abstract - today at minesinside.mines.edu/userfiles/image/geophysics... · abstract this report...

Abstract

AbstractThis report summarizes the data acquisition, processing, and final results from the

Colorado School of Mines Department of Geophysics 2017 summer Field Camp. Thepurpose of Field Camp is to conduct a geophysical investigation into the geologic andgeothermal systems of Archuleta County, Colorado. Beginning May 14th, students workedto acquire data in Archuleta County using a variety of different geophysical methods,assisted by a dedicated team of TAs, faculty, and technical volunteers. Following up thedata acquisition in Pagosa Springs, the students returned to Colorado School of Minescampus to begin processing and interpreting the collected data. Results of the acquisition,processing, and interpretation were assembled in this final report and a final presentationon June 9th at Colorado School of Mines.

The acquisition process focused on two sites to conduct the investigation. The mainsite was along County Road 542, also known as Montezuma Road, near the town ofChromo, Colorado. The main line survey was 12 kilometers, beginning at the intersectionwith Service Road 653 and continuing to connect with the 2015 Field Camp main linenear the intersection of County Road 359 and County Road 391. The student sitewas at Reservoir Hill, located just east of downtown Pagosa Springs. Reservoir Hill wasinvestigated by the 2012 Field Camp, and was revisited this year for further understandingof a potential fault. Students used various methods to investigate these sites, includingDC resistivity, deep and hammer seismic, spontaneous potential (SP), differential GPS,magnetotellurics (MT), gravity, and electromagnetics (EM).

The results at the Chromo site revealed a discontinuity in the basement topped witha low density material. This was located off the west side of a minor dike. Data for thisresult was initially provided by deep seismic, and supported by findings from gravity andMT. According to DC and SP data, the two volcanic dikes that intersected the main linedid not have an observable effect on the regional fluid flow. At the student site, the DClines showed evidence of a large conductive anomaly within a relatively resistive layer.This data was supported by information from EM, while the SP data showed evidence ofupwelling at the same location.

3

Based on the results from the main line, it is possible to conclude that there is evidenceof a fractured or faulted crystalline basement. Seismic, MT, and gravity data support theexistence of this geological structure. The inconclusive data surrounding the volcanicdikes is grounds for further study. Student site results indicate that there is most likelya fault located on Reservoir Hill. This result was supported across multiple datasets,including results found by previous field camps. The final results from both the mainline and the student site imply the possibility of a fractured or faulted subsurface thatcould be indicative of regional and local fluid flow.

DisclaimerThis report and its contents are derived from a summer field camp for undergraduate students inthe Department of Geophysical Engineering at the Colorado School of Mines. The primary objectiveof this field camp is education, focusing on the instruction of applied field geophysics. All datacontained in this report was acquired, processed, and interpreted primarily by students from theColorado School of Mines. Therefore, all results and conclusions should be regarded appropriately.The Colorado School of Mines and its Geophysics Department do not guarantee the validity of theinformation or results contained in the remainder of this report.

Acknowledgments

The students involved in the 2017 Colorado School of Mines Geophysical Engineering Field Campwould like to recognize the individuals and organizations that provided generous financial, equip-mental, and intellectual support. This valuable experience would not have been possible withoutthe help of everyone involved in the process.

We would like to thank the following faculty and staff for supplying us with the knowledgenecessary to participate in this project. Their time and assistance is an extremely integralelement of our success.

Department Head• Roel Snieder

Field Camp Director• Andrei Swidinsky

Professors• Rich Krahenbuhl• Bob Raynolds• Paul Sava• Yaoguo Li• Ali Tura• Ebru Bozdag• Brandon Dugan

Adjunct Faculty• Bob Basker• Seth Haines• Cici Martinez• Matthew Wisneiwski• Gene Wolfe

Equipment Manager• Brian Passerella

Administrative Staff• Joana Perez• Michelle Szobody

Teaching Assistants• Liz Maag• Joe Capriotti• Stephen Cuttler• Andy McAliley• Aspen Anderson• Hayden Powers

5

We are very grateful for the generous financial contributions provided by:

Colorado School of Mines College of Earth Resource Sciences & Engineering (CERSE)Colorado School of Mines, Department of Geophysics

Halliburton

Chevron

Shell Oil Company

ConocoPhillips

Bob Basker

Anadarko PetroleumCorporation

Society of Exploration Geophysicists (SEG) Foundation

6

We thank the following individuals and organizations for providing equipment, software,services, time, support, and knowledge during our project:

