ore fluid. Econ. Geol., 83, 1–17.

Bell A.R. & Hopkins D.A. (1988) From farmlandto tarmac. Extractive Industry Geology 1985, 19–26.

Bell T. (1989) Ore Reconciliation and StatisticalEvaluation of the Trinity Silver Mine, PershingCounty, Nevada. Unpublished MSc Dissertation,University of Leicester.

Bell T. & Whateley M.K.G. (1994) Evaluation ofgrade estimation techniques. In Whateley M.K.G.& Harvey P.K. (eds), Mineral Resource EvaluationII, Methods and Case Histories, spec. publ. 79,Geol. Soc. London.

Bennett M.J., Beer K.E., Jones R.C., Turton K., RollinK.E., Tombs J.M.C. & Patrick D.J. (1981) MineralInvestigations Near Bodmin, Cornwall. Part3 – The Mulberry and Wheal Prosper Area. BritishGeological Survey Mineral ReconnaissanceReport 48, Keyworth.

Berger B.R. & Bagby W.C. (1991) The geology andorigin of Carlin-type gold deposits. In Foster R.P.(ed.) Gold Metallogeny and Exploration, 210–48.Blackie, Glasgow.

Berkman D.A. (1989 & 2001) Field Geologist’sManual, 3rd and 4th Editions. Australasian In-stutute of Mining and Metallurgy, Parkville.

Berry L.G. (1941) Geology of the Bigwater Lake area.Ontario Department of Mines Annual Report, 48,

Page 466: Introduction to mineral exploration

REFERENCES 449

1–9 including map 48n, Bigwater Lake area, 1 inchto 1 mile, Toronto.

Berry L.G., Mason B. & Dietrich R.V. (1983) Minera-logy. Freeman, San Francisco.

Beus A.A. & Grigorian S.V. (1977) GeochemicalMethods for Mineral Deposits. Applied Publish-ing, Wilmette, lllinois.

BHP Billiton (2003) BHP Billiton Website. http://www.bhpbilliton.com.

Bieniawski Z.T. (1976) Rock mass classification inrock engineering. In Bieniawski Z.T. (ed.), Ex-ploration for Rock Engineering, vol. 1, 97–106.A.A Balkema, Cape Town.

Bishop J.R. & Lewis R.J.G. (eds) (1996) DHEM SpecialVolume. Explor. Geophys., 27, 37–177.

Blain C.F. (2000) Fifty-year trends in minerals dis-covery – commodity and ore-type targets. Explor.Mining Geol., 9, 1–11.

Blanchard R. (1968) Interpretation of Leached Out-crops. Nevada Bureau of Mines Bull. 66, Reno.

Blanchard R. & Boswell P.F. (1934) Additionallimonite types of galena and sphalerite derivation.Econ. Geol., 29, 671–690.

Bleeker W. & Hester B. (1999) Discovery of the KiddCreek massive sulphide orebody: a historical per-spective. In Hannington M.D. & Barrie C.T. (eds)The Giant Kidd Creek Volcanogenic MassiveSulfide Deposit, Western Abitibi Subprovince,Canada. Economic Geology Monograph 10, 31–42. Economic Geology Publishing Co., Littleton,Colorado.

Bleeker W. & Parrish R.R. (1996) Stratigraphy andU-Pb zircon geochronology of Kidd Creek: impli-cations for the formation of giant volcanogenicmassive sulphide deposits and the tectonic historyof the Abitibi greenstone belt. Can. J. Earth Sci.,33, 1213–1231.

Bloom L. (ed.) (2001) Writing Geochemical Reports,2nd edn. Association of Exploration Geochem-ists, spec. publ. 15. Nepean, Ontario, Canada.

Bonham H.F. (1988) Silver deposits of Nevada.In Silver: exploration, mining and treatment,25–31. Institution of Mining and Metallurgy,London.

Bonham H.F. Jr (1989) Bulk mineable gold deposits ofthe Western United States In Keays R.R., RamsayW.R.H. & Groves D.I. (eds) The Geology of GoldDeposits – the perspective in 1988, 193–207. Eco-nomic Geology Monograph 6. El Paso.

Bonham-Carter G.F. (1994) Geographic InformationSystems for Geoscientists: modelling with GIS.Pergamon, New York.

Bonham-Carter C.F., Agterberg F.P. & Wright D.F.(1988) Integration of geological datasets for gold

exploration in Nova Scotia. PhotogrammetricEngineering Remote Sens., 54, 1585–1592.

Boyd F.R. & Gurney J.J. (1986) Diamonds and theAfrican lithosphere. Science, 232, 472–477.

Boyd F.R. & Nixon P.H. (1978) Ultramafic nodulesfrom the Kimberley pipes, South Africa. Geochim.Cosmochim. Acta, 42, 1367–1371.

Bradshaw P.M.D. (ed.) (1975) Conceptual models inexploration geochemistry – the Canadian Shieldand Canadian Cordillera. J. Geochem. Explor., 4,1–213.

Brady B.H.G. & Brown E.T. (1985) Rock Mechanicsfor Underground Mining. George Allen andUnwin, London.

Brassington R. (1988) Field Hydrogeology. Geolo-gical Society of London Professional HandbookSeries, Wiley, Chichester.

Brinkmann R. (1976) Geology of Turkey. Elsevier,Amsterdam.

Brisbin D., Kelly V. & Cook R. (1991) Kidd CreekMine. In Fyon J.A. & Green A.H. (eds), Geologyand Ore Deposits of the Timmins District,Ontario, 66–71. Geological Survey of CanadaOpen File Report 2161.

British Standard 812 (1975) Method for Samplingand Testing of Mineral Aggregates, Sands and Fill-ers. British Standards Institution, London.

Brook Hunt (2004) Brook Hunt website.www.brookhunt.com.

Brooks R.R. (1983) Biological Methods of Prospect-ing for Minerals. Wiley, New York.

Brown E.T. (ed.) (1981) Rock Characterization,Testing and Monitoring. International Society forRock Mechanics, Pergamon Press, Oxford.

Brown W.M., Gedeon T.D., Groves D.I. & BarnesR.G. (2000) Artificial neural networks: a newmethod for mineral prospectivity mapping.Australian J. Earth Sci., 47, 757–770.

Brummer J.J. (1978) Diamonds in Canda. C.I.M.Bull., 71, 64–70.

Brundtland G.H. (1987) Our Common Future. WorldCommission on Environment and Development,Oxford University Press, Oxford.

Burke K., Kidd W.S.F. & Kusky T.M. (1986) Archaeanforeland basin tectonics in the Witwatersrand,South Africa. Tectonics, 5, 439–456.

Bustillo R.M. & Lopez J.C. (1997) Manual deEvaluacion y Diseño de Explotaciones Mineras.Entorno Grafico, Madrid.

Butt C.R.M. & Smith R.E. (1980) Conceptual modelsin exploration geochemistry – Australia. J.Geochem. Explor., 12, 89–365.

Butt C.R.M. & Zeegers H. (eds) (1992) RegolithExploration Geochemistry in Tropical and

Page 467: Introduction to mineral exploration

450 REFERENCES

Subtropical Terrains. Handbook of ExplorationGeochemistry, vol. 4. Elsevier, Amsterdam.

Cady J.W. (1980) Calculation of gravity and magneticanomalies of finite length right polygonal prisms.Geophysics, 45, 1507–1512.

Cameron E.M., Hamilton S.M., Leybourne M.I.,Hall G.E.M. & McClenaghan M.B. (2004) Find-ing deeply buried deposits using geochemistry.Geochem. Explor. Environ. Analysis, 4, 7–32.

Cameron R.I. & Middlemis H. (1994) Computermodelling of dewatering a major open pit mine:case study from Nevada, USA. In WhateleyM.K.G. & Harvey P.K. (eds), Mineral ResourceEvaluation II: methods and case histories, 205–215, spec. publ. 79. Geological Society, London.

Camisani-Calzolari F.A.G.M., Ainslie L.C. & vander Merwe P.J. (1985) Uranium in South Africa1985. South African Atomic Energy Corporation ofSouth Africa, Pelindaba.

Camm T.W. (1991) Simplified Cost Models forPrefeasibility Mineral Evaluations. Bureau ofMines Information Circular 9298. United StatesDepartment of the Interior, Washington DC.

Campbell A.S. (ed.) (1971) Geology and history ofTurkey. Petroleum Exploration Society of Lybia,13th Annual Conference.

Campbell G. & Crotty J.H. (1990) 3-D seismic map-ping for mine planning purposes at the South Deepprospect. In Ross-Watt R.A.J. & Robinson P.D.K.(eds), Technical Challenges in Deep Level Mining,569–597. South African Institute of Mining andMetallurgy, Johannesburg.

Campbell I.H., Lesher C.M., Coad P., Franklin J.M.,Gorton M.P. & Thurston P.C. (1984) Rare-earth mobility in alteration pipes below Cu-Zn-sulphide deposits. Chem. Geol., 45, 181–202.

Cartwright A.P. (1967) Gold Paved the Way. Purnell,Johannesburg.

Carver R.N., Chenoweth L.M., Mazzucchelli R.H.,Oates C.J. & Robbins T.W. (1987) Lag – ageochemical; sampling medium for arid regions. J.Geochem. Explor., 28, 183–199.

Cas R.A.F. & Wright J. (1987) Volcanic Successions:modern and ancient. Allen & Unwin, London.

Cathles L.M. (1993) Oxygen isotope alteration inthe Noranda Mining District, Abitibi GreenstoneBelt, Quebec. Econ. Geol., 88, 1483–1511.

Chadwick J. 2002. Northparkes. Mining Magazine,187, 8–14.

Chaussier J-B. & Morer J. (1987) Mineral ProspectingManual. Elsevier, Amsterdam.

Chen W. (1988) Mesozoic and Cenozoic sandstone-hosted copper deposits in South China. Mineral.Deposita, 23, 262–267.

Chinn C.T. & Ascough G.L. (1997) Mineral poten-tial mapping using an expert system and GIS.In Gubins A.G. (ed.) (1997) Geophysics andGeochemistry at the Millenium: Proceedings ofExploration 97, 89–86. Prospectors & DevelopersAssociation of Canada, Toronto.

Christison Scientific Equipment Catalogue (2002)Particle Size Reduction. http://www.christison.com/CAT7.html. Gateshead, UK.

Clark I. (1979) Practical Geostatistics. ElsevierApplied Science, Amsterdam.

Clark R.J. McH., Homeniuk L.A. & Bonnar R. (1982)Uranium geology in the Athabaska and a compari-son with other Canadian Proterozoic Basins. CIMBull., 75(April), 91–98.

Clarke P.R. (1974) The Ecstall Story: introduction.Can. Inst. Min. Metall. Bull., 77, 51–55.

Clifford J.A., Meldrum A.H., Parker, R.H.T. & EarlsG. (1990) 1980–1990: a decade of gold explorationin Northern Ireland and Scotland. Trans. Inst.Min. Metall., Sect. B, 99, B133–138.

Clifford J.A., Earls G., Meldrum A.H. & Moore N.(1992) Gold in the Sperrin mountains, NorthernIreland: an exploration case history In BowdenA.A., Earls G., O’Connor P.G. & Pyne J.F. (eds) TheIrish Minerals Industry 1980–1990, 77–87. IrishAssociation of Economic Geology, Dublin.

Clement C.R., Harris J.W., Robinson D.N. &Hawthorne J.B. (1986) The De Beers kimberlitepipe – a historic South African diamond mine.In Anhausser C.R. & Maske S. (eds), MineralDeposits of South Africa, 2193–2214. GeologicalSociety of South Africa, Johannesburg.

Clifford T.N. (1966) Tectono-metallogenic unitsand metallogenic provinces of Africa. Earth Plan.Sci Lett., 1, 421–434.

Clifton H.E., Hunter R.E., Swanson F.J. & PhillipsR.L. (1969) Sample Size and Meaningful GoldAnalysis. US Geol. Surv. Prof. Paper 625-C.

Closs L.G. & Nichol I. (1989) Design and planningof geochemical programs. In Garland G.D. (ed.)Proceedings of Exploration ‘87, 569–583, spec. vol.3. Ontario Geological Survey, Toronto.

Coad P.R. (1984) Kidd Creek Mine. In Gibson H.L.et al. (eds) Surface Geology and Volcanogenic BaseMetal Massive Sulphide Deposits and Gold De-posits of Noranda and Timmins. Geological Asso-ciation of Canada Field Trip Guide Book14.

Coad P.R. (1985) Rhyolite geology at Kidd Creek – aprogress report. Can. Inst. Min. Metall. Bull., 78,70–83.

Cohen A.D., Spackman W. & Raymond R. (1987)Interpreting the characteristics of coal seams from

Page 468: Introduction to mineral exploration

REFERENCES 451

chemical, physical and petrographic studies ofpeat deposits. In Scott A.C. (ed.), Coal and Coal-bearing Strata: recent advances, 107–125, spec.publ. 32. Geological Society, London.

Coker W.B., Hornbrook E.H.W. & Cameron E.M.(1979) Lake sediment geochemistry applied tomineral exploration. In Hood P.J. (ed.), Geophysicsand Geochemistry in the Search for Metallic Ores,435–447. Econ. Geol. Rept 31. Geological SurveyCanada.

Colman T.B. (2000) Exploration for Metalliferousand Related Minerals: a guide, 2nd edn. BritishGeological Survey, Keyworth.

Cook D.R. (1987) A crisis for economic geologistsand the future of the Society. Econ. Geol., 82, 792–804.

Coope J.A. (1991) Monitoring Laboratory Perform-ance with Standard Reference Samples. Abstract,15th Geochemical Exploration Symposium,Reno.

Cooper P.C. & Sternberg M. (1988) Deep directionalcore drilling with Navi-Drill. Engrg. Min. J., 189(7),48–49, 59.

Corbett J. (1990) Overview of geophysical methodsapplied to gold exploration in Nevada. LeadingEdge, 9, 17–26.

Corner B. & Wilshire W.A. (1989) Structure of theWitwatersrand Basin derived from the interpreta-tion of aeromagnetic and gravity data. In GarlandG.D. (ed.), Proceedings of Exploration ’87, 532–545, spec. vol. 3. Ontario Geological Survey.

Cox D.P. & Singer D. (1986) Mineral DepositModels. US Geol. Surv. Bull. 1693. Reston.

Cox F.C., Bridge D.McC. & Hull J.H. (1977)Procedure for the Assessment of LimestoneResources. Mineral Assessment Report 30. Insti-tute of Geological Sciences, London.

Craig J.R. & Vaughan D.J. (1994) Ore Microscopy andOre Petrography, 2nd edn. Wiley, New York.

Craig Smith R. (1992) PREVAL: Prefeasibility Soft-ware Program for Evaluating Mineral Properties.Information Circular 9307. United States Depart-ment of Interior, Washington DC.

Crosson C.C. (1984) Evolutionary development ofPalabora. Trans. Instn Min. Metall., Sect. A, 93,A58–A69.

Crowson P.C.F. (1988) A perspective on worldwideexploration for minerals. In Tilton J.E., EggertR.G. & Landsberg H.H. (eds), World MineralExploration, 21–103. Resources for the Future,Washington.

Crowson P. (1998) Inside Mining: the economics ofthe supply and demand of minerals and metals.Mining Journal Books, London.

Crowson P. (2003) Astride Mining: issues andpolicies for the mining industry. Mining JournalBooks, London.

CRU (2004) Commodities Research Unit website.http://www.crugroup.com

Cummings J.D. & Wickland A.P. (1985) DiamondDrill Handbook. J.K. Smit and Sons, Toronto.

Curtin G.C., King H.D. & Moiser E.L. (1974)Movement of elements into the atmospherefrom coniferous trees in the subalpine forestsof Colorado and Idaho. J. Geochem. Explor., 3,245–263.

Danchin P.D. (1989) Obituary of J. Dalrymple. Geo-bulletin 2nd Quarter, 21–22.

Davenport P.H., Friske P.W.B. & DiLabio R.N.W.(1997) The application of lake sediment geochem-istry to mineral exploration: recent advancesand examples from Canada. In Gubins, A.G. (ed.)Geophysics and Geochemistry at The Millenium:Proceedings of Exploration 97, 249–260. Pro-spectors & Developers Association of Canada,Toronto.

David M. (1988) Handbook of Applied AdvancedGeostatistical Ore Reserve Estimation. Devel-opments in Geomathematics, 6. Elsevier,Amsterdam.

Davidson C.F. (1965) The mode of origin of banketorebodies. Trans. Instn Min. Metall., 74, 319–338.

Davies M.H. (1987) Prospects for copper. Bull.Institn Min. Metall., September, 3–6.

Davies P. (2000) Discovery of Marsfontein M-1 pipe.http://www.msaprojects.co.za/discm1pipe.htm.

Davis J.C. (1986, 2003) Statistics and Data Analysisin Geology. Wiley, New York.

Dawson C.W. & Tokle V. (1999) Directional CoreDrilling: the state of the art. In Hilton D.E. &Samuelson K. (eds) Proceedings of the 1999 RapidExcavation and Tunnelling Conference, 351–362.Society for Mining, Metallurgy and Exploration,Littleton, CO.

Dawson J.B. & Stephens W.E. (1975) Statisticalclassification of garnets from kimberlite andassociated xenoliths. J. Geol., 83, 589–607.

De Beer J.H. & Eglington B.M. (1991) Archaeansedimentation on the Kaapvaal Craton in relationto tectonism in granite-greenstone terrains: geo-physical and geochronological constraints. J. Afr.Earth Sci., 13, 27–44.

De Beers (2001) Technical and Financial Report(background to offer by DBI). http://www.angloamerican.co.uk.

De Beers (2003a) De Beers Exploration and Opera-tions Overview. www.angloamerican.co.uk.

Page 469: Introduction to mineral exploration

452 REFERENCES

De Beers (2003b) Venetia Analysts Visit. www.angloamerican.co.uk.

Decisioneering (2003), Crystal Ball. http://www. de-cisioneering.com/monte-carlo-simulation.html.

Dentith M.C., Frankcombe K.F., Ho S.E., ShepherdJ.M., Groves D.I. & Trench A. (eds) (1995)Geophysical Signatures of Western AustralianMineral Deposits. Geology and GeophysicsDepartment (Key Centre) & UWA Extension, Uni-versity of Western Australia, Pubn 26, and Aus-tralian Society of Exploration Geophysicists, spec.publ. 7, 454 pp.

Department of Land Information (2003) Governmentof Western Australia, Perth. http://www.dola.wa.gov.au/.

D’Ercole C. Groves D.I. & Knox-Robinson C.M.(2000) Using fuzzy logic in a geographic informa-tion system environment to enhance conceptuallybased prospectivity analysis of Mississippi Valley-type mineralization. Aust. J. Earth Sci., 47, 913–927.

Deutsch C.V. (2002) Geostatistical Reservoir Model-ling. Oxford University Press, New York.

Deutsch C.V. & Journel A.G. (1998) GSLIB, Geo-statistical Software Library and User’s Guide, 2ndedn. Oxford University Press.

Deutsch M., Wiesnet D.R. & Rango A. (eds)(1981) Satellite Hydrogeology. American WaterResources Association, Minneapolis, MN.

Diamond Facts (2003) North West TerritoriesGovernment Web Site. http://www.gov.nt.ca/RWED/diamond/.

Diamond Geology (2003) De Beers Group Website.http://www.debeersgroup.com/exploration.

Dickinson R.T., Jones C.M. & Wagstaff J.D. (1986)MWDatanet – acquisition and processing of down-hole and surface data during drilling. Trans. InstnMin. Metall., Sect. B, 95, B140–148.

DigitalGlobe (2004) QuickBird Specifications.http://www.digitalglobe.com/products/quickbird.shtml, Longmont, CO.

Dines H.G. (1956) The Metalliferous Mining Regionof South-west England. HMSO, London.

Donohoo H.V., Podolsky G. & Clayton R.H. (1970)Early geophysical exploration at Kidd Creek Mine.Mining Congress Journal, May, 44–53.

Dorey R., Van Zyl D. & Kiel J. (1988) Overviewof heap leaching technology. In Van Zyl D.,Hutchison I. & Kiel J. (eds), Introduction to Evalu-ation, Design and Operation of Precious MetalHeap Leaching Projects, 3–22. Society of MiningEngineers, Denver, CO.

Dozy J.J., Erdman D.A., Jong W.J., Krol G.L. &.Schouten C. (1939) Geological results of the

Carstenz Expedition 1936. Leidsche GeologischeMededeelingen, 11, 68–131.

Drury S.A. (1986) Remote sensing of geologicalstructure in European agricultural terrains. Geol.Mag., 123, 113–121.

Drury S.A. (2001) Image Interpretation in Geology,3rd edn. Allen & Unwin, London.

Du Toit A.L. (1954) The Geology of South Africa.Oliver and Boyd, Edinburgh.

Dunn C.E. (2001) Biogeochemical methods in theCanadian Shield and Cordillera. In McClenaghanM.B., Bobrowsky P.T., Hall G.E.M. & Cook S.J.(eds) Drift Exploration in Glacial Terrain, 151–164, spec. publ. 185. Geological Society, London.

Dunn P.R., Battey G.C., Miezitis Y. & McKay A.D.(1990a) The distribution and occurrence ofuranium. In Glasson K.R. & Rattigan J.H. (eds),Geological Aspects of the Discovery of SomeImportant Mineral Deposits in Australia, 455–462. Australasian Institution of Mining andMetallurgy, Melbourne.

Dunn P.R., Battey G.C., Miezitis Y. & McKay A.D.(1990b) The uranium deposits of the North-ern Territory. In Glasson K.R. & Rattigan J.H.(eds), Geological Aspects of the Discovery of SomeImportant Mineral Deposits in Australia, 463–476. Australasian Institute of Mining andMetallurgy, Melbourne.

E3 (2004) Environmental Excellence in Explora-tion (Prospectors & Developers Association ofCanada) Web Site. http://www.e3mining.com.

Eaton D.W., Milkereit B. & Salisbury M. (2003)Seismic methods for deep mineral exploration:mature technologies adapted to new targets. Lead-ing Edge, 22, 580–585.

Eckstrand O.R. (ed.) (1984) Canadian MineralDeposit Types: a geological synopsis. Geol. Surv.Canada Econ. Rept 36. Ottawa.

Eckstrand O.R., Sinclair W.D. & Thorpe R.I. (eds)(1995) Geology of Canadian Mineral DepositTypes. Geology of Canada 8. Geological Survey ofCanada.

Economagic (2003) Producer Price Index Data.http://www.economagic.com/blsppi.htm.

Edwards A.C. (ed.) (2001) Mineral Resource and OreReserve Estimation: the AusIMM guide to goodpractice. Australasian IMM monograph series, 23,421–434. Australasian Institute of Mining andMetallurgy, Carlton, Victoria.

Eggert R.G. (1988) Base and precious metal explora-tion by major corporations. In Tilton J.E., EggertR.G. & Landsberg H.H. (eds), World MineralExploration, 105–144. Resources for the Future,Washington DC.

Page 470: Introduction to mineral exploration

REFERENCES 453

Eggington H.F. (ed.) (1985) Australian DrillersGuide. Australian Drilling Training Committee,Macquarie Centre, New South Wales.

Eimon P.I. (1988) Exploration for and evaluationof epithermal deposits. In Gold Update 87–88,Epithermal Gold, 68–80. Course Manual. Depart-ment of Geology, University of Southampton,Southampton.

Eldorado Gold (2003) Eldorado Gold Website. http://www.eldoradogold.com.

Ellis D. (1988) Well Logging for Earth Scientists.Elsevier, Amsterdam.

Ellis Consulting (2000) The Diavik DiamondsProject: the distribution of the project resource in-come. http://www.diavik.ca/pdf/ellisreport.pdf.

Emerson D.W. (1980) The Geophysics of the EluraOrebody. Australian Society of Exploration Geo-physicists, Sydney.

Engineering Group Working Party (1977) Quart. J.Eng. Geol., 10, Geol. Soc., London.

Englebrecht C.J. (1986) The West Wits Line. InAntrobus E.S.A. (ed.), Witwatersrand Gold – 100Years, 199–225. Geological Society of SouthAfrica, Johannesburg.

Englebrecht C.J., Baumbach G.W.S., Matthysen J.L.& Fletcher P. (1986) The West Wits Line. InAnnhaeusser C.R. & Maske S. (eds), MineralDeposits of Southern Africa, 599–648. GeologicalSociety of South Africa, Johannesburg.

Environment Agency (2002) Environmental ImpactAssessment (EIA): a handbook for scoping pro-jects. Environment Agency, London.

Erlich E.I. & Hausel W.D. (2002) Diamond Deposits:origin, exploration, and history of discovery.Society for Mining, Metallurgy, and Exploration,Littleton, CO.

Erseçen N. (1989) Known Ore and MineralResources of Turkey. M.T.A., Ankara.

ESRI (2004) ESRI Website. http://www.esri.com.European Space Agency (2004) ERS 1 and 2. Paris,

France.http://www.esa.int/export/esaSA

GGGWBR8RVDC_earth_0.html .Evans A.M. (1968) Charnwood Forest. In Sylvester-

Bradley P.C. & Ford T.D. (eds), The Geology of theEast Midlands, 1–12. Leicester University Press,Leicester.

Evans A.M. (1993) Ore Geology and IndustrialMinerals – an introduction. Blackwell ScientificPublications, Oxford.

Falcon (2003) BHP Billiton Falcon Website. http://falcon.bhpbilliton.com/.

Farmer I. (1983) Engineering Behaviour of Rocks,2nd edn. Chapman & Hall, London.

Fauth H., Hindel R., Siewers U. & Zinner J.(1985) Geochemischer Atlas BundesrepublikDeutschland. Bundesanstalt fur Geowissens-chaften und Rohstoffe, Hannover.

Feather C.E. & Koen G.M. (1975) The mineralogyof the Witwatersrand reefs. Miner. Sci. Engrg, 7,189–224.

Fenwick K.G., Newsome J.W. & Pitts A.E. (1991)Report of Activities 1990. Ontario Geological Sur-vey Misc. Paper 152. Toronto.

Fipke C.E., Gurney J.J. & Moore R.O. (1995)Diamond Exploration Techniques EmphasizingIndicator Mineral Chemistry and CanadianExamples. Geological Survey of Canada Bulletin423.

Fletcher W.K. (1981) Analytical Methods in Explora-tion Geochemistry. Handbook of ExplorationGeochemistry, vol. 1, Elsevier, Amsterdam.

Fletcher W.K. (1987) Analysis of soil samples. InFletcher W.K., Hoffman S.J., Mehrtens M.B.,Sinclair A.J. & Thomson I. (eds), ExplorationGeochemistry: design and interpretation ofsoil surveys, 73–96. Reviews in Economic Geo-logy, vol. 3. Society of Economic Geologists, ElPaso.

Fletcher W.K. (1997) Stream sediment geochem-istry in today’s exploration world. In GubinsA.G. (ed.), Geophysics and Geochemistry at theMillenium: Proceedings of Exploration 97, 249–260. Prospectors & Developers Association ofCanada, Toronto.

Fletcher W.K., Hoffman S.J., Mehrtens M.B., SinclairA.J. & Thomson I. (1987) Exploration Geochem-istry: design and interpretation of soil surveys. Re-views in Economic Geology vol. 3. Society of Eco-nomic Geologists, El Paso.

Fluor Mining and Metals Inc. (1978) The Use ofGeostatistics in Exploration and Development.Unpublished report, Fluor Metals Inc.

Forstner U. & Wittman G.T.W. (1979) Metal Pollu-tion in the Aquatic Environment. Springer-Verlag,Berlin.

Fortescue J.A.C. & Hornbrook D.H.W. (1969) Twoquick projects: one at a massive sulphide orebodynear Timmins, Ontario and the other at a copperdeposit in Gaspe Park, Quebec. In Progress Reporton Biogeochemical Research at the GeologicalSurvey of Canada, 39–63. Geological Survey ofCanada Paper 67-23. Ottawa.

FracSIS (2002) Fracsis Website. http://www.fractal.csiro.au/fracsis/.

François-Bongarçon D.M. & Gy P. (2002) The mostcommon error in applying ‘Gy’s Formula’ in thetheory of mineral sampling and the history of the

Page 471: Introduction to mineral exploration

454 REFERENCES

liberation factor. J.S. Afr. Inst. Min. Metall., 102,475–484.

Francké J.C. & Yelf R. (2003) Applications of GPRfor surface mining. In Proceedings of the 10thInternational Conference on Ground Pemetrat-ing Radar (GPR 2004). Delft University of Tech-nology, Delft, Netherlands

Franklin J.M. (1997) Lithogeochemical and minera-logical methods for base metal and gold ex-ploration In Gubins A.G. (ed.) Geophysics andGeochemistry at the Millenium: Proceedings OfExploration 97, 191–208. Prospectors & Develop-ers Association of Canada, Toronto.

Fraser Institute (2003) The Fraser Institute AnnualSurvey of Mining Companies. http://www.fraserinstitute.ca.

Friend T. (1990) The Vegetation Unit – caring for theenvironment. Mining Survey, 1990(1), 18–22.

Frimmel H.E. & Minter W.E.L. (2002) Recent de-velopments concerning the geological history ofthe Witwatersrand gold deposits, South Africa. InGoldfarb R. & Nielsen R.L. (eds) Global Explora-tion 2002– integrated methods for discovery.Special Publ. 9, 17–45. Society for EconomicGeology, Littleton.

Frolick V. (1999) Fire Into Ice. Raincoast Books,Vancouver.

Fry N. (1991) The Field Description Of MetamorphicRocks, 2nd edn. Wiley, New York.

Fyon A. (1991) Volcanic-associated massive basemetal sulphide mineralization. In Fyon J.A. &Green (eds), Geology and Ore Deposits of theTimmins District, Ontario (Field Trip 6). Geol.Surv. Canada Open File Rept 216.

Gableman J.W. & Conel J.E. (1985) Section 7,Uranium Commodity Report. In Abrams M.J.,Conel J.E. & Lang H.R. (eds), The Joint NASA/Geosat Test Case Report, 3 vols. American Asso-ciation of Petroleum Geologists, Tulsa.

Gardiner P.R. (1989) The recent Irish experiencein promoting mineral exploration. In ChadwickJ.R. (eds), Mineral Exploration Programmes ’89,paper 28. International Mining, Madrid.

Garrett R.G. (1989) The role of computers in explora-tion geochemistry. In Garland G.D. (ed.) Proceed-ings of Exploration ’87, spec. vol. 3, 586–608.Ontario Geological Survey, Toronto.

Gatzweiler R., Scheling B. & Tan B. (1981) Explora-tion of the Key Lake uranium deposits,Saskatchewan, Canada. In IAEA (eds), UraniumExploration Case Histories, 195–220. Interna-tional Atomic Energy Agency, Vienna.

Geelhoed B. & Glass H.J. (2001) A new modelfor sampling of particulate materials and deter-

mination of the minimal sample size. Geostand.Newsl. J. Geostand. Geoanalysis, 25, 65–81.

Gentry D.W. (1988) Minerals project evaluation – anoverview. Trans. Instn Min. Metall., Sect. A, 97,A25–A35.

Geocover (2004) Geocover (Earthsat Corp.) Website.http://www.geocover.com.

Geosoft (2004) Geosoft Website. http://www.geosoft.com.

Geovariances (2001) Isatis Software Manual, 3rdedn. Ecole des Mines de Paris, Fontainebleau.

Gibson H.L., Morton R.L. & Hudak G.J. (1999)Submarine volcanic processes, deposits and envir-onments favourable for the location of volcanic-as-sociated massive sulphide deposits In Barrie C.T.& Hannington M.T. (eds) Volcanic AssociatedMassive Sulfide Deposits and Examples in Mod-ern And Ancient Settings, 13–52. Reviews in Eco-nomic Geology, 8. Economic Geology PublishingCo., Littleton, CO.

Gilluly J., Waters A.C. & Woodford A.O. (1959)Principles of Geology. Freeman, San Francisco.

Glasson K.R. & Rattigan J.H. (eds) (1990) GeologicalAspects of the Discovery of Some Important Min-eral Deposits in Australia. Australasian. Instituteof Mining and Metallurgy, Melbourne.

GLCF (Global Land Cover Facility) (2004) GlobalLand Cover Facility. http//esip.umiacs.umd.edu.

Globe (Globe Team) (2004) Globe Website. http://www.ngdc.noaa.gov/seg/topo/globe.shtml.

Gocht W.R., Zantop H. & Eggert R.G. (1988) Inter-national Mineral Economics. Springer-Verlag,Berlin.

Goetz A.F.H., Rock B.N. & Rowan L.C. (1983)Remote sensing for exploration: an overview.Econ. Geol., 78, 573–590.

Goetz A.F.H. & Rowan L.C. (1981) Geologic remotesensing. Science, 211, 781–791.

Gökçen N. (1982) The ostracod biostratigraphy ofthe Denizli–Mugla Neogene Sequence. Bull. Inst.Earth Sci. 9, 111–131.

Golder Associates (1983) Feasibility Report on theSoma-Isiklar Lignite Deposit, Manisa Province,Western Turkey. Unpublished Report for TurkiyeKomur Isletmeleri Kurumu, 8 vols.

Goldfarb R.J. & Neilsen R.L. (eds) (2002) IntegratedMethods for Discovery: global exploration inthe twenty-first century. Society of EconomicGeologists, spec. publ. 9. Society of EconomicGeologists, Littleton, CO.

Goldspear (1987) Gold Panning. Practical and usefulinstructions. Goldspear (UK) Ltd, Beaconsfield.

Goldstein J.I., Newbury D.E., Echlin P., Joy D.C.,Fiori C.E. & Lifshin E. (1981) Scanning Electron

Page 472: Introduction to mineral exploration

REFERENCES 455

Microscopy and X-ray Microanalysis. Plenum,New York.

Goode J.R., Davie M.J., Smith L.D. & Lattanzi C.R.(1991) Back to basics: the feasibility study. Can.Inst. Min. Metall. Bull., 84, 53–61.

Goodman R.E. (1976) Methods in GeologicalEngineering in Discontinuous Rocks. West Pub-lishing Co., St Paul.

Goodman R.E. (1989) Introduction to RockMechanics, 2nd edn. Wiley, New York.

Goold D. & Willis A. (1998) The Bre-X Fraud.McClelland & Stewart, Toronto.

Gorman P.A. (1994) A review and evaluation ofthe costs of exploration, acquisition and devel-opment of copper related projects in Chile. InWhateley M.K.G. & Harvey P.K. (eds), MineralResource Evaluation II: Methods and Case His-tories, 123–128, spec. publ. 79. Geological Society,London.

Govett G.J.S. (1983) Rock Geochemistry in MineralExploration. Handbook of Exploration Geo-chemistry, vol. 3. Elsevier, Amsterdam.

Govett G.J.S. (1989) Bedrock geochemistry inmineral exploration. In Garland G.D. (ed.) Pro-ceedings of Exploration ’87, 273–299, spec. vol. 3.Ontario Geological Survey, Toronto.

Grant F.S. & West G.F. (1965) Interpretation The-ory in Applied Geophysics. McGraw-Hill, NewYork.

Graton L.C. (1930) Hydrothermal origin of the Randgold deposits, Part I, testimony of the conglomer-ates. Econ. Geol. 25 (suppl. to no. 3) 185pp.

Grayson R.L. (2001) Hazard identification, riskassessment and hazard control. In Karmis M. (ed.)Mine Health and Safety Management, 275–289.Society for Mining, Metallurgy and Exploration,Littleton, CO.

Gribble P. (1993) Fault Interpretation from CoalExploration Borehole Data using Surpac Software,abstract volume. Mineral Resource Evaluation1993 Conference, Leicester University.

Gribble P. (1994) Fault interpretation from coalexploration borehole data using Surpac software.In Whateley M.K.G. & Harvey P.K. (eds), MineralResource Evaluation, 29–35, spec. publ. 79. Geo-logical Society, London.

GSC (1995) Generalized Geological Map of theWorld, CD-ROM. Geological Survey of CanadaOpen File 2915d.

Gu Y. (2002) Automated scanning electron micro-scope based mineral liberation analysis. J. Min.Mater. Character. Engrg, 2, 33–41.

Gubins A.G. (ed.) (1997) Geophysics and Geochem-istry at the Millenium: Proceedings of Exploration

97. Prospectors & Developers Association ofCanada, Toronto.

Gulson B.L. (1986) Lead Isotopes in Mineral Explora-tion. Elsevier, Amsterdam.

Gunn A.G. (1989) Drainage and overburden geo-chemistry in exploration for platinum-groupelement mineralisation in the Unst Ophiolite,Shetland, U.K. J. Geochem. Explor., 31, 209–236.

Gurney J.J. (1984) A correlation between garnets anddiamonds in kimberlites. In Glover J.E. & HarrisP.G. (eds) Kimberlite Occurrence and Origin: abasis for conceptual models in exploration, 8,143–166. University of Western Australia, Perth,Western Australia.

Gurney J.J. & Zweistra P. (1995) The interpretationof the major element compositions of mantle min-erals in diamond exploration. J. Geochem. Explor.,53, 293–310.

Gy M. (1992) Sampling of Heterogeneous andDynamic Material Systems. Elsevier, New York.

Hale M. & Plant J.A. (1994) Drainage Geochem-istry. Handbook of Exploration Geochemistry.Elsevier, Amsterdam.

Hallbauer D.K. (1975) The plant origin of Wit-watersrand carbon. Miner. Sci. Eng., 7, 111–131.

Hallberg J.A. (1984) A geochemical aid to igneousrock identification in deeply weathered terrain. J.Geochem. Explor., 20, 1–8.

Handley G.A. & Carrey R. (1990) Big Bell gold de-posit. In Hughes F. (ed.) Geology of the MineralDeposits of Australia and Papua New Guinea,217–220. Australas. Inst. Min. Metall. Monograph14. Melbourne.

Handley J.R.F. (2004) Historic Overview of theWiwatersrand Goldfields. Handley, Howick,Natal, South Africa.

Hannington M.D. & Barrie C.T. (1999) The GiantKidd Creek Volcanogenic Massive SulfideDeposit, Western Abitibi Subprovince, Canada.Economic Geology Monograph 10. EconomicGeology Publishing Co., Littleton, CO.

Hanula M.R. (ed.) (1982) The Discovers. Pitt Publish-ing, Toronto.

Harben P.W. & Kuzvart M. (1997) IndustrialMinerals: a global geology, 3rd edn. InternationalMinerals Information, London.

Hart M. (2002) Diamond, the History of A Cold-Blooded Love Affair. Fourth Estate, London.

Haslett M.J. (1994) The South Deep project – geologyand planning for the future. In Anhaeusser CR(ed.) Proceedings XVth CMMI Congress, Vol. 3 –Geology, 71–83. S. Afr. I.M.M. Symposium Series14.

Page 473: Introduction to mineral exploration

456 REFERENCES

Hatton W. (1994a) INTMOV. A program for the in-teractive analysis of spatial data. In WhateleyM.K.G. & Harvey P.K. (eds), Mineral ResourceEvaluation II: methods and case histories, 37–43.Spec. Publ. 79. Geological Society, London.

Hatton W. (1994b) Exploration Risk Assessment inthe UK Coal Measures. Unpublished PhD, Univer-sity of Leicester.

Hawkes H.E. (1976) The downstream dilution ofstream sediment anomalies. J. Geochem. Explor.,6, 345–358.

Hawkins M.A. (1991) Large fine fraction stream sam-ples – a practical method to maximize catchmentsize, minimize nugget effect and allow detectionof coarse or fine gold. In 15th International Geo-echemical Symposium, Reno, abstract.

Heald P., Foley N.K. & Hayba D.O. (1987) Com-parative anatomy of volcanic-hosted epithermaldeposits: acid sulphate and adularia-sercite types.Econ. Geol., 82, 1–26.

Hedenquist J.W., Arribas R.A. & Gonzalez-Urien E.(2000) Exploration for epithermal gold deposits. InHagemann S.G. & Brown P.E. (eds) Gold in 2000,245–279. Reviews in Economic Geology, 13. Eco-nomic Geology Publishing Co., Littleton, CO.

Hefferman V. (1998) Worldwide Mineral Explora-tion. Financial Times Energy, London.

Heitt D.G., Dunbar W.W., Thompson T.B., JacksonR.G. (2003) Geology and geochemistry of the DeepStar gold deposit, Carlin Trend, Nevada. Econ.Geol., 98, 1107–1135.

Helmstaedt H. (1993) Primary diamond deposits –what controls their size grade and location? InWhiting B.H., Mason R. & Hodgson C.J. (eds)Giant Ore Deposits. Society of Economic Geo-logists Special Publication 2, 13–80. Society ofEconomic Geologists, Littleton, Colorado.

Henley R.W. (1991) Epithermal gold deposits involcanic terranes In Foster R.P. (ed.) GoldMetallogeny and Exploration, 133–164. Blackie,Glasgow.

Henley S. (1981) Nonparametric Geostatistics.Applied Science Publishers, London.

Heyl A.V. (1972) The 38th Parallel Lineament and itsrelationship to ore deposits. Econ. Geol., 67, 879–894.

Hill I.A. (1990) Shallow Reflection Seismic Methodsin the Extractive and Engineering Industries.Unpublished Workshop Notes, University ofLeicester.

Hitzman M.W., Proffett J.M. Jr, Schmidt J.M. &Smith T.E. (1986) Geology and mineralization ofthe Ambler District, Northwestern Alaska. Econ.Geol., 81, 1592–1618.

Hodgson C.J. (1989) Use (and abuses) of ore depositmodels in mineral exploration. In Garland G.D.(ed.), Proceedings of Exploration ’87, 31–45, spec.vol. 3. Ontario Geological Survey, Toronto.

Hoek E. & Bray J. (1977) Rock Slope Engineering.Institution of Mining and Metallurgy, London.

Hoek E. & Brown E.T. (1980) Underground Excava-tion in Rock. Institution of Mining and Metal-lurgy, London.

Hoeve J. (1984) Host rock alteration and its applica-tion as an ore guide at the Midwest Lake Uraniumdeposit, northern Saskatchewan. Can. Inst. Min.Metall. Bull., 77(August), 63–72.

Hoffman S.J. (1987) Soil sampling. In Fletcher W.K.,Hoffman S.J., Mehrtens M.B., Sinclair A.J. &Thomson I. (eds), Exploration Geochemistry:design and interpretation of soil surveys, 39–77.Reviews in Economic Geology, vol. 3. Society ofEconomic Geologists, El Paso.

Hoffmann P.F. (1989) Precambrian geology andtectonic history of North America. In Bally A.W.& Palmer A.R. (eds), The Geology of NorthAmerica: an overview, 447–512. Geological Soci-ety of America, Boulder, CO.

Hofstra A.H. & Cline J.S. (2000) Characteristics andmodels for Carlin-type deposits. In Hagemann S.G.& Brown P.E. Gold in 2000. Reviews in EconomicGeology, vol. 13, 163–220. Economic GeologyPublishing Co., Littleton, CO.

Hoover D.B., Heran W.D. & Hill P.L. (eds) (1992)The Geophysical Expression of Selected MineralDeposit Models. United States Department of theInterior, Geological Survey Open File Report92–557.

Horikoshi E. & Sato T. (1970) Volcanic activity andore deposition in the Kosaka Mine. In Tatsumi T.(ed.), Volcanism and Ore Genesis, 181–195. Uni-versity of Tokio Press, Tokio.

Hornbrook D.H.W. (1975) Kidd Creek Cu-Zn-Agdeposit Ontario (case history) J. Geochem. Explor.,4, 165–168.

Hosking K.F.G. (1971) Problems associated with theapplication of geochemical methods of explorationin Cornwall, England. In Boyle R.W. (ed.),Geochemical Exploration, spec. vol. 11, 176–189.Canadian Institution of Mining and Metallurgy,Toronto.

Howarth R.J. (ed.) (1982) Statistics and DataAnalysis in Exploration Geochemistry, vol. 2.Handbook of Exploration Geochemistry. Elsevier,Amsterdam.

Howarth R.J. (1984) Statistical applications in geo-chemical prospecting: a survey of recent develop-ments. J. Geochem. Explor., 21, 41–61.

Page 474: Introduction to mineral exploration

REFERENCES 457

Humphreys D. (2003) The Relocation of GlobalIndustrial Production: what it means for mining.http://www.riotinto.com.

Huston D.L. & Large R.R. (1989) A chemical modelfor the concentration of gold in volcanogenicmassive sulphide deposits. Ore Geol. Rev., 4, 171–200.

Hustrulid W.A. & Bullock R.L. (Eds) (2001) Under-ground Mining Methods: engineering funda-mentals and international case studies. Societyfor Mining, Metallurgy and Exploration, Littleton,CO.

Hutchinson R.W. (1980) Massive base metal sul-phide deposits as guides to tectonic evolution. InStrangeway D.W. (ed.), The Continental Crust andIts Mineral Deposits, 659–684. Geol. Assoc.Canada, Spec. Pap. 20.

Huchinson R.W. & Grauch R.I. (eds) (1991) His-torical Perspectives of Genetic Concepts and CaseHistories of Famous Discoveries. Monograph 8,Econ. Geol.

Hutchinson R.W. & Viljoen R.P. (1988) Re-evaluation of gold sources in Witwatersrand ores.S. Afr. J. Geol., 91, 157–173.

Hutchison C.S. (1974) Laboratory Handbook ofPetrographic Techniques. Wiley, New York.

Hutchison C.S. (1983) Economic Deposits and TheirTectonic Setting. Macmillan, London.

IG (1989) Code of Practice for Geological Visitsto Quarries, Mines and Caves. Institute ofGeologists, London.

IMM Working Group (2001) Code For Reporting OfMineral Exploration Results, Mineral Resourcesand Mineral Reserves (The Reporting Code) http://www.imm.org.uk/rescode/reportingcode.doc.

Independent Environmental Monitoring Agency(1998) Independent Environmental Monitor-ing Agency Report for 1998. http://www.monitoringagency.net/.

Ineson P.R. (1989) Introduction to Practical OreMicroscopy. Longman Scientific & Technical,Harlow.

Infomine (2004) Infomine Web Site, Vancouver.www.infomine.com.

Infoterra (2004) Ikonos. http://www.infoterra-global.com/ikonos.htm.

Institution of Mining and Metallurgy (1987) Financefor the Mining Industry. Trans. Instn Min. Metall.,Sect. A, 96, A7–A36.

Isaaks E.H. & Srivastava R.M. (1989) An Introduc-tion to Applied Geostastistics. Oxford UniversityPress, Oxford.

Ishikawa Y., Sawaguchi T., Iwaya S. & Horiuchi M.(1976) Delineation of prospecting targets for

Kuroko deposits based on modes of volcanism ofthe underlying dacite and alteration halos. Min.Geol., 26, 105–117 (in Japanese).

Isles D.J., Harman P.G. & Cunneen J.P. (1989) Thecontribution of high resolution aeromagneticsto Archaean gold exploration in the Kalgoorlieregion, Western Australia. In Keays R.R., RamsayW.R.H. & Groves D.I. (eds) The Geology of GoldDeposits – the perspective in 1988, 389–397. Econ.Geol. Monograph 6. El Paso.

Ixer R.A. & Duller P.R. (1998) Virtual Atlas ofOpaque and Ore Minerals in their Associations.http://www.smenet.org/opaque-ore/.

Jackson D.G. & Andrew R.L. (1990) Kintyre uraniumdeposit. In Hughes F. (ed.), Geology of theMineral Deposits of Australia and Papua NewGuinea, 653–658. Australas. Inst. Min. Metall.Monograph 14. Melbourne.

Jacobsen J.B.E. & McCarthy T.S. (1976) The copper-bearing breccia pipes of the Messina District,South Africa. Mineral. Deposita, 11, 33–45.

Jagger F. (1977) Bore core evaluation for coal minedesign. In Whitmore R.L. (ed.), Coal BoreholeEvaluation, 72–81. Symposium, Australasian In-stitution of Mining and Metallurgy, Parkville.

Janisch P.R. (1986) Gold in South Africa J.S. Afr. Inst.Min. Metall., 86, 273– 316.

Janse A.J.A. & Sheahan P.A. (1995) Catalog of worldwide diamond and kimberlite occurrences – aselective and annotative approach. J. Geochem.Explor., 53, 73–111.

Jaques A.L. & Milligan P.R. (2003) Patterns and con-trols on the distribution of diamond pipes inAustralia. In 8th International Kimberlite Confer-ence, Victoria, abstracts.

Jeffery R.G., Sampsom D.B., Seymour K.M. &Walker I.W. (1991) A review of current explorationand evaluation practices in the search for gold inthe Archaean of Western Australia. In World Gold’91, 313–322. Australasian Institution of Miningand Metallurgy, Cairns.

Jenner J.W. (1986) Drilling and the year 2000. Trans.Instn Min. Metall., Sect. B, 95, B77–144.

Johnson I.M. & Fujita M. (1985) The Hishikari golddeposit: an airborne EM discovery. CIM Bull., 78,61–66.

Johnson M.G. (1977) Geology and Mineral Depositsof Pershing County, Nevada, bull. 80. NevadaBureau of Mines and Geology.

Jolley S.J., Freeman S.R., Barnicoat A.C., et al. (2004)Structural controls on Witwatersrand gold miner-alization. J. Struct. Geol., 26, 1067–1086.

Jones M.P. (1987) Applied Mineralogy. Graham &Trotman, London.

Page 475: Introduction to mineral exploration

458 REFERENCES

Jordaan M. & Austin M. (1986) Gold-placer Sedi-mentology. Excursion guide 1A Geocongress ’86.Geological Society of South Africa, Johannesburg.

Journel A.G. & Huijbregts C-H.J. (1978) MiningGeostatistics. Academic Press, London.

Kalogeropolous S.I. & Scott S.D. (1983) Minera-logy and geochemistry of tuffaceous exhalites(Tetsuekiei) of the Fukazwa Mine, Hokuroku Dis-trict, Japan. In Ohmoto H. & Skinner B.J. (eds),Kuroko and Related Volcanogenic Massive Sul-fide Deposits. Econ. Geol. Monogr., 5, 412–432.

Kauranne K., Salimen R. & Eriksson K. (1992)Regolith Exploration Geochemistry in Arctic andTemperate Terrains. Handbook of ExplorationGeochemistry, vol. 5. Elsevier, Amsterdam.

Kearey P., Brooks M. & Hill I. (2002) An Intro-duction to Geophysical Exploration, 3rd edn.Blackwell Science Ltd, Oxford.

Kelly T., Buckingham D., DiFrancesco C., et al.(2001) Historical Statistics for Mineral and Mate-rial Commodities in the United States. USGeological Survey Open-File Report 01-006.//minerals.usgs.gov/minerals/pubs/of 01-006/.

Kernet C. (1991) Mining Equities: evaluation andtrading. Woodhead Publishing Ltd, Cambridge.

Kerr D.J. & Gibson H.L. (1993) A comparison of theHorne volcanogenic massive sulfide deposit andintracauldron deposits of the Mine Sequence,Noranda, Quebec. Econ. Geol., 88, 1419–1442.

Kesler S.E. (1994) Mineral Resources, Economics,and the Environment. Macmillan, Oxford.

Kilburn L.C. (1990) Valuation of mineral propertieswhich do not contain exploitable reserves. CIMBull., 83, 90–93.

King B.M. (2000) Optimal Mine Scheduling Policies.Unpublished PhD thesis, University of London.

Kirk J., Ruiz J., Chesley J., Walshe J. & England G.(2002) A major Archean, gold- and crust-formingevent in the Kaapvaal craton, South Africa. Sci-ence, 297, 1856–1858.

Kirkham R.V., Sinclair W.D., Thorpe R.I. & DukeJ.M. (eds) (1993) Mineral Deposit Modeling. Geo-logical Association of Canada Special Paper 40,St John’s, Newfoundland, Canada.

Kirkley M., Mogg T. & McBean D. (2003) Snap Lakefield trip guide. In Kjarsgaard B.A. (ed.) VIIIth inter-national Kimberlite Conference Field Trip Guide-book. Geological Survey of Canada, CD-ROM.

Klemd R. & Hallbauer D.K. (1987) Hydrothermallyaltered peraluminous Archaean granites as a prov-enance model for Witwatersrand sediments. Min-eral. Deposita, 22, 227–235.

Knox-Robinson C.M. (2000) Vectoral fuzzy logic: anovel technique for enhanced mineral prospect-

ivity mapping, with reference to orogenic goldmineralisation potential of the Kalgoorlie ter-rane, Western Australia. Australian J. Earthsci.,47, 929–941.

Knudsen H.P. (1988) A Short Course on Geostatist-ical Ore Reserve Estimation. Unpublished Rept,Montana Tech., Butte.

Knudsen H.P. & Kim Y.C. (1978) A comparativestudy of the geostatistical ore reserve estimationmethod over the conventional methods. Min. Eng.,30(1), 54–58.

Krajick K. (2001) Barren Lands: an epic search fordiamonds in the North American Arctic. W.H.Freeman & Co., New York.

Kreiter V.M. (1968) Geological Prospecting andExploration. Mir, Moscow.

Krige D.G. (1978) Log Normal – De Wijsiangeostatistics for ore evaluation. South AfricanInstitution of Mining and Metallurgy Monograph1. Johannesburg.

Kruse F.A. & Boardman J.W. (2000) Characterizationand mapping of kimberlites and related diatremesusing hyperspectral remote sensing. In Proceed-ings, 2000 IEEE AeroSpace Conference, 18–24March 2000, Big Sky, Montana.

Krynine D. & Judd W.R. (1957) Principles ofEngineering Geology and Geotechnics. MaGraw-Hill, New York.

Kunasz I.A. (1982) Foote Mineral Company – KingsMountain Operation. In Cerny P. (ed.), ShortCourse in Granitic Pegmatites in Science andIndustry, 505–512. Mineralogical Society ofCanada, Winnipeg.

Lafitte P. (ed.) (1970) Metallogenic Map of Europe, 9Sheets. UNESCO, Paris.

Lane K.F. (1988) The Economic Definition of Ore –Cut-off grades in theory and practice. MiningJournal Books, London.

Larocque A.C.L. & Cabri L.J. (1998) Ion-microprobequantification of precious metals in sulphide min-erals. In McKibben M.A., Shanks W.C. & RidleyW.I. (eds), Applications of Microanalytical Tech-niques to Understanding Mineralizing Processes.Reviews in Economic Geology 7, 155–168.

Larson L.T. (1989) Geology and gold mineralizationin west Turkey. Min. Eng., 41, 1099–1102.

Lawrence M.J. (1998) Australian project valuationlessons for Canadian developers. In PDAC (Pros-pectors & Developers Association of Canada)Mineral Property Valuation and Investor Concer,69–96. Short Course Notes, March 1998.

Leaman D.E. (1991) Exploration significance of grav-ity surveys, Roseberry Mine, Tasmania. Explora-tion Geophysics, 22, 231–234.

Page 476: Introduction to mineral exploration

REFERENCES 459

Leake R.C., Cameron D.G., Bland D.J., Styles M.T.& Rollin K.E. (1992) Exploration for Gold in theSouth Hams District of Devon. Mineral Recon-naissance Programme Report 121. British Geolo-gical Survey.

Leat P.T., Jackson S.E., Thorpe R.S. &Stillman C.J. (1986) Geochemistry of bimodalbasalt-subalkaline/peralkaline rhyolite provinceswithin the southern British Caledonides. J. Geol.Soc. London, 143, 259–273.

Lebrun S. (1987) The Development of a Menu DrivenComputer Program, ‘The Geostats Program’, toCarry Out Geostatistical Analysis of GeologicalData, 2 vols. Unpubl. MSc Dissertation, Univer-sity of Leicester.

Leca X. (1990) Discovery of a concealed massive sul-phide deposit at Neves-Corvo, southern Portugal –a case history. Trans. Instn Min. Metall. Sect B, 99,B139–B152.

Lednor M. (1986) The West Rand Goldfield. InAntrobus E.S.A. (ed.), Witwatersrand Gold – 100Years, 49–110. Geological Society of SouthAfrica, Johannesburg.

Levinson A.A. (1980) Introduction to ExplorationGeochemistry. Applied Publishing, Wilmette,IL.

Levinson A.A. (1998) Diamond sources and theirdiscovery. In Harlow G.E. (ed.) The Nature of Dia-monds, 72–104. Cambridge University Press.

Levinson A.A., Bradshaw P.M.D. & Thomson I. (eds)(1987) Practical Problems in Exploration Geo-chemistry. Applied Publishing, Wilmette, IL.

Lillesand T.M., Kieffer R.W. & Chipman J.W. (2004)Remote Sensing and Image Interpretation, 5thedn. Wiley, New York.

Lindley I.D. (1987) The discovery and exploration ofthe Wild Dog gold-silver deposit, east New Britain,Papua New Guinea. In Pacific Rim Congress ’87,283–286, Australasian Institution of Mining andMetallurgy, Melbourne.

Liu G., Diorio P., Stone P., et al. (2001) Detect-ing kimberlite pipes at Ekati with airbornegravity gradiometry. Abstract ASEG 15th Geo-physical Conference and Exhibition, August 2001,Brisbane.

LME (2003) Data & Prices, London Metal Exchange.http://www.lme.co.uk/data_prices/home.html.

Lo C.P. (1986) Applied Remote Sensing. LongmanScientific & Technical, New York.

Lock N.P. (1985) Kimberlite exploration in the Kala-hari region of Southern Botswana with emphasison the Jwaneng Kimberlite Province. In Prospect-ing in Areas of Desert Terrain, 183–190. Institu-tion of Mining and Metallurgy, London.

Longley P.A., Goodchild M.F., Maguire D.J. & RhindD.W. (1999) Geographical Information Systems, 2vols. Wiley, New York.

Longley P.A., Goodchild M.F., Maguire D.J. & RhindD.W. (2001) Geographic Information Systems andScience. Wiley, New York.

Lovell J.S. & Reid A.R. (1989) Carbon dioxide/oxygenin the exploration for sulphide mineralization. InGarland G.D. (ed.) Proceedings of Exploration ’87,457–472, spec. vol. 3. Ontario Geological Survey,Toronto.

Lovering T.G. & McCarthy J.H. Jr (eds) (1977) Con-ceptual models in exploration geochemistry –the Basin and Range Province of the westernUnited States and Northern Mexico. J. Geochem.Explor., 9, 89–365.

Lumina (2003) Analytica. http://lumina.com/software/.

Lydon J.W. (1989) Volcanogenic massive sulphide de-posits parts 1 & 2. In Roberts R.G. & Shehan P.A.(eds), Ore Deposit Models, 145–181. GeologicalAssociation of Canada, Memorial University,Newfoundland.

Lynn M.D., Wipplinger P.E. & Wilson M.G.C. (1998)Diamonds. In Anhaeusser C.G. & Wilson M.G.C.(eds) The Mineral Resources of South Africa.Handbook Council for Geoscience 16, 232–258.

Maas R., McColloch M.T., Campbell I.H. & CoadP.R. (1986) Sm-Nd and Rb-Sr dating of an Archaeanmassive sulfide deposit: Kidd Creek, Ontario. Ge-ology, 14, 585–588.

MacDonald A. (2002) North America: an industryin transition. Mines and Minerals Sustainable De-velopment Project. Earthscan, London, CD-ROM.

Machamer J.F., Tolbert G.E. & L’Esperance R.L.(1991) Discovery of Serra dos Carajas. InHuchinson R.W. & Grauch R.I. (eds), HistoricalPerspectives of Genetic Concepts and Case Histo-ries of Famous Discoveries, 275–285. Monograph8. Economic Geology Publishing Co., El Paso.

Mackenzie B. & Woodall R. (1988) Economic produc-tivity of base metal exploration in Australia andCanada. In Tilton J.E., Eggert R.G. & LandsbergH.H. (eds), World Mineral Exploration, 21–104.Resources for the Future, Washington.

Macnae J. (1995) Applications of geophysics for thedetection and exploration of kimberlites andlamproites. J. Geochem Expl., 53, 213–244.

Magri E.J. (1987) Economic optimization ofboreholes and deflections in deep gold exploration.J. S. Afr. Inst. Min. Metall., 87, 307–321.

Majoribanks R.W. (1997) Geological Methods inMineral Exploration and Mining. Chapman &Hall, London.

Page 477: Introduction to mineral exploration

460 REFERENCES

Maptek (2004) Maptek Web Site. http://www.maptek.com/pdf/VCS_southdeep.pdf.

Mason A.A.C. (1953) Thc Vulcan tin mine. InEdwards A.B. (ed.), Geology of Australian Ore De-posits, 718–721. Australasian Institution of Min-ing and Metallurgy, Melbourne.

Mather P.H. (1987) Computer Processing ofRemotely Sensed Images. John Wiley & Sons,Chichester.

Matulich A., Amos A.C., Walker R.R., et al. (1974)The Ecstall Story; the geology department.Can. Inst. Min. Metall. Bull., 77, 228–235.

Maynard J.B., Ritger S.D. & Sutton S.J. (1991)Chemistry of sands from the modern IndusRiver and the Archean Witwatersrand Basin: im-plications for the composition of the Archeanatmosphere. Geology, 19, 265–268.

Mazzucchelli R.H. (1989) Exploration geochemistryin areas of deeply weathered terrain: weatheredbedrock geochemistry. In Garland G.D. (ed.) Pro-ceedings of Exploration ’87, 300–311, spec. vol. 3.Ontario Geological Survey, Toronto.

McCabe P.J. (1984) Depositional environments ofcoal and coal-bearing strata. In Rahmani R.A. &Flores R.M. (eds), Sedimentology of Coal andCoal-bearing Strata, 13–42. Spec. Publ. 7. Interna-tional Association of Sedimentologists.

McCabe P.J. (1987) Facies studies of coal andcoal-bearing strata. In Scott A.C. (ed.), Coal andCoal-bearing Strata: recent advances, 51–66.Spec. Publ. 32. Geological Society, London.

McCabe P.J. (1991) Geology of coal: environments ofdeposition. In Gluskoter H.J., Rice D.D. & TaylorR.B. (eds), Economic Geology, US. The Geologyof North America, vol. P-2, 469–482. GeologicalSociety of America, Boulder, CO.

McClay K.R. (1991) The Mapping of GeologicalStructures. Wiley, New York.

McClenaghan M.B. & Kjarsgaard B.A. (2001) Indica-tor mineral and geochemical methods for diamondexploration in the glaciated terrain of Canada. InMcClenaghan M.B., Bobrowsky P.T., Hall G.E.M.& Cook S.J. (eds), Drift Exploration in Glacial Ter-rain, 83–124, spec. publ. 185. Geological Society ofLondon.

McLean A.C. & Gribble C.D. (1985) Geology forCivil Engineers, 2nd edn. George Allen & Unwin,London.

McMullan S.R., Matthews R.B. & RobertshawP. (1989) Exploration geophysics for Athabascauranium deposits. In Garland G. (ed.), Proceed-ings of Exploration ’87, 547–68. Ontario Geolo-gical Survey, Toronto.

McNish J. (1999) The Big Score. Doubleday, Canada.

McVey H. (1989) Industrial Minerals – canwe live without them? Industr. Min., 259,74–75.

Mellor E.T. (1917) The Geology of the Witwaters-rand: Map and Explanation. Spec. Pub. 3. Geolog-ical Survey of South Africa, Johannesburg.

Meyer F.M., Tainton S. & Saager R. (1990) Themineralogy and geochemistry of small-pebbleconglomerates from the Promise Formation in theWest Rand and Klerksdorp areas. S. Afr. J. Geol.,93, 103–117.

Micromine (2004) Micromine Website. http://www.micromine.com.au.

Milsom J. (2002) Field Geophysics, 3rd edn. JohnWiley & Sons, Chichester, 232 pp.

Minecost (2004) World Mine Cost Data Exchange.www.minecost.com.

Mining Journal (weekly) Mining Journal Website,London. www.mining-journal.com.

Minter W.E.L., Hill W.C.N., Kidger R.J., KingsleyC.S. & Snowden P.A. (1986) The Welkom Gold-field. In Annhaeusser C.R. & Maske S. (eds),Mineral Deposits of Southern Africa, 497–539.Geological Society of South Africa, Johannesburg.

Mitcham T.W. (1974) Origin of breccia pipes. Econ.Geol., 69, 412–413.

Mitchell R.H. (1986) Kimberlites. Plenum Press,London.

Mitchell R.H. (1991) Kimberlites and lamproites:primary sources of diamond. Geosci. Canada, 18,1–16.

Mitchell R.H. (1995) Kimberlites, Orangeites, andRelated Rocks. Plenum Press, London.

MMSD (2002) Breaking New Ground: minesand minerals sustainable development project.Earthscan, London. (http://www.iied.org/mmsd/finalreport/index.html).

Moeskops P.G. (1977) Yilgarn nickel gossan geo-chemistry – a review with new data. J. Geochem.Explor., 8, 247–258.

Moon C.J. (1973) A Carborne Radiometric Survey ofthe Prince Albert-Beaufort West area, Cape Prov-ince. Open File Report Geol. Surv. S. Africa, G227.Pretoria.

Moon C.J. (1999) Towards a more quantitativerelationship for the downstream dilution of pointsource geochemical anomalies. J. Geochem.Expl.,65, 111–132.

Moon C.J. & Whateley M.K.G. (1989) A retro-spective analysis of commercial explorationstrategies for uranium in the Karoo Basin, SouthAfrica. In Chadwick J.R. (ed.), Mineral Explora-tion Programmes ’89. Mining International,Madrid.

Page 478: Introduction to mineral exploration

REFERENCES 461

Moon W.M. (1990) Integration of geophysical andgeological data using evidential belief function.IEEE Trans. Geosci. Remote Sensing, 28, 711–720.

Moore J.M. & Moore A.E. (2004) The roles of primarykimberlitic and secondary Dwyka glacial sourcesin the development of alluvial and marine dia-mond deposits in Southern Africa. J. Afr. EarthSci., 38, 115–134.

Moore P.D. (1987) Ecological and hydrologicalaspects of peat formation. In Scott A. (ed.), Coaland Coal-bearing Stata, 17–24. Spec. Publ. 32.Geological Society, London.

Moore R.L. (1998) Case study – the impact of mineraldeposit models on exploration strategies of amajor mining company. In Prospectors & Develop-ers Association of Canada (ed.) The Funda-mentals of Exploration and Mining. Short CourseNotes, 95–104. Toronto.

Morgan J.D. (1989) Stockpiling. In Carr D.D. & HerzN. (eds), Concise Encyclopedia of MineralResources, 288–292. Pergamon Press, Oxford.

Morris R.O. (1986) Drilling – potential develop-ments. Trans. Instn Min. Metall., Sect. B, 95, B77–78.

Morrissey C.J. (1986) New trends in geological con-cepts. Trans. Instn Min. Metall., Sect. B, 95, B54–57.

Moseley F. (1981) Methods in Field Geology.Freeman, Oxford.

Muggeridge M.T. (1995) Pathfinder sampling tech-niques for locating primary sources of diamond:recovery of indicator minerals, diamonds andgeochemical signatures. J. Geochem Expl., 53,125–144.

Mular A.L. (1982) Mining and Mineral ProcessingEquipment Costs and Preliminary Cost Estima-tions, spec. vol. 25. Canadian Institution ofMining and Metallurgy, Montreal.

Mullins M.P. (ed.) (1986) Gold Placer Sedimento-logy. Excursion Guidebook 1A. Geological Societyof South Africa, Johannesburg.

Munha J., Barriga F.J.A.S. & Kerrich R. (1986) High18O/16O ore-forming fluids in volcanic-hosted basemetal massive sulphide deposits: geologic, 18O/16O, and D/H evidence from the Iberian Pyrite Belt;Crandon, Wisconsin; and Blue Hill, Maine. Econ.Geol., 81, 530–552.

NAMHO (1985) NAMHO Guidelines. NationalAssociation of Mining History Organizations,Matlock.

Nash C.R., Boshier P.R., Coupard M.M., TheronA.C. & Wilson T.G. (1980) Photogeology andsatellite image interpretation in mineral explora-tion. Miner. Sci. Eng. 12, 216–244.

Nebert K. (1978) Das Braunkohlenfuhrende Neo-gengebeit von Soma, West Anatolien. Bull. Min.Res. Explor. Inst. Turkey, 90, 20–72.

Niblak W. (1986) An Introduction to Digital ImageProcessing. Prentice Hall, New York.

Nichol I., Closs L.G. & Lavin O.P. (1989) Samplerepresentativity with reference to gold explora-tion. In Garland G.D. (ed.), Proceedings ofExploration ’87, 609–624, spec. vol. 3. OntarioGeological Survey, Toronto.

Nixon P.H. (1995) The morphology and nature ofprimary diamondiferous occurrences. J. Geochem.Explor., 53, 41–71.

Noetstaller R. (1988) Industrial Minerals: a tech-nical review. The World Bank, Washington.

Northcote A.E.A. (1998) Scoping of feasibilitystudies In: The Mining Cycle. Proceedings of theAusIMM 1998 annual conference held in MountIsa, Queensland, 19–23 April 1998. AustralasianIMM publication series, no. 2/98, 105–110.Australasian Institute of Mining and Metallurgy,Carlton, Victoria.

Northern Miner (weekly) Northern Miner Website.http://www.northernminer.com.

Northern Miner (1990) Buried treasures. NorthernMiner Magazine, March, 32–37.

Nowicki T.E., Carlson J.A., Crawford B.B., LockhartG.D., Oshurst P.A. & Dyck D.R. (2003) FieldGuide to the Ekati Mine. In Kjarsgaard B.A. (ed.)VIIIth International Kimberlite Conference FieldTrip Guidebook. Geological Survey of Canada,CD-ROM.

Nunes P.D. & Pyke D. (1981) Time–stratigraphiccorrelation of the Kidd Creek Orebody withvolcanic rocks south of Timmins, Ontario, asinferred from zircon U-Pb ages. Economic Geo-logy, 76, 944–951.

Obert L. & Duval W.I. (1967) Rock Mechanics andthe Design of Structures in Rock. Wiley, NewYork.

O’Hara T.A. (1980) Quick guide to the evaluation oforebodies. Can. Inst. Min. Metall. Bull., Feb., 87–99.

Onions R.I. & Tweedie J.R. (1992) Development ofa field computer data logger and its integra-tion with the DATAMINE mining software. InAnnels A.E. (ed.) Case Histories and Methods inMineral Resource Evaluation, 125–133. Geol. Soc.spec. publ., 63. Bath, UK.

Papenfus J.A. (1964) The Black Reef Series within theWitwatersrand Basin with special reference to theoccurrence at Government Gold Mining Areas. InHaughton S.H. (ed.), The Geology of Some Ore De-posits in Southern Africa, 191–218. Geological So-ciety of South Africa, Johannesburg.

Page 479: Introduction to mineral exploration

462 REFERENCES

Parasnis D.S. (1996) Principles of Applied Geophys-ics, 5th edn. Chapman & Hall, London.

Paterson N.R. & Halof P.G. (1991) Geophysicalexploration for gold. In Foster R.P. (ed.), GoldMetallogeny and Exploration, 360–398. Blackie,Glasgow.

PDAC (Prospectors & Developers Association ofCanada) (1998) Mineral Property Valuation andInvestor Concern. Short Course Notes, March1998.

Pearson G.N., Kelley S.P., Pokhilenko N.P. & BoydF.R. (1997) Laser 40Ar/39Ar dating of phlogopitesfrom southern African and Siberian kimberlitesand their xenoliths: constraints on eruption ages,melt degassing and mantle volatile compositions.Russian Geol. Geophys., 38, 106–118 (Englishedition).

Pemberton R.H. (1989) Geophysical response ofsome Canadian massive sulphide deposits. InGarland G.D. (ed.), Proceedings of Exploration ’87,517–531, spec. vol. 3. Ontario Geological Survey,Toronto.

Percival J.A. & Williams H.R. (1989) Late ArchaeanQuetico metasedimentary belt, Superior Province,Canada. Can. J. Earth Sci., 26, 677–693.

Perry G. (1989) Trinity Mine Operations Report.Unpublished Report for US Borax.

Peters E.R. (1983) The use of multispectral satel-lite imagery in the exploration for petroleum andminerals. Phil. Trans. R. Soc. Lond., A309, 243–255.

Peters W.C. (1987) Exploration and Mining Geology,2nd edn. Wiley, New York.

Phillips G.N. & Law J.D.M. (2000) WitwatersrandGold Fields: geology, genesis and exploration.In Hagemann S.G. & Brown P.E. Gold in 2000. Re-views in Economic Geology, 13, 439–500. Eco-nomic Geology Publishing Co., Littleton, CO.

Phillips G.N., Myers R.E. & Palmer J.A. (1987) Prob-lems with the placer model for Witwatersrandgold. Geology, 15, 1027–1030.

Pintz W. (1984) Ok Tedi – Evaluation of a ThirdWorld Mining Project. Mining Journal Books,Edenbridge, Kent.

Piper D.P. and Rogers P.J. (1980) Procedure for theAssessment of the Conglomerate Resources of theSherwood Sandstone Group. Mineral AssessmentReport 56. Institute of Geological Sciences,Keyworth.

Pirajno F. (1992) Hydrothermal Mineral Deposits:principles and fundamental concepts for theexploration geologist. Springer-Verlag, Berlin.

Pirow H. (1920) Distribution of the pebbles in theRand banket and other features of the rock. Trans.Geol. Soc. S. Afr., 23.

Pitard F.F. (1993) Pierre Gy’s Sampling Theory andPractice. CRC Press, Boca Raton.

Popoff C. (1966) Computing Reserves of MineralDeposits; principles and conventional methods.US Bureau of Mines Information Circular 8283.

Potts D. (1985) Guide to the Financing of MiningProjects. Trans. Instn Min. Met., Sect. A, 94,A127–A133.

Prain K.A.R. (1989) Application of MWD in theoil field drilling industry. Miner. Ind. Int., 987(March), 10–12.

Preston K.B. & Sanders R.H. (1993) Estimating thein-situ relative density of coal. Aust. Coal Geol., 9,22–26.

Pretorius C.C., Jamison A.A. & Irons C. (1989)Seismic exploration in the Witwatersrand Basin,Republic of South Africa. In Garland G.D. (ed.)Proceedings of Exploration ’87, 241–253, spec. vol.3. Ontario Geological Survey, Toronto.

Pretorius C.C., Trewick W.A. & Irons C. (1997)Application of seismics to mine planning at VaalReefs Gold Mine, Number 10 Shaft, Republic ofSouth Africa. In Gubins A.G. (ed.) Geophysics andGeochemistry at the Millenium: Proceedings ofExploration ‘97, 399–408. Prospectors & Develop-ers Association of Canada, Toronto.

Pretorius D.A. (1991) The sources of Witwaters-rand gold and uranium: a continued difference ofopinion. In Huchinson R.W. & Grauch R.I. (eds),Historical Perspectives of Genetic Concepts andCase Histories of Famous Discoveries, 139–163.Monograph 8. Econ. Geol.

Price M. (1985) Introducing Ground Water.Chapman & Hall, London.

Reed S.J.B. (1993) Electron Microprobe Analysis.Cambridge University Press, Cambridge.

Reedman J.H. (1979) Techniques in Mineral Explora-tion. Applied Science Publishers, London.

Reeves C.V. (1989) Geophysical mapping ofPrecambrian granite-greenstone terranes as an aidto exploration. In Garland G.D. (ed.), Proceedingsof Exploration ’87, 254–266, spec. vol. 3. OntarioGeological Survey, Toronto.

Reeves P.L. & Beck L.S. (1982) Famous miningcamps: a history of uranium exploration in theAthabasca Basin. In Hanula M.R. (ed.), The Dis-covers, 153–161. Pitt Publishing, Toronto.

Reflections (2004) Reflections Mining and Explora-tion Website. http://www.reflections.com.au/MiningandExploration/.

Regan M.D. (1971) Management of Exploration inthe Metals Mining Industry. MS thesis, Massachu-setts Institute of Technology, Cambridge.

Reim K. (1988) Trinity Ore Reserve, Task ForceReport. Unpublished Report for US Borax.

Page 480: Introduction to mineral exploration

REFERENCES 463

Reimann C. (1989) Reliability of geochemicalanalysis: recent experiences. Trans. Instn Min.Metall., Sect. B, 98, B123–B129.

Reimer T.O. & Mossman D.J. (1990) Sulfidizationof Witwatersrand black sands: from enigma tomyth. Geology, 18, 426–429.

Reinecke L. (1927) The location of payable ore-bodies in the gold-bearing reefs of the Witwaters-rand. Trans. Geol. Soc. S. Afr., 30, 89–119.

Richardson S.H., Erlank A.J., Harris J.W. & Hart S.R.(1990) Eclogitic diamonds of Proterozoic age fromCretaceous kimberlites. Nature, 346, 54–56.

Richardson S.H., Gurney J.J., Erlank A.J. & HarrisJ.W. (1984) Origin of diamonds in old enrichedmantle. Nature, 310, 198–202.

Rickard D. (1987) Proterozoic volcanogenic miner-alization styles. In Pharaoh T.C., Beckinsale R.D.& Rickard D. (eds), Geochemistry and Mineral-ization of Proterozoic Volcanic Suites, 23–35,spec. publ. 33. Geological Society, London.

Riddler G.P. (1989) The impact of corporate amdhost country strategy on mineral explorationprogrammes. In Chadwick J.R. (ed.), Paper 15,Mineral Exploration Programmes ’89. Interna-tional Mining, Madrid.

Ritchie W., Wood M., Wright R. & Tait D. (1977)Surveying and Mapping for Field Scientists.Longman, New York.

Robb L. (2004) Introduction to Ore-forming Pro-cesses. Blackwell Publishing, Oxford.

Robb L.J., Davis D.W., Kamo N.L., et al. (1990) U-Pbages on single detrital zircon grains from theWitwatersrand Basin, South Africa: constraintson the age of sedimentation and the evolution ofgranites adjacent to the basin. J. Geol., 98, 311–328.

Robb L.J. & Meyer F.M. (1990) The nature of theWitwatersrand hinterland: conjectures on thesource area problem. Econ. Geol., 85, 511–536.

Robb L.J. & Robb V.A. (1995) Gold in theWitwatersrand basin. In: Wilson M.G.C. &Anhaeusser C.R. (eds) The Mineral Resources ofSouth Africa. Council for Geoscience Handbook16, 294–349. Pretoria, South Africa.

Robinson D.N., Gumey J.J. & Shee S.R. (1984)Diamond eclogite, graphite eclogite xenolithsfrom Orapa, Botswana. In Kornprobst J. (ed.),Kimberlites II: the mantle, crust-mantle relation-ships, 11–24. Elsevier, Amsterdam.

Roberts R.G. & Sheahan P.A. (1988) Ore DepositModels. Geological Association of CanadaReprint Series 3, St Johns, Newfoundland.

Robinson S.C. (1952) Autoradiographs as a means ofstudying distribution of radioactive minerals inthin section. Am. Miner., 37, 544–547.

Rogers P.J., Chaterjee A.K. & Aucott J.W. (1990)Metallogenic domains and their reflection inregional lake sediment surveys from the MegumaZone, southern Nova Scotia. J. Geochem. Explor.,39, 153–174.

Rombouts L. (2003) Assessing the diamond poten-tial of kimberlites from discovery to evalua-tion bulk sampling. Mineral. Deposita, 38, 496–504.

Rona P.A. (1988) Hydrothermal mineralization atoceanic ridges. Can. Mineral., 26, 431–465.

Rose A.W., Hawkes H.E. & Webb J.S. (1979)Geochemistry in Mineral Exploration. AcademicPress, London.

Ross G.M., Parrish R.R., Villeneuve M.E. & BowringS.A. (1991) Geophysics and geochronology of thecrystalline basement of the Alberta Basin, westernCanada. Can. J. Earth Sci., 28, 512–522.

Rothschild L. (1978) Risk. Listener, 30 November,715.

Roux A.T. (1969) The application of geophysicsto gold exploration in South Africa. In Mining andGroundwater Geophysics, 425–438. Econ. Geol.Rept 26. Geological Survey, Canada.

Royle A.G. (1995) The Sampling and Evaluation ofGold Deposits. Unpublished PhD thesis, Univer-sity of Leeds.

Rush P.M. & Seegers H.J. (1998) Ok Tedi Copper-Gold Deposits. In Hughes F.E. (ed.), Geology of theMineral Depoits of Australia and Papua NewGuinea, Monograph 14, 1747–1754. AustralasianInstitute of Mining and Metallurgy, Melbourne,Victoria.

Saager R. (1981) Geochemical studies of the originof detrital pyrites in the conglomerates ofthe Witwatersrand Goldfields, South Africa. InArmstrong F.C. (ed.), Genesis of Uranium- andGold-Bearing Precambrian Quartz-pebble Con-glomerates, L1–L17, prof. paper 1161. US Geolo-gical Survey.

Sabins F.F. (1997) Remote Sensing. Principles andinterpretation, 5th edn. Freeman, New York.

Salminen R. and the FOREGS Geochemistry Group(2004) Geochemical Atlas of Europe. http://www.eurogeosurveys.org/foregs/.

Sangster D.F. (1980) Quantitative characteristicsof volcanogenic massive sulphide deposits: 1.metal content and size distribution of massive sul-phide deposits in volcanic centres. Can. Inst. Min.Metall., 73, 74–81.

Sangster D.F. & Scott S.D. (1976) Precambrian,strata-bound, massive Cu-Zn-Pb sulphide ores inNorth America. In Wolf K.H. (ed.), Handbook ofStrata-bound and Stratiform Deposits, vol. 6,129–222. Elsevier, Amsterdam.

Page 481: Introduction to mineral exploration

464 REFERENCES

Sato T. (1977) Kuroko deposits: their geology,geochemistry and origin. In Volcanic Processes inOre Genesis, spec. publ. 7. Geological Society,London.

Sawkins F.J. (1984, 1990) Metal Deposits in Relationto Plate Tectonics. Springer-Verlag, Berlin.

Schandl E.S. & Wicks F.J. (1993) Carbonate and asso-ciated alteration of ultramafic and rhyolitic rocksat the Hemingway property, Kidd Creek volcaniccomplex, Timmins, Ontario. Econ. Geol., 88,1615–1635.

Schowengerdt R.A. (1983) Techniques for ImageProcessing and Classification in Remote Sensing.Academic Press, New York.

Schunnesson H. & Holme K. (1997) Drill monitor-ing for geological mine planning in the ViscariaCopper Mine, Sweden. CIM Bull. Sept., 83–89.

Scott A.C. (ed.) (1987) Coal and Coal-bearing Strata:recent advances, spec. publ. 32. Geological Soci-ety, London.

Scott F. (1981) Midwest Lake uranium discovery,Saskatchewan, Canada. In IAEA (eds), UraniumExploration Case Histories, 221–239. Interna-tional Atomic Energy Agency, Vienna.

Scott-Russell H., Wanblad G.P. & De Villiers J.S.(1990) The impact of modified oilfield technologyon deep level mineral exploration and ultimatemining systems. In Ross-Watt R.A.J. & RobinsonP.D.K. (eds), Technical Challenges in Deep LevelMining, 429–439. South African Institution ofMining and Metallurgy, Johannesburg.

Select Committee (1982) Memorandum by theInst. Geol. Sci., Strategic Minerals. In StrategicMinerals Select Committee, HL Rep; 1981–82(217) xii.

Selley R.C. (1989) Deltaic reservoir prediction fromrotational dipmeter patterns. In Whateley M.K.G.& Pickering K.T. (eds), Deltas: sites and traps forfossil fuels, 89–95, spec. publ. 41. Geological Soci-ety, London.

Sengor A.M.C., Gorur N. & Saroglu F. (1985)Strike-slip faulting and related basin formationin zones of tectonic escape: Turkey as a case study.In Biddle K.T. & Christie-Blick N. (eds), Strike-slipDeformation Basin Formation and Sedimenta-tion, 227–264, spec. publ. 37. Soc. Econ. Paleo.Mineral, Houston.

Settle M., Abrams M.J., Conel J.E., Goetz A.F.H. &Lang H.R. (1984) Sensor assessment report. InAbrams M.J. (ed.), Joint NASA/Geosat Test CaseReport, 2–1 to 2–24. American Association ofPetroleum Geologists.

Severin P.W.A., Knuckey M.J. & Balint F. (1989)The Winston Lake, Ontario, massive sulphide

discovery – a successful result of an integratedexploration program. In Garland G.D. (ed.), Pro-ceedings of Exploration ’87, 60–69, spec. vol. 3.Ontario Geological Survey, Toronto.

Sichel H.S. (1966) The estimation of means andassociated confidence limits for small samplesfrom log normal populations. In Symposiumon Mathematical Statistics and ComputerApplications in Ore Valuation, 106–122. SouthAfrican Institution of Mining and Metallurgy,Johannesburg.

Siegal B.S. & Gillespie A.R. (1980) Remote Sensing inGeology. Wiley, New York.

Siegel S. (1956) Nonparametric Statistics for theBehavioural Sciences. McGraw-Hill, New York.

Sillitoe R.H. (1991) Intrusion-related gold deposits.In Foster R.P. (ed.) Gold Metallogeny and Explora-tion, 165–209. Blackie, Glasgow.

Sillitoe R.H. (1995) Exploration and Discoveryof Base- and Precious-metal Deposits in theCircum-Pacific Region During the Last 25 Years.Resource Geology spec. issue 19. Tokyo.

Sillitoe R.H. (2000) Exploration and Discoveryof Base- and Precious-Metal Deposits in theCircum-Pacific Region: late 1990s update.Resource Geology spec. issue 21. Tokyo.

Sinclair A.J. (1976) Probability Graphs, spec. publ.4. Association of Exploration Geochemists,Toronto.

Sinclair A.J. (1991) A fundamental approach tothreshold estimation in exploration geochem-istry: probability plots revisited. J. Geochem.Explor., 41, 1–22.

Slade M.E. (1989) Prices of metals: history. In CarrD.D. & Herz N. (eds), Concise Encyclopedia ofMineral Resources, 343–356. Pergamon Press,Oxford.

Smith A. (1988a) Cyanide degradation and detoxifi-cation in a heap leach. In Van Zyl D., Huchison I.& Kiel J. (eds) Introduction to Evaluation, Designand Operation of Precious Metal Heap LeachingProjects, 293–305. Society of Mining Engineers,Denver.

Smith B.H. (1977) Some aspects of the use ofgeochemistry in the search for nickel sulphidesin lateritic terrain in Western Australia. J.Geochem. Explor., 8, 259–281.

Smith C.B., Atkinson W.J. & Tyler E.W.J. (1990) Dia-mond exploration in Western Australia, NorthernTerritory and South Australia. In Glasson K.R. &Rattigan J.H. (eds), Geological Aspects of the Dis-covery of Some Important Mineral Deposits inAustralia, 429–454. Australasian Institution ofMining and Metallurgy, Melbourne.

Page 482: Introduction to mineral exploration

REFERENCES 465

Smith D.M. (1988b) Geology of silver deposits alongthe western cordilleras. In Silver Exploration, Min-ing and Treatment, 11–24. Institution of Miningand Metallurgy, London.

Smith J. (1991) How companies value properties.Can. Inst. Min. Metall. Bull., 84, 953, 50–52.

Smith M.R. & Collis L. (2001) Aggregates: sand,gravel and crushed rock aggregates for construc-tion purposes, 3rd edn, spec. publ. 17. GeologicalSociety of London Engineering Geology.

Smith R.J. & Pridmore D.F. (1989) Electromagneticexploration for sulphides in Australia. In GarlandG.D. (ed.), Proceedings of Exploration ’87, 504–516, spec. vol. 3. Ontario Geological Survey,Toronto.

Smith R.S. & Fountain D.K. (2003) Geophysics anddiamond exploration – a review. FUGRO Website.http://www.dighem.com.

Smith R.S., Annan A.P., Lemieux J. & Pedersen R.N.(1996) Application of a modified Geotem systemto reconnaissance exploration at Point Lake Area,N.W.T., Canada. Geophysics, 61, 82–92.

Snow G.G. & Mackenzie B.W. (1981) The environ-ment of exploration: economic, organizationaland social constraints. In Skinner B.J. (ed.),Economic Geology: Seventy-Fifth AnniversaryVolume, 871–896. Society of Economic Geo-logists, Littleton, Colorado.

Snyder J.P. (1987) Map Projections: a workingmanual. U.S.G.S. Professional Paper 1395. Wash-ington, DC.

Solomon M. (1976) ‘Volcanic’ massive sulphidedeposits and their host rocks – a review and anexplanation. In Wolf K.H. (ed.) Handbook ofStrata-Bound and Stratiform Deposits, vol. 6, 21–54. Elsevier, Amsterdam.

Solovov A.P. (1987) Geochemical Prospecting forMineral Deposits. Mir, Moscow.

Souch B.E., Podolsky T. & Geological Staff (1969)The sulphide ores of Sudbury: their particularrelationship to a distinctive inclusion-bearingfacies of the nickel irruptive. In Wilson H.D.B.(ed.), Magmatic Ore Deposits Symposium, 252–261, Econ. Geol. Monogr. 4. Society of EconomicGeologists, Littleton, Colorado.

South Deep (1990) Prospectus for Flotation of SouthDeep Exploration Company Ltd. South DeepExploration Co., Johannesburg.

South Deep (1992) Annual Report for 1991. SouthDeep Mining and Exploration Co., Johannesburg.

South Deep (2001) Annual Report for 2000. SouthDeep Mining and Exploration Co., Johannesburg.

South Deep (2004) South Deep Web Site. http://www.placerdome.com/southdeep.

Spectrum Mapping (2003) Remote Sensing: Hyper-spectral Imaging: AISA Sensor, Denver, Colorado.http://www.spectrummapping.com/rem-hyper-aisa.html.

Speight J.G. (1983) The Chemistry and Technologyof Coal. Marcel Dekker, New York.

Spock L.E. (1953) Guide to the Study of Rocks.Harper, New York.

Spooner E.T.C. & Barrie C.T. (1993) A special issuedevoted to Abitibi ore deposits in a modern con-text. Econ. Geol. 88, 1307–1322.

Sprigg G. (1987) Advances in image analysers.Microscopy and Analysis, Sept., 11–13.

Springett M. (1983a) Sampling and ore reserve esti-mation for the Ortiz Gold Deposit, New Mexico,USA. In AIME Precious Metal Symposium,Sparks, Nevada, USA.

Springett M. (1983b) Sampling Practices andProblems, preprint 83–395. Society of MiningEngineers of AIME, New York.

Stach E. (1982) Stach’s Textbook of Coal Petrology.Borntraeger, Berlin.

Stanistreet I.G. & McCarthy T.S. (1991) Changingtectono-sedimentary scenarios relevant to thedevelopment of the late Archean WitwatersrandBasin. J. Afr. Earth Sci., 13, 65–81.

Stanton R.L. (1978) Mineralization in island arcswith particular reference to the south-westPacific region. Proc. Australas. Inst. Min. Metall.,268, 9–19.

Stanton R.L. (1991) Understanding volcanic massivesulfides – past, present and future. In HuchinsonR.W. & Grauch R.I. (eds), Historical Perspect-ives of Genetic Concepts and Case Histories ofFamous Discoveries, 82–95, monograph 8. Econ.Geol., El Paso.

Steiger R. & Bowden A. (1982) Tungsten mineral-ization in southeast Leinster, Ireland. InBrown A.G. (ed.), Mineral Exploration in Ireland,Progress and Developments 1971–1981, 108–114. Irish Association of Economic Geology,Dublin.

Stephens J.D. (1972) Microprobe applications in min-eral exploration and development programmes.Miner. Sci. Eng., 3, 26–37.

Stephenson P.R. (2003) The JORC Code –maintaining the standard. http://www.jorc.org/pdf/stephenson3.pdf

Stewart B.D. (1981) Exploration of the uranium reefsof Cooke Section, Randfontein Estates GoldMining Company (Witwatersrand) South Africa.In International Atomic Energy Agency (eds), Ura-nium Exploration Case Histories, 141–170. Inter-national Atomic Energy Agency, Vienna.

Page 483: Introduction to mineral exploration

466 REFERENCES

Stewart J.H. (1980) Geology of Nevada, spec. publ. 4.Nevada Bureau of Mines and Geology, Reno,Nevada.

St Pierre M. (1999) Geophysical characteristicsof BHP/Dia Met Kimberlites, NWT, Canada. InLowe C., Thomas M.D. & Morris W.A. (eds) Geo-physics in Mineral Exploration: fundamentals andcase histories. Geological Association of CanadaShort Course Notes 14. St John’s, Newfoundland.

Strauss H. (1989) Carbon and sulfur isotope datafor carbonaceous metasediments from the KiddCreek massive sulfide deposit and vicinity,Timmins, Ontario. Econ. Geol., 84, 959–962.

Sutherland D.G., & Dale M.L. (1984) Methods ofestablishing the minimum size for sampling allu-vial diamond deposits. Trans. Instn Min. Metall.,Sect. B, 93, B55–58.

Swan A.R.H. & Sandilands M. (1995) IntroductionTo Geological Data Analysis. Blackwell Science,Oxford.

Taufen P.M. (1997) Ground waters and surfacewaters in exploration geochemical surveys. InGubins A.G. (ed.) Geophysics and Geochemistryat the Millenium: Proceedings Of Exploration 97,271–284. Prospectors & Developers Association ofCanada, Toronto.

Tankard A.J., Jackson M.P.A., Eriksson K.A., HobdayD.K., Hunter D.R. & Minter W.E.L. (1982) CrustalEvolution of Southern Africa. Springer-Verlag,Berlin.

Taylor B.E. & Huston D.L. (1998) Hydrothermalalteration of oxygen isotope ratios in quartzphenocrysts, Kidd Creek mine, Ontario: magmaticvalues preserved in zircons. Geology, 26, 763–764.

Taylor C.D., Zierenberg R.A., Goldfarb R.J., KilburnJ.E., Seal R.R. II & Kleinkopf M.D. (1995)Volcanic-associated massive sulfide depositsIn du Bray E. (ed.) Preliminary Compilation ofDescriptive Geoenvironmental Mineral DepositModels, 137–144. USGS Open File 95-831 http://pubs.usgs.gov/of/1995/ofr-95-0831

Taylor H.K. (1989) Ore reserves – a general overview.Miner. Ind. Int., 990, 5–12.

Telford W.M., Geldart L.P., Sheriff R.E. & Keys D.A.(1990) Applied Geophysics, 2nd edn. CambridgeUniversity Press, Cambridge.

Thatcher J., Struhsacker D.W. & Kiel J. (1988) Regu-latory aspects and permitting requirements forprecious metal heap leach operations. In Van ZylD., Hutchison I. & Kiel J. (eds), Introduction toEvaluation, Design and Operation of PreciousMetal Heap Leaching Projects, 40–58. Society ofMining Engineers. Society of Mining Engineers,Littleton, Colorado.

Theron J.C. (1973) Sedimentological evidence for theextension of the African continent during thelate Permian–early Triassic times. In CampbellK.S.W. (ed.), Gondwana Geology, 61–71. A.N.U.Press, Canberra.

Thiann R. Yu (1983) Rock mechanics to keep a mineproductive. Can. Min. J., April, 61–66.

Thomas E. (2003) Exploration for Diamonds inCanada. Powerpoint presentation. Resource Expo2002, Calgary.

Thomas L. (1992) Handbook of Practical Coal Geo-logy. Wiley, Chichester.

Thompson A.J.B. & Thompson J.F.H. (eds) (1996)Atlas of Alteration: a field and petrographic guideto hydrothermal alteration. Geological Asso-ciation of Canada, Mineral Deposits Division,St John’s, Newfoundland.

Thompson M. (1982) Control procedures in explora-tion geochemistry. In Howarth R.J. (ed.), Statisticsand Data Analysis in Exploration Geochemistry.Handbook of Exploration Geochemistry, vol. 2,39–58. Elsevier, Amsterdam.

Thompson M. & Walsh N. (1989) Handbookof Inductively Coupled Plasma Spectrometry.Blackie, London.

Thomson I. (1987) Getting it right. In Fletcher W.K.,Hoffman S.J., Mehrtens M.B., Sinclair A.J. &Thomson I. (eds), Exploration Geochemistry:design and interpretation of soil surveys. Reviewsin Economic Geology, vol. 3, 1–17. Society of Eco-nomic Geologists, Littleton, Colorado.

Thorpe R. & Brown G.C. (1993) The Field Descrip-tion of Igneous Rocks. Wiley, New York.

Tilton J.E., Eggert R.G. & Landsberg H.H. (eds) (1988)World Mineral Exploration. Resources for theFuture, Washington.

Tona F., Alonso D. & Svab M. (1985) Geology andmineralization in the Carswell Structure – a gen-eral approach. In Laine R., Alonso D. & Svab M.(eds), The Carswell Structure Uranium Deposits,1–18, spec. paper 29. Geological Association ofCanada, St Johns, Newfoundland.

Travis G.A., Keays R.R. & Davison R.M. (1976)Palladium and iridium in the evaluation of nickelgossans in Western Australia. Econ. Geol., 71,1229–1242.

Tregoning G.W. & Barton V.A. (1990) Design of awide orebody mining system for a deep-level mine.In Ross-Watt R.A.J. & Robinson P.D.K. (eds), Tech-nical Challenges in Deep Level Mining, 601–623.South African Institution of Mining and Metal-lurgy, Johannesburg.

Tucker M.E. (1988) Techniques in Sedimentology.Blackwell Scientific Publications, Oxford.

Page 484: Introduction to mineral exploration

REFERENCES 467

Tucker M.E. (2003) Sedimentary Rocks in the Field,3rd edn. Wiley, New York.

Tucker R.F. & Viljoen R.P. (1986) The geology ofthe West Rand Goldfield. In Annhaeusser C.R.& Maske S. (eds), Mineral Deposits of SouthernAfrica, 649–698. Geological Society of SouthAfrica, Johannesburg.

Turnbull N. (1999) Internal Control: guidance fordirectors on the combined code. ConsultativeCommittee of Accounting Bodies, London.

Turner B.R. (1985) Uranium mineralization in theKaroo Basin, South Africa. Econ. Geol., 80, 256–269.

USBM (1976) Coal Resource Classification Systemof the US Bureau of Mines and US Geological Sur-vey. US Geol. Surv. Bull. 1450-B.

USGS (1976) Principles of a Resource/ReserveClassification for Minerals. US Geol Surv Bull.1450-A.

USGS (2003) Gemstones 2003. minerals.usgs.gov/minerals/pubs/commodity/gemstones/.

USGS (2004) US Geological Survey Mineral DepositModels Web Page. minerals.cr.usgs.gov/team/depmod.html.

Ussher W.A.E. (1912) The Geology of the CountryAround Ivybridge and Modbury. Memoirs of theGeological Survey of England and Wales, Sheet349. HMSO, London.

Van Blaricom R. (1993) Practical Geophysics II.Northwest Mining Association, Spokane.

Vearncombe S. & Vearncombe J.R. (2002) Tectoniccontrols on kimberlite location, southern Africa. J.Struct. Geol., 24, 1619–1625.

Vikre P.G. (1989) Fluid mineral relations in theComstock Lode. Econ. Geol. 84, 1574–1613.

Viljoen R.P. (1990) Deep level mining – a geologicalperspective. In Ross-Watt R.A.J. & RobinsonP.D.K. (eds), Technical Challenges in Deep LevelMining, 411–427. South African Institution ofMining and Metallurgy, Johannesburg.

Viljoen R.P., Viljoen M.J., Grootenboer J. &Longshaw T.G. (1975) ERTS-1 imagery – anappraisal of applications on geology and mineralexploration. Min. Sci. Eng., 7, 132–168.

Visidata (2004) Visidata Website. http://www.visidata.com.au/.

Walker J. & Howard S. (2002) Finding the WayForward – how could voluntary action movemining towards sustainable development? WorldBusiness Council for Sustainable Development,Conches, Switzerland. (http://www.wbcsd.ch/).

Walker R.R., Matulich A., Amos A.C., Watkins J.J.& Mannard G.W. (1975) The geology of the KiddCreek Mine. Econ. Geol., 70, 80–89.

Walsham B.T. (1967) Exploration by diamonddrilling for tin in west Cornwall. Trans. Instn Min.Metall., Sect. A, 76, A49–56.

Walters W.H. (1999) Building and maintaining anumeric data collection. J. Documentation, 55,271–287.

Wang C. (1995) Application of Geographical Information Systems to the Interpretation ofExploration Geochemical Data and Modellingof Gold Prospects, South Devon, England. PhDThesis, University of Leicester.

Wanless R.M. (1982) Finance for Mine Management.Chapman & Hall, London.

Ward C.R. (ed.) (1984) Coal Geology and Coal Tech-nology. Blackwell Scientific Publications, Oxford.

Ward M-C. & Lawrence R.D. (1998) Comparabletransaction analysis: the market place is alwaysright. In PDAC (Prospectors & Developers Asso-ciation of Canada) Mineral Property Valuationand Investor Concern. Short Course Notes, March1998.

Weaver T.A., Freeman S.H., Broxton D.E. & BolivarS.L. (1983) The Geochemical Atlas of Alaska. LosAlamos National Laboratory, Los Alamos.

Webb J.S., Howarth R.J., Thompson M., et al. (1978)Wolfson Geochemical Atlas of England andWales. Clarendon Press, Oxford.

Wellmer F.-W. (1998) Problems related to cut-offlevels. In Statistical Evaluations in Explorationfor Mineral Deposits, 151–159. Springer-Verlag,Berlin.

Wellmer F.W. & Becker-Platen J.D. (2002) Sustain-able development and the exploitation of mineraland energy resources: a review. Int. J. Earth Sci.,91, 723–745.

Wells J. (1999) Bre-X: the inside story of the world’sbiggest mining scam. Texere Publishing, NewYork.

Werdmuller V.W. (1986) The Central Rand. InAntrobus E.S.A. (ed.), Witwatersrand Gold – 100Years, 7–48. Geological Society of South Africa, Jo-hannesburg.

West G. (1991) The Field Description of EngineeringSoils and Rocks. Open University Press, MiltonKeynes.

Western Areas (2003) Western Areas Annual Report2002. http://www.westernareas.co.za.

Western Mine Engineeering (2004) Website.www.westernmine.com.

Western Mining (1993) News Release to Share-holders, Western Mining Corp., Adelaide.

Whateley M.K.G. (1991) Geostatistical determina-tion of contour accuracy in evaluating coal seamparameters: an example from the Leicestershire

Page 485: Introduction to mineral exploration

468 REFERENCES

Coalfield, England. Bull. Soc. Géol. France, 162,209–218.

Whateley M.K.G. (1992) The evaluation of coalborehole data for reserve estimation and minedesign. In Annels A.E. (ed.), Mineral ResourceEvaluation, 95–106, spec. publ. 63. GeologicalSociety, London.

Whateley M.K.G. (2002) Measuring, understandingand visualising coal characteristics – innovationsin coal geology for the 21st century. Int. J. CoalGeol., 50, 303–315.

Whateley M.K.G. & Harvey P.K. (eds) (1994) MineralResource Evaluation II: methods and case his-tories, spec. publ. 79. Geological Society, London.

Whateley M.K.G. & Spears D.A. (eds) (1995) Euro-pean Coal Geology, spec. publ. 82. GeologicalSociety, London.

Wheat T.A. (1987) Advanced ceramics in Canada.Can. Inst. Min. Metall. Bull., 80(April), 43–48.

Whitchurch K.D., Gillies Saunders A.D. & Just G.D.(1987) A geostatistical approach to coal reserveclassification. Pacific Rim Congress, 87, 475–482.

White A.H. (1997) Management of Mineral Explora-tion. Andrew White and Associates, Moggill,Queensland & Australian Mineral Foundation,Glenside, South Australia.

White C.J. (1989) A Summary of the Mineral SandsIndustry of Western Australia Including an Explo-ration and Feasibility Project. Unpublished BScDissertation, Leicester University.

White M.E. (2001) Feasibility studies: scope and ac-curacy. In: Edwards A.C. (ed.), Mineral Resourceand Ore Reserve Estimation: the AusIMM guide

to good practice. Australasian IMM monograph se-ries, 23, 421–434. Australasian Institute of Miningand Metallurgy, Carlton, Victoria.

Whitely R.J. (ed.) (1981) Geophysical Case Study ofthe Woodlawn Orebody, Australia. PergamonPress, Oxford.

Williams H.R. & Williams R.A. (1977) Kimberlitesand plate tectonics in West Africa. Nature, 270,507–508.

Willis R.P.H. (1992) The integration of new tech-nology in South African gold mines as a survivalstrategy. In Jones M.J. (ed.), Mines, Materials andIndustry, 509–524. Institution of Mining and Met-allurgy, London.

Wills B.A. (1985, 1988) Mineral Processing Tech-nology. Pergamon Press, Oxford.

Wills B.A. (1997) Mineral Processing Technology.3rd edn. Pergamon Press, Oxford.

Woodall R. (1984) Success in mineral exploration.Geoscience Canada, 11, 41–46, 83–90, 127–133.

Woodall R. (1992) Challenge of minerals explorationin the 1990s. Min. Eng., July, 679–684.

Zarraq G. (1987) Comparison of Quality and ReserveEstimation Methods on Soma Lignite Deposit,Turkey. Unpublished MSc Dissertation, Univer-sity of Leicester.

Zeegers H. & Leduc C. (1991) Geochemical expora-tion for gold in temperate, arid, semi-arid and rainforest terrains. In Foster R.P. (ed.), Gold Metal-logeny and Exploration. Blackie, Glasgow, 309–335.

Zussman J. (1977) Physical Methods in Determinat-ive Mineralogy. Academic Press, London.

Page 486: Introduction to mineral exploration

INDEX 469

Index

alteration 392, 394argillic 383, 394hydrothermal 72, 365, 378quartz-adularia-sericite (QAS)

383, 396regional, propylitic 394see also silicificationsee also wall rock alteration

alteration halo 365, 366aluminum 19, 20

price 9, 9production 6

alunite 383analytical errors 215–16

see also accuracy and precisionanalytical techniques 78Anglo American plc 12, 13

AngloGold Ashanti 12, 13Anglo American Platinum 12,

13anisotrophy, representation of

concept of 239, 239anomalies

statistically defined 162–4antimony 20aquifers 249Arc-SDM 192arenaceous hosts 37argillaceous hosts 37argillites 37arsenic 20, 25Argyle, Australia 430–1, 431asbestos 20assay cut-off 232assay portion 212Aster imagery 105, 105, 114, 116asymmetrical distribution 201,

201

Athabasca Basin, unconformityrelated uranium deposits 40,41–5, 42–3, 45

atomic absorption spectrometry(AAS) 78, 160–2, 161, 161,390

AAS assays 390graphite furnace 385

attribute data 179audio-frequency magnetotellurics

146–7, 390Auger drilling 218–19, 283–4, 284Auld Lang Syne Group (ALSG)

391–3AUTOCAD 72autoradiography 27Average Annual Production Rate

(AAPR) 254

Banded Iron Formation 38Barrick Gold 12, 13Basal Reef 322, 324, 329, 339basal till sampling 173, 174base case 269B

cash flow prediction 269–70base metals

exploration, Australia cf.Canada 17–18

and silver mineralisation 395basement, Mesozoic 298Bathurst area, Canada 197bauxite 20, 39Beaufort Group 64bedrock diamond sources 414–15Besshi-type massive sulfides 45,

358, 360between-batch effects 160BHP Billiton 12, 13, 433

Page numbers in italics refer tofigures; page numbers in bold referto tables. The letter B after a pagenumber refers to boxes.

Aber Resources 440Abitibi Greenstone Belt 370–1,

370problems of surficial

geochemical exploration 374accuracy and precision 159–60acquisition value (of

mineralisation) 268aerial photographs 58, 64, 104,

117–26coverage 118, 119, 120distortion 121, 122flying, problems with 122interpretation 122–5low sun angle 125oblique 117photographic resolution 121scale of 118, 120, 120use of 122–5vertical 117

aeromagnetic anomalies 58Africa, southern, kimberlites 418,

420Aggregate reserves, development

of 281–93airborne surveys 58, 127–9, 128,

144, 151airborne gravity surveys 133Alcoa 12Alligator River, Northern

Territory, unconformityrelated uranium deposits 40,41–5, 42–3

Page 487: Introduction to mineral exploration

470 INDEX

Billa Barra Hill 283biogeochemistry 176–7bismuth 20bit problems, drilling at Stud Farm

285Black Reef mineralisation 323black smokers 364blue ground 415Blusson, Stuart 431, 432Boolean algebra 193Boolean methods 192, 193borehole sampling 310borehole surveying 229Bothaville Gap 322Botswana

Jwaneng mine 428, 429kimberlite exploration 427,

429Orapa mine 428, 429

Bouguer anomalies 133–4massive pyrite orebodies

134boulder tracing 82, 82, 173

see also glacial boulder trainsboxworks 80, 80brannerite 329breccia 394

argillite 393hydrothermal 382stockwork 395

brecciation 364crackle 376

Bre-X Minerals 100–2, 101brittle failure 395budgeting 70–1

for exploration planning 55–7,57

Bulk Leach Extractable Gold(BLEG) 385

Busang 100–2Bushveld Complex, South Africa

39, 112

calc-silicate minerals 36caliper tool 297capital and operating costs 264–5Carajas iron deposits, Brazil 78carat 414

definition 414Carbon Leader Reef 322, 337–8

facies of 326, 327Carlin Trend, Nevada 382Carlin type deposits 382–3

formation of 383

cash flow 261B, 261–72base case cash flow prediction

268B, 268–70in DCF analysis 268–70differential inflation in 271–2,

272cassiterite 25cathodoluminescence 27cauldron subsidence areas, VMS

deposits in 363Central Rand Group 322–3, 324,

337–8depositional setting 322stratigraphy 349

chalcopyrite 25channel sampling 207Charnwood Forest 281chemical sediments 36–7chemical tests, in the field 81chemical treatment, after cyanide

leaching 386, 397Chile, copper 9Chinese demand for minerals,

impact on exploration 5chip sampling 207, 385chloritization 43, 362, 378chlorosis 176chromite 20, 39, 421, 422, 423chromite ores, use of microscopy

31clay minerals 27–8Cliffe Hill Quarry (Leics.),

development of aggregatereserves 281–93

desk study 282–3environmental impact

292–3follow-up work, initial and

detailed 283–91project implementation

291–2Clifford’s Rule 419cluster development, and

favorable horizons (VMSdeposits) 363–4, 364, 372

coalproperties and quality of 294,

309–10see also Soma Lignite Basin

Codelco 12color anomaly, of outcropping

mineral deposits 79, 80Comex (New York Commodity

Exchange) 6

commodity prices 6–11forces determining prices 7–9trend 258, 258world markets 6–8

comparative production costranking 266B

compressive strength, unconfined246–7

Computer Compatible Tapes(CCTs) 107, 110

computer modeling, to integratedata 150

see also GIScomputer software

data handling 179–80mining 179

Comstock Lode 383cone and quartering 211–12confidence interval, two-sided

202, 202, 202conglomerate

oligomictic 328uraniferous 37–8see also Witwatersrand gold

constant money 268, 271–2contamination, affecting stream

sampling 165contractors, negotiations with

152contracts, long term supply 6Controlled Source Audio Magneto

Telluric (CSAMT) survey145, 385, 386

Cooke Section, West Rand 323,346–52

development of 352drilling 350geophysics 350interpretation of results 350–2

coordinate systems, map 186copper 21, 35, 37, 170, 365

discovery rate 15, 15price 9, 9production 6stratiform, China 37supply and demand 4

copper shales 37, 41copper-lead-zinc deposits 36–7copper-lead-zinc-silver deposits

358–64see also Trinity Silver Mine

copper-zinc deposits 38–9, 361–2core logging 225–6, 226core recovery, effective 225, 226

Page 488: Introduction to mineral exploration

INDEX 471

core, whole assays 101Coronation Hill Au-Pd deposit,

Northem Territory 45corporate exploration expenditure

55–6, 56costs

adjustment of 270–1development, greenfield site to

productive quarry 291–4,292, 293, 294

on project basis 56–7of working Witwatersrand gold

330–1Coulomb strength envelope 246Coulomb’s shear strength

criterion 246–7crackle brecciation 376cross-validation 239cross-seam mining 312, 313Crown lavas 322crustal extension 391current money 272cut-off grade 1cyanidation (cyanide leaching)

68, 162, 335–6, 343, 386, 397cyanide degradation, natural and

chemical 397, 399, 398cyanide leachable assays 162, 396Cyprus-type sulfide deposits 358,

360Czech Republic, ban on

cyanidation 68

Daisy Creek, Montana, statisticalexamination of copper data163–4, 163, 164

Darcy’s law 248–9data

attribute 179spatial component 179

data assessment, Soma LigniteBasin 303–7

data capture 179, 185digitization 185field 185

data model 179–80data sources 186–7data storage 181–2

flat file 181–2, 182relational database 182

data typeslines 180, 180points 180, 180polygons 180, 180

databasenormalization 182–4object oriented 182

debt finance 274–5, 275deep overburden geochemistry 82deformation, polyphase 391deleterious (undesirable)

substances 25–6, 31Demir Basin 302, 302Dempster Shafer method 194density logs 297depositional environment,

Witwatersrand Supergroup322, 326, 328

depositional model, Soma Basin302–3

desk studies 57–61Cliffe Hill Quarry 282–3example, target areas for

epithermal gold, westernTurkey 59–62, 60, 62

dewateringof the open pit, Soma lignite

309dewatering cones 249Dia Met Minerals 433, 435diamond

clarity 414color 414cut 414deposits 414–15jewellery 413market 413–14size 414world production 414, 415

diamond core drilling 220–5diamond drill bits 223

impregnated or surface-set 223diamond drilling 86, 220–5,

389–90diamond exploration 413–43diamondiferous pipes

Grib 66see also kimberlites and

lamproitesdiamonds, in kimberlites 35Diavik mine, construction 440,

441, 443environmental agreement 440,

443Diavik mine, exploration 438–43,

439A154N pipe 440, 441, 441A154S pipe 440, 441, 441

evaluation 440, 443geophysics 440indicator minerals 438, 440resources and reserves 441

Diavik mine, projected cashflow442

differential thermal analysis27

digital image processing 104,110–15

digital imagescharacteristics of 107, 109image parameters 110, 110pixel parameters 109–10

diopside, chrome 421, 422, 423,dipole–dipole (Pant’s-leg) anomaly

140, 140discount factor 267discount rates, adjustment of 267,

267discounted cash flow rate of

retum (DCF ROR) 266B,268–70

dispersion fans 374–5dispersion trains, glacial 173,

174disseminated deposits 35disseminated precious metals

382–412deposit types 382–8geochemistry 385geophysics 385mining and metallurgy 385–6Trinity Silver Mine, Nevada

386–412distributions, normal and

asymmetrical 200–1Dominion Group 320downhole motors, as directional

control devices 229downhole pulse EM surveys

148–9, 367–8, 368drag bit drilling 285drag bits 285drainage patterns, on aerial

photographs 123, 124, 125drillhole information,

interpretation and plottingof 86–90, 87–90

drillhole patterns 84–6, 86,229–30

Busang 102drillhole samples, logging of

225–8

Page 489: Introduction to mineral exploration

472 INDEX

drilling ambiguities concerninginterpretation 86–90, 90

Cooke Section, West Rand350–2

Doris Creek vein gold deposit,Western Australia 87

Kidd Creek Mine 372Stud Farm 284–5Trinity Silver Mine 389–90,

390when to start, when to stop

90–3Witwatersrand 335–6, 335

drilling contracts 228–9drilling patterns 86, 87–91drilling programs

monitoring of 86, 89–90setting up 84, 86

drilling targetsfinding of 70–84geological mapping 72laboratory program 76–8mapping and sampling of old

mines 72–4merging data 83–4organization and budgeting

70–1physical exploration 82–3prospecting 78–82topographical surveys 71–6

DTM 129, 181, 189Dummett, Hugo 132dykes

intruding Witwatersrandsediments 326

rhyolitic 393

Earth Resources TechnologySatellite (ERTS) 105

East Rand Basin, gold orebodies37–8, 38

Ecca Group 326economic climate, affecting

mineral extraction 256, 258Ecstall orebody see Kidd Creek

Mine, Ontarioeddy currents 141Ekati mine

environmental report 68Ekati mine construction 437

ice road 437Ekati mine, exploration 431–8

drilling 435–6environmental study 436–7

evaluation 436geophysics 435, 435indicator minerals 433–5,

434Koala pipe 436, 438Panda pipe 436, 438

electrode arrays, resistivitysurveys 137–9, 137

electromagnetic prospectingsystem 141–5

electromagnetic spectrum 108electromagnetic surveys 141–5electron probe microanalyzer

26–7, 78electronic distance measuring

(EDM) equipment 71–2elemental analysis 160–2, 161,

161elemental associations 155–7,

156, 171Elmcik High 302Elsburg Reefs 347, 350–7, 357

Massive and Individual 353Elura, geophysical signature 127Emirli Antimony Mine, gold at

61environmental concerns and

considerations andexploration areas 13–14,67–8

goldfields 334Trinity Silver Mine, Nevada

397, 399environmental cost 259environmental impact 67–8, 259environmental impact statements

67–8environmental permits 13–14epigenetic deposits 40equal variance lines 209equity finance 274Ernest Henry mine 67Ertsberg copper deposit, Irian Jaya

78Esquel prospect, Argentina 69estimation variance, geostatistical

methods 240evaluation techniques 199–252

drilling methods 218–30geotechnical data 243–9kimberlites 425–6, 426mineral resource and ore

reserve estimation, gradecalculations 230–43

pitting and trenching 218sampling 199–218use of Gy’s sampling formula

250–2BEvander Sub-basin 339exchange rates 271exhalites, cherty 365exploration expenditure 17–18,

17factors governing choice of area

15–18, 53exploration geochemistry

155–78analysis 160–2follow-up sampling 177interpretation 162–4Kidd Creek Mine, Ontario

374–5, 378planning 155–9reconnaissance techniques

164–77in search for VMS deposits

364–5exploration geophysics

Kidd Creek Mine, Ontario373–4

see also geophysicsexploration methods,

Witwatersrand 334–44drilling 341, 343gravity and discovery of Orange

Free State Goldfield 338–9,340

magnetics and the West Witsline 336–8

retreating old mine dumps343–4

surface mapping and down-dipprojection 335–6

exploration modelsWitwatersrand 333–4

exploration planning 53–61budgets 55–7geological models 40–5host country factors 53–4organization 54–5, 55

exploration programsSoma Lignite Basin 295, 297–8

exploration spending, corporate55–6, 56

exploration strategy, choice of53–4

exploration techniques,integration of 367–9

Page 490: Introduction to mineral exploration

INDEX 473

failurebrittle 395footwall 246–7Soma Basin in rocks

fan delta complexes, Central RandGroup 322

faults and faulting 285, 344basin and range 393block 322strike-slip 300, 393thrust 393Witwatersrand Supergroup 326,

344feasibility studies 11, 12, 262–5,

263inflation 271–2mineral projects finance 272–6mineralisation valuation of

260, 262risk 270–1Soma Lignite Basin 310–19types of 262–5, 263valuation and selection criteria

265–71, 266Bvalue of 253–60

field tests 80–1financing

of mineral projects 272–6, 273Btraditional 272

fining-upwards sequence 302Finsch mine, South Africa 428,

429Fipke, Chuck 431, 432fire assay 396flight lines 118, 119float mapping 82footwall anomalies 362–5footwall failure, Soma Basin 295footwall stability, Soma Basin

308–9Fort a la Corne 415forward modeling 134–5froth flotation circuit 257fundamental error (FE) 204–15,

213variance of, precious metals

214–15, 215Bfuzzy logic 194, 195, 196

gamma logs 297–8gamma ray spectrometer 125–6gangue 4, 19gangue minerals 19, 25, 25, 34Gannicott, Robert 440

gaps (between goldfields) 322, 334garnets 421, 423, 422

G10 423, 424G9 423, 424

gases, sampling of 175geobotany 175–6geochemical analysis

analytical methods 160–2sample collection and

preparation 160geochemical exploration for

diamondsconventional 424indicator mineral chemistry

423–4, 424indicator mineral preservation

423indicator minerals 421, 423,

422sample preparation 423

geochemical maps 157geochemical surveys 61, 83–4,

157, 164–78regional 58reporting of 157, 159

geochemists 151, 155geological and geophysical data,

integration of 150geological information,

background 58geological logging 225–8geological mapping 72

computerized drafting 72geological models, exploration

planning 41–5geological surveys 58geologist, role of, geophysical

exploration 150–2geology, airborne 61geophysical exploration for

kimberlites 424–5EM 424–5flight spacing 424–5gravity 424magnetics 424seismic 424

geophysical logging, down-the-hole 148–9, 298

geophysical methods 83–4,126–53

airborne surveys 58, 61, 127–9,128

electromagnetic methods (EM)141–5, 365–6, 373–4

gravity method 133–5, 134, 365ground radar 147induced polarization (IP)

139–41magnetic surveys 59, 129–34,

365, 367radiometrics 61, 63–5, 64,

135–6resistivity 136–9, 286seismic methods 58, 145–7,

285–7spontaneous polarization (SP)

139transient electromagnetics

(TEM) 144–5VLF and magnetotellurics 83–4

geophysical signatures, VMSdeposits 365, 367

geophysical surveys 83–4, 149–50geophysics

airborne 58, 61, 373–4, 374borehole 148Cliffe Hill Quarry 285–7Cooke Section, West Rand 350ground 83–4, 149regional 58Trinity Silver Mine 389

geostatistical methods, tonnagecalculation 235–41

geotechnical data 243–9hydrogeology 248–9quantitative assessment of rock

244–9, 244subjective 243, 245

geotechnical investigation, SomaLignite Basin 307–9

GIS 83–4, 104, 179definition 186layer representation 186–7software 186South Devon gold prospect

187–97glacial boulder trains 58glacial deposits, blanketing effect

of 374–5, 378glacial dispersion trains 82, 82global estimation errorgold 21, 27, 29

in alkalic volcanics 383alluvial 37, 102conglomerate, Witwatersrand

320–57deposit discovery rate 15, 15epithermal deposits 383

Page 491: Introduction to mineral exploration

474 INDEX

gold (cont’d)errors in assays on prepared

sample 216, 216fraud 100–2grade controlled by

sedimentological factors 325,327, 327

grain shape 101history since World War II 10in massive sulfide deposits 362N Ireland 83–4pathfinder elements 385placer gold 45price 10, 10western Turkey 59–62see also disseminated precious

metals; porphyries, gold-rich;precious metals;Witwatersrand Basin

gold films 382gold reefs 328–9

sheet and channelized 326, 328East Rand Basin 37–8see also Witwatersrand Basin

gold standard 337gold and uranium distributions,

sedimentological controlson 326, 327

goldfieldsdetection of extensions to 333–4environmental concerns 334Witwatersrand, location of 320

gossans 78–82examination of 79–80, 172gold-rich 82recognized from satellite

imagery 114, 116relic textures 79–80, 172

government action, affectingprices 13–14

government controls 258–9GPS (global positioning satellites)

129grab sampling 207graben 302grade 73, 76grade calculation methods

computational 241–3conventional 241

grade control 396, 396grade cut-offs 73–6, 76

Trinity Silver Mine 402–4grade estimation, Trinity Silver

Mine 399–402

gradient arrays 137, 140in IP reconnaissance 140

grain size, crushing and grindingcosts 28–9

graphite furnace AAS 385graphitic sediments, problem of

367Grasberg mine, Indonesia 100gravity anomalies, St Helena gold

mine and discovery of OrangeFree State Goldfield 338–9,340

gravity methods and surveys133–5, 134, 343

geophysical interpretation 134greenstone belts 367, 370–1

sourcing some Witwatersrandgold 330

see also Abitibi Greenstone Beltgrid sampling 29groundwater movement 248–9grouting 298Gurney, John 433Gy’s sampling (reduction) formula

208Buse of 250–2B

hammer drills 219hand sampling techniques

211–13hanging wall anomalies 365Hardgrove Grindability Index

(tests) 297, 310Harmony Formation 391health and safety 69heap detoxification, natural and

chemical 397–9, 398heap leaching (cyanide leaching)

see cyanidationheavy minerals 37, 145, 167, 169,

170helicopters 129

electromagnetic system 372for radiometric surveys 135

hematitization 41homogeneity and heterogeneity

200Hooke’s Law 245Horne Orebody, Noranda, Quebec

362host country stability 258–9host rocks

igneous 38–9Kidd Creek area 375–7

metamorphic 39sedimentary 36–7

hot spring deposits, preciousmetals in 383

hydraulic conductivity 249hydrofracturing 395hydrogeochemistry 173, 175hydrogeology 248–9

Soma Lignite Basinhydrothermal deposits 78hydrothermal solution, in

formation of VMS deposits362–4

hydrothermal vents see blacksmokers

Hyperspectral imagery 117

Iberian Pyrite Belt 363ICP-ES see inductively coupled

plasma emissionspectrometry (ICP-ES)

ICP-MS see inductively coupledplasma sourced massspectrometry (ICP-MS)

igneous activity, alkaline 39igneous host rocks 38–9Ikonos imagery 104, 105,

116–17ilmenite 37, 175, 421, 423,

422image enhancement 113image processing

software 104image restoration 113Impala Platinum 12, 13impact diamonds 421in situ stress determination

247–8in-seam mining 311–12, 313Inco Ltd. 12, 13indicator plants 175–6Indonesia 100–2induced polarization (IP) 139–41

time-domain effects 140use of gradient array 140

INduced PUlse Transient system(INPUT) 144, 144

inductively coupled plasmaemission spectrometry (ICP-ES) 78, 80, 161, 161, 162

inductively coupled plasmasourced mass spectrometry(ICP-MS) 78, 161, 161, 162,385

Page 492: Introduction to mineral exploration

INDEX 475

industrial minerals 4, 8, 10–11, 17definition 4exploration expenditure 17, 17importance of 4, 8, 10–11

inflation 271–2insider trading 102Instantaneous Field of View

(IFOV) 109instrument splitters 213intergrowth pattems 30International Geomagnetic

Reference Field (IGRF) 131interval estimates 201–3

one-sided 202–3two-sided 202, 202

inverse distance weighting 311inverse power of the distance

(l/Dn) method, reservescalculation 235, 410

ion chromatography 162ion probe microanalyzer 26–7Ireland

impact of tax holiday onexploration 14

iron ore 21price 9, 9production 5supply and demand 4

ironstonesoverlying sulfides 80Phanerozoic 36–7

isopach mapsoverburden, Stud Farm 285–7,

286reefs, Witwatersrand 351, 351Soma Lignite Basin 304–5, 305

isotopic analysis 162

jasperoid 383, 385Jennings, Chris 438, 440joint ventures 12–13, 93

Kyrgyzstan 66Jones Riffle 213, 214junior mining companies

see mineral explorationcompanies

JORC code 3–4

Kaapvaal Craton 320, 321kaolin deposits 39kaolinization 41, 43Karoo Basin, South Africa 63, 63

search for sandstone-hosteduranium deposits 63–5

Karoo Supergroup 63Kennecott Canada 440, 443Kidd Creek anomaly 371Kidd Creek Mine, Ontario

exploration geochemistry374–5

exploration geophysics 373exploration history 371–2genesis of orebody 378–9mine geology 375–8mining operations 379–80regional setting 370–1rock mechanics 381

Kimberley 415blue ground 427Cecil Rhodes 427discovery 426–7kimberlite pipes 428, 429

Kimberley Reef 322kimberlites 78, 415–20

dating 418–19diamond-bearing, exploration

for, Botswana 173, 175, 427,429

dykes 415inclusions 417–19morphology 415–16, 416,

417petrology 416size 415, 416, 417, 428structural control 419–20xenoliths 417–19

King’s Mountain Operation Tin–Spodumene Belt, NorthCarolina 19

Klerksdorp area 323, 338kriging 239–40, 314, 341kriging variance 240–1Kumtor gold mine 66Kupferschiefer – copper shales

37Kuroko-type VMS deposits 358,

360, 361, 361

Lac de Gras 432, 433, 434, 435–40lacustrine sediments 415lake sediments, sampling of 167lamproite pipes 415

geology 416land acquisition 65–7Landsat imagery 64, 105–12, 111,

112, 114advantages of 107applications of 115–16

Landsat Systemavailability of data 113digital image processing 110,

113–14extracting geological

information 115–16spectral bands 108

laterite 78lead 21

deposit discovery rate 15, 15price 9, 9production 5supply and demand 4

lead (Pb) anomalies 389lead-zinc mineralisation

Mount Isa, Queensland 36–7Sullivan, British Columbia 36,

37legal disputes and land acquisition

66–7leveling 72lignite deposits see Soma Lignite

Basin, Turkeylignite quality, Soma Lignite

Basin 309–10, 310lignite reserve estimates, Soma

Basin 314–15confidence in 315estimation methods compared

314limestone hosts 37lineaments 115, 125linear trends 382liquation deposits 39lithium 21loans, interest on 274location, of exploration areas

15–16lodes 33–4

see also veinslogistic regression 196London Bullion Market 6London Metal Exchange 6long vs. short run considerations

269–70Louvicourt VMS deposit, Quebec

367–8, 369

McArthur River deposit,Australia, grain sizeproblem 28

maghemite 131, 367magnetic anomalies 131magnetic field 130–1, 130

Page 493: Introduction to mineral exploration

476 INDEX

magnetic maps, interpretation of131–2

Euler deconvolution 133Werner deconvolution 133

magnetic storms 131magnetics and magnetic surveys

airbome 127–9ground and airbome 129–33,

338–9mapping of Yilgam block 36,

367and the West Wits line 337–8,

337magnetite 39, 131magnetotellurics 145Main Reef 322, 336–7Malmani Subgroup 323manganese

price 9, 9, 22production 5

marker bands 350market conditions, and mineral

extraction 256, 258markfieldite, Charnwood Forest

281–91, 282Marmoraton skarn iron deposit

29Marsfontein mine, South Africa

428, 429massive sulfide deposits

Besshi-type 360Canada 177–8classifications 358, 359, 360,

361Cyprus-type 360electromagnetic prospecting

system for 61, 141–6exhalative, associated with

black smokers 364exploration model for genesis

of 364in island arc sequences 358Kidd Creek Mine 370–81Kuroko-type 358, 360, 361, 361Primitive type 360shape of 363use of transient electromagnetic

surveys 146volcanic-associated 38–9, 38,

358–64zoned 362

massive sulfide lenses 358, 361measurements while drilling,

developments in 224

Mercia Mudstone 283, 285, 287mercury 22, 25mercury impurity, use of

microscopy 32metal prices 9–11, 9–11

and mineral prices 10–11metallogenic maps 59Metallogenic Map of Europe 60

Turkey 60metals

optimum target 10relative proportions in

orebodies important 25–6supply and demand, cyclicity

10world production 4–9, 5, 6

metamorphic aureoles 39metamorphism 326

contact 392regional 392

microdiamonds 426, 427microfractures 246, 393microscopy, examples of use of in

ore evaluation and mineralprocessing 31–2

Middle Elsburg Reefs 347–56mine dumps 19

re-treated, Witwatersrand343–4

mineral deposit models 40–51classification of Cox and Singer

46–51pitfalls of 45

mineral depositsgeology of 33–9mineralogy of 19–32

mineral economics 3–11mineral exploration 12–18

companies 12–13data handling 179–97discovery rate 16, 16, 18economic influences 18exploration expenditure 17–18,

17exploration productivity 15–16,

16stages 12success rate 14–15time span 12

mineral identification 26–8mineral liberation size 28, 30mineral and metal prices 6–11mineral occurrence maps 58–9

Turkey 62

mineral processingTrinity Silver Mine 397, 399

mineral projectsfinancing of 272, 274–5valuation and selection criteria

253–60mineral rights, ownership of 65–6

Russia 66Turkish Witwatersrand 332

mineralisationacquisition value 268assessment of 89–90, 91continuity of 90, 149definition in JORC code 4geological controls on 86, 90geological relationships of 72Kidd Creek area 377–9primary, trace element

signature preservation 155,156, 171–3

quality of 256relative value of 262–5Trinity Silver Mine, Nevada

394–5valuation of 260–70value of 253–60Witwatersrand Supergroup,

relation to sedimentology326–8, 327

mineralisation envelope 232mineralogical form 19, 25mineralogical investigations 19,

25–32economic significance of

textures 30–1quantitative analysis 28–30sampling 26

mineralscolors of in outcrop 80factors in economic recovery of

11–12identification of 26–8interlocking 30–2

mines 11closure 11life of 253–4surface mine design, Soma

Basin 316Mines & Quarries (Tips)

Regulation 291report 291

mininggovernment aspects 69social aspects 68–9

Page 494: Introduction to mineral exploration

INDEX 477

mining at depth, constraints on333

mining companiesmajor 12–13, 13junior, see mineral exploration

companiesmining equipment, surface mine,

Soma Basin 319mining methods, surface mine,

Soma Basin 315–16mining rights see mineral rightsMir mine, Russia 428, 429“mise-à-la masse” method 149Misery Point, Lac de Gras 433modal estimates, visual, in the

field 29Mohr diagrams 247, 247molybdenite deposits, Colorado

35molybdenum 35Mulberry Prospect, re-

examinations of 93, 94, 96,97, 99

British Geological Survey 93Central Mining and Finance Ltd

(CMF) 93, 97, 98, 99Consolidated Gold Fields (CGF)

93, 94Noranda- Kerr Ltd 95, 96

multi-client surveys 367multi-temporal photography 124

net present value (NPV) 268–9neural network 194, 197neutron activation analysis (NAA)

76, 161, 162, 385neutron tool 297Neves-Corvo, Portugal

base metal deposit discovery13, 133, 134

VMS deposits 362Newmont Mining Corp. 12, 13nickel 22, 25, 80

deposit discovery rate 15, 15nickel-copper sulfide orebodies

29non-core drilling 218–20non-metallicsNoranda area, Quebec, massive

sulfides 362–3, 361, 367Norilsk Nickel 12normal distribution 163North Anatolian Fault 298“noseeum” gold 382, 385

oceanic crustsea water circulation through

364Ok Tedi Mine, Papua New

Guinea, porphyry deposit 82,273B, 383

old minesmapping and sampling of 72–3,

76OPEC shock, effects of 8open pit mining 33, 362, 386

Kidd Creek 378–9Soma Lignite Basin 310–11,

316–19operating costs and capital costs

261B, 264–5operating margin to initial

investment ratio 266B, 267Orange River 427ore 3–4, 231ore deposits

size and shape 33–4ore minerals 3–4, 19, 20–4ore reserves 3, 231

see also reserve estimationsore shoots, ribbonorebodies 3

concordant 36–9discordant 33–6igneous host rocks 38–9irregularly shaped 35–6low grade, large scale mining of

33metallic, significant discoveries

15metamorphic host rocks 39nature and morphology of 33–9regularly shaped 33–5residual deposits 39sedimentary host rocks 36–8tabular 33–4terms used in description 34,

34tubular 34–5

orientation surveys 157, 159orthomagmatic deposits 39outcrops, weathered, recognition

of 78–80overburden

dispersion through 155, 157Soma Lignite Basin 305–7,

306Stud Farm 283–4, 286, 286–7see also glacial deposits

overburden geochemistry 167,169, 173

surface soil sampling 167,169–72

transported overburden 172–3overburden ratio maps see

stripping ratio mapsoverburden sampling 167, 169–73

deep overburden sampling172–3

overcoring principle 247–8, 248overlaying (data), usefulness of

83–4see also GIS

packer tests, in boreholes 248Pacific Rim Metallogenic Belt

391palaeoplacer theory (gold) 329–30

modified 330panning 166–7, 168Panvlakte Fault 351parallax 120pathfinder elements 156, 385pay shoots

control on 336extensions of 335–6see also ore shoots

payback 266, 266B, 274pegmatites, grid sampling 29penalty payments 276percussion drilling 86, 219–20,

228B, 396permeability 249Pershing County, Trinity District,

geology of 385, 387, 391structural phases 391

photogeology 117–25photogrammetry 125photointerpretation equipment

121piezometric surface 249, 249pinch-and-swell structures 34,

34pipes and chimneys 34–5, 35, 396pitting 82–3, 218placer deposits 37–8

diamond deposits 420, 421gold placers 320, 322, 326

placers, sheet and channelized326, 327

plastic flow 247plastics, and metal substitution

8–9

Page 495: Introduction to mineral exploration

478 INDEX

Platinum Group Elements (PGE)8, 22

plutonic hosts 39plutons

delineation of mineralized177

political factors, affectingexploration 14

pollutionWitwatersrand, little concem

over 334population and sample 199–200porosity 248porphyries 35

copper 273gold-rich 383see also Ok Tedi Mine

posterior probability 194, 195,195

potash 22Potchefstroom Gap 322, 323precious metals

analysis of 162disseminated 382–6epithermal deposits of 383exploration of 385

pre-feasibility study stage inmineral explorationprograms 12

see also feasibility studiesPremier mine, Cullinan, South

Africa 423, 427discovery 427indicator minerals 423

preparation errors 204, 209–11pressure leach extraction, for gold

386price determination

by government action 7–8by recycling 8by substitution and new

technology 8by supply and demand 7, 46see also metal prices

primary dispersion 155, 156primary halo 172, 362, 365product price see commodity

pricesproduction, world mineral

by tonnage 5by value 5

production rate 254, 256profitability, of Witwatersrand

gold 330–1, 331

project development, differingemphasis of banks andsponsor 274–5

project evaluation 265–7, 266Bproject finance 274–6

risk in 275–6project implementation, Stud

Farm Quarry 291–2projections, map 186prospecting 78–82protore 38pyrite 329, 362

arsenianpyrrhotite 31, 131, 362

quantitative analysis of minerals26–30

grain size and shape 28–9modal analysis 29–30

quarry planning 289–91Quickbird imagery 104, 105,

117

radar imagery 105, 124radiometric surveys 61, 63–5,

135–6, 136airborne 64–5, 64carborne 62–3, 63

radon 135random sampling 198rare earth elements (REE) 22raster representation 179–80,

181ratios, digital image processing

113–15reconnaissance, aim of 61reconnaissance exploration 12,

61–5desk studies 12, 57–61exploration planning 12, 54–61exploration players 53land acquisition 65–7

reconnaissance projects 61, 63–5reconnaissance techniques,

exploration geochemistry164–77

recovery 28recycling prospects 93–9remote sensing 104–14

data collection 104–5kimberlites 424Landsat System 105–16

replacement deposits, irregular35–6

reserve estimationslignite, Soma Basin 314–16Stud Farm Quarry 289Trinity Silver Mine 409–12Witwatersrand Basin 344–5

reserves, vs. resources 3–4,231–2

residual deposits 39resistivity surveys 136–9resistivity tool 297resource evaluation 197

software 197–8resources, mineral

indicated 3measured 3

retained profits financing 274retreat mining 353reverse circulation drilling 220,

221rhyolite 363, 389, 393, 395

as a host rock 395rhythmic layering 39riffle splitters 213, 214Rio Tinto plc 12, 13risk 270–1

in project finance 275–6risk analysis 271rock, quantitative assessment of

243–5, 244rock geochemistry 177–8, 365

mine scale uses 177–8rock mass rating (RMR) 244rock mechanics 381Rock Quality Designation (RQD)

243, 244rock sludge 226–7rock strength 243, 245rockbursts 333room (bord) and pillar mining

333, 353rotary drilling 219rudaceous hosts 37–8run of mine product (ROM)

309–10Russian diamond exploration

429–30Grib pipe 430

safety line, sample reduction209–11, 210

variation of 211St Austell Granite 93salt 23salting, core 101

Page 496: Introduction to mineral exploration

INDEX 479

sample acquisition 206–8,216–17

after sampling 217before sampling 216

sample density–volumerelationship 206

sample extraction 205–8sample reduction, graphical

solution of 209–11, 211sample reduction system, design

of 209–11sample spacing, soil collection

169sample testing, Stud Farm 288–9sample volume–variance

relationship 205–6samples and sampling 26

analytical errors 215–16borehole sampling 226containing native gold and

platinum, preparation of215–16

follow-up sampling 177for grade control 396–7introduction to statistical

concepts 199–203and the laboratory program 76,

78lignite cores 298point and interval estimates of

201–3problems in transported

overburden 173sample acquisition 216–17sample extraction 205–8sample preparation 209–15sampling error 203–5sampling procedures 205–7splitting of during reduction

211–15, 212, 213, 214underground, old mines 72–3

sampling strategies, acrossmineralised veins 73, 75, 76

satellite imagery 58, 104–17different resolution and

frequencies 106see also Landsat imagery

satellite orbits 107, 107scanning electron microscopy

(SEM) 27, 78Schlumberger arrays 136–8, 137scintillometers 135secondary dispersion 156–7sedimentary host rocks 36–8

seismic surveys 145–7, 147reflection surveys 145–6refraction surveys 146–7, 1482-D 339, 341, 342, 3433-D 339

Sekkdy Formation 298–9selection errors 203–5

preparation errors, possiblesources 209–15

semi-variograms 235–40interpretation of 239–40Trinity Silver Mine 405, 408

sensitivity analysis 268–9, 269Bseparation process, lead and zinc

sulfides 257Sen Basin 302sericitization 41, 383, 394shafts, sunk through water-

bearing formations 337shear strength 246–7shear stress 246–7shear zones 395

recognition of 145shearing 246–7sheeted vein complex 93side-looking airborne radar (SLAR)

58, 124silicification 78, 382, 385, 394silver 23

deposit categories 382–5Nevada deposits 382–5, 392Trinity Silver Mine, Nevada

392silver deposits, classifications of

391silver doré bullion 397skarn deposits 35, 36, 39social license to operate 68–9soil collection, temperate terrains

169–70, 171soil geochemistry surveys 172,

173Solomon Islands, airborne

radiometric survey 152Soma Graben 300Soma Lignite Basin, Turkey

294–319data assessment 303–7depositional model for 302–3exploration programs 295–8geology of 298–303geotechnical investigation

307–9lignite quality 309–14

lignite reserve estimates314–15

structure 300–2surface mine evaluation 315–19

sonic log 298South Deep mine, West Rand

353–7, 354, 355, 356wide orebody mining 353

South Devon gold prospect187–97

Arc-SDM 192Boolean algebra 193Boolean methods 192, 193complex methods 192, 194, 194data display 187–8data overlay and buffering

191–2DEM 189, 189fuzzy logic 194, 195, 196geochemical data

representation 189–91, 190,191

lineaments 192, 192logistic regression 196polygon number 188, 188posterior probability 194, 195,

195remote sensing 189, 190

spatial component of data 179splitters, asymmetrical and rotary

213, 214spontaneous polarization (SP) 139SPOT 116staking claims 433statistics

in geochemical exploration162–4, 163, 164

parametric and nonparametric200–1

Steyn Reef 322stockwork breccias 395stockworks 35, 35, 38, 39

Kidd Creek area 377with zone of acid leaching 383

stopingbreast stoping 322narrow orebody 353

strain 277–8strain gauges 248stratified sampling 207stratiform deposits 36–7, 358stream sediment sampling 36–8,

82sample collection methods 165

Page 497: Introduction to mineral exploration

480 INDEX

stream sediment surveys 165–7,166, 169

stress 246–8stress–strain curve 247strip drill logs 86, 89–90, 89stripping ratio

maps, Soma Lignite Basin305–7, 306

structural information, in mineplanning 244–5

structural maps, from Landsatimagery 115

structure contour maps, SomaLignite Basin 303–4, 304

Stud Farm (New Cliffe HillQuarry) 281–93

quarry planning 289–91reserve estimation 289

sub-bottom profiling 145sulfides

authigenic 329direct detection of 367weathered 78–9

sulfur, in coal, use of microscopy32

supergene enrichment 81–2, 81surface mining

lignite quality 310–11selection criteria 309–10Soma, evaluation of 304–7

surface rights, ownership of 65surface soil sampling 167, 169–72survey lines, location and height

149surveying methods 71–2sustainable development 13–14,

67–9

tape and compass surveys 73, 74, 75target testing stage in exploration

programs 12taxation

affecting choice of explorationarea 14

of gold mining 331TEM see transient

electromagneticstensile strength 245terranes, favorable for VMS

deposits 371textures

on aerial photographs 123economic significance of 30–1relic 78–80, 80

Thematic Mapper 105–11, 105,106, 108, 110, 114

recognizing hydrothermalalteration 115–16

38th Parallel Lineament 115Thomas, Grenville 440Thringstone Fault 281–2thrusting 391, 393time value of money techniques

267–9Timmins massive sulfide deposits

144tin 23, 25, 32

noncassiterite tin and problems25

tin mineralisation 171–2tin ores, use of microscopy 32tin production

international tin agreement andprice fluctuations 8

tin-tungsten mineralisation,Ballinglen, geochemistry of169, 171–2, 171, 172, 173

titanium 10, 24, 29Tli Kwi Cho kimberlite 440tonnage calculation methods

233–40classical statistical 233–4conventional methods 234–5geostatistical methods 235–40

topographical surveys 71–2, 75topography, aerial photographs

122–3Total Initial Capital (Capex) 266Btotal sampling error (TE) 204–5

comparison of calculated andobserved 213–14

trace elements 177background concentrations

158trace metals, temperate terrains

169tracking and data relay systems

(TDRSS) 107trackless mining 333, 352, 354trade union policytransient electromagnetics (TEM)

144–5airborne 144–5ground 145

transition metal anomalies 365Transvaal Supergroup 323, 347,

350trenching 82–3, 83

triangulated irregular network(TIN) 181

triangulation 71triaxial compressive strength test

245–6, 246Trinity Silver Mine, Nevada

386–412defined by assay cut-off 396development and mining 395–7drilling 389–91environmental considerations

397–9estimate for mine planning

410exploration 387–91factors affecting reserve

calculations 411–12geochemistry and geophysics

389geology alteration of 394geology of 392–5grade estimation 399–409mapping 389mineral processing 397mineralisation 394–5ore reserve estimation 409–12outlier and cut-off grade

evaluation 402–4spatial distribution 404–9stratigraphy 392–3structural geology 393–4

tungsten 169, 171–2Turgut Formation 298, 301Turkey 59–62

epithermal gold deposits59–62

mining rights 59, 295new mineral law 59see also Soma Lignite Basin,

Turkey

unconformities 283unconformity vein deposits

41–5underground mining

Kidd Creek 379–81Soma Basin, selection criteria

311–12underground mining costs 254,

256univariate statistics, for grade

estimation 399–404upward fining cycles 302uraninite 329

Page 498: Introduction to mineral exploration

INDEX 481

uranium 24, 38, 320as a byproduct 320controlled by sedimentological

factors 329radiometric surveys for 135–6

uranium depositsdeposit model for 41–5, 41,

42–3found by airborne radiometrics

61–5sandstone hosted, search for,

Karoo Basin 62–5, 63unconformity-related 41–5Yeeleerie, Western Australia,

discovery of 61, 136US Borax exploration, Trinity

Mine 387–91

Vaal Reef 322, 329Vaal Reef Mine complex 338Valmy Formation 391valuation of prospects 99–100vanadium 24variance

estimation variance 240–1of fundamental error 215Bobserved 209reduction of 208

variance units 201vector representation 179–80, 181

advantages of 180vegetation

in exploration geochemistry176–7

and land use, from aerialphotographs 123–5

veins 34Ventersdorp Contact Reef (VCR)

322, 337South Deep mine area 353, 354

Ventersdorp Supergroup 322, 338,350

vertical crater mining 353, 355very low frequency

electromagnetic resistivitysurveys 84

VLF magnetotellurics 145VMS see massive sulfide deposits,

volcanic-associatedVoisey’s Bay mine, Labrador 100volcanic complexes, Abitibi

Greenstone Belt 371volcanic domes

and Kuroko-type deposits 361,363

volcanic hosted systemsacid sulphate type 383adularia-sericite type 383alkalic 383

volcanic hosts 38–9volcanics and volcanism

alkaline 383bimodal 391felsic 38–9Kidd Creek 378rhyolitic, brecciated 314Tertiary 393Trinity Silver Mine 393, 395wall rock alteration 362, 368see also alteration

Wales, Snowdonia National Park,copper deposit 13

wasting assets, orebodies as 14water table 248–9Wenner arrays 137–9, 137, 138West Rand 336–7

surface geology 346, 347West Rand Fault 337West Rand Goldfield, exploration

and development of southernpart 345–57

West Rand Group 322exhalative gold reworked 330

West Wits Line 322, 323, 327,336–8

down-dip potential 338Western Deep Levels 338Wheal Prosper, Cornwall 93, 95wide orebody miningwindow statistics, silver grades

404–5, 406, 407

Witpan placer, grade-locationrelation 328

Witpoortje Horst 347, 348, 351Witwatersrand Basin 320–57,

321, 322economics of mining 330–1exploration methods 334–44exploration models 333–5exploration and development

southern part, West RandGoldfield 345–57

geology 320–30reserve estimation 344–5

Witwatersrand Supergroupmineralisation 326–8mineralogy 328–9stratigraphy of 320–2structure and metamorphism

326theories of genesis 329–30

xenoliths 416dating 418–19ecologite 418peridotite 418

X-ray diffraction (XRD) 26X-ray fluorescence (XRF) 162,

161, 161

Yatagan Formation 299yellow ground 415Young’s modulus 245

Zambian Copperbelt 176host rocks 37

Zechstein Sea 37zinc 24

price 9, 9production 5supply and demand 4see also copper-zinc deposits

zinc loss, use of microscopy 32zirconium 24zonation

in massive sulfide deposits362

Page 499: Introduction to mineral exploration