Marvin Johnson and Gene Wolfe of Halliburton deserve special recognition

for their time and expertise with seismic acquisition and on-site processing

Chevron Corporation• Jeremy Zimmerman

Dawson Geophysical• Stuart Wright

• Steve Kite• Mike Kite

Halliburton• Gene Wolfe

Sercel Corporation• Alba Guerrero

Center for Gravity,Electrical & Magnetics

Studies at Colorado Schoolof Mines (CGEM)

Reservoir CharacterizationProject (RCP)

• Ali Tura• Tom Davis

Pagosa Verde

United States GeologicalSurvey

• Seth Haines

Geophysical Technology, Inc.• John Gonzales

7

Additional critical resources and support were provided by:

• CSU/Archuleta County Extension Office (Terry Schaaf, Becky Johnson)

• Pagosa Springs High School (J.D. Kurz)

• Pagosa Springs Middle School (Kristen Hentschel, Linda Reed)

• Wyndham Pagosa

• Town of Pagosa Springs

Finally, we would like to thank the following Pagosa Springs community members forproviding us with access to roads and property to perform our geophysical investigations:

•Glacier Bank (Bruce Penny)

• US Forest Service (Kevin Khung)

• Archuleta County Road and Bridge (Tim Hatch)

• City of Pagosa Springs (Darren Lewis)

• Ken Levine

• Judy Schofield and family

• David L Smith

• Natalie Woodruff

• David Fairclough

• Elma Garcia

8

Colorado School of Mines Geophysics Field Camp 2017 Students:

• Project ManagerKristen Marberry

• Assistant Project ManagerM. Adam Kinard

• Geology and MiscellaneousDana Sirota Group LeaderCourtney BoneFelicia NurindrawatiNadima DwihusnaTravis HastingsMohamed Al ReyamiVictoria FisherSkyler Neumeyer

• MagnetotelluricsDelaney Marsh Group LeaderSean SmithCesar ValasquezJonathan MorsicotoMadison Johnson

• DC Resistivity and Self PotentialChloe Hampton Group LeaderLarsen KronstadChaeli TrostNahjee Maybin

• Hammer SeismicChloe Hampton Group LeaderCam HarveyBenjamin Federspiel

• ElectromagneticsDevon Dunmire Group LeaderChristopher TietzPatrick CarabelloAli J. AlzayerKent Lewis

• Gravity and MagneticsHeather Schovanec Group LeaderBane SullivanMax PaceDaniel Walker

• Deep SeismicIga Pawelec Group LeaderJalen ChampagneJohn WimanYasmine Yus’AimeiKenneth LiKeenan BarkerJing JianSkyler Neumeyer

Unless otherwise noted, photography provided by the following participants of Field Camp:

• Aspen Anderson• Michelle Szobody• Yoagou Li• Joana Perez• Kristen Marberry• Victoria Fisher• Devon Dunmire

• Heather Schovenac• Paul Sava• Iga Pawelec• Stephen Cuttler• Nadima Dwihusna• Larsen Kronstad

Introduction

IntroductionTasked with the prospect of accurately interpreting the geological origin of the town of Pagosa

Springs geothermal system, the 37 undergraduates of the 2017 Colorado School of Mines sum-mer field session enlisted the aid of township personnel, industry analysists, support staff andfaculty. Using statistical data analysis of prior geophysics testing and modeling endeavors, theassembled team produced a survey route that accommodated a north-westerly progression of thecamps 2014 and 2015 seismic main lines.

The 2017 main line, nestled between the northeast quadrant of Archuleta Mesa and theChromo Mountain, was plotted along County Road 542, intersecting two dikes with northeast-southwest strike directions and ending near the intersection of County Roads 542 and 359.This 12 km line employed a variety of geophysical methods in order to investigate the area anddescribe indicators that pertain to underground-fluid motion.

In order to capitalize on the ability of our geophysical survey to produce comprehensive re-sults, the 2017 team selected an additional survey site. Understanding that the collection ofnumerous types of geophysical data has the potential to deliver significantly more evidence thana lone survey site, Reservoir Hill was selected due to its close proximity to the Mother Springand possible supposition to the 2012 field camps characterization studies. Reservoir Hill, morecommonly referred to as the student site, is bordered by the San Juan River, Mill Creek, andHighways 84 and 160.

The significance of the Mother Spring is due to its persistent production of geothermal fluids;this spring is a gauge and an indicator of the locations potential capacity to be harnessed by thetown as a viable and renewable energy resource.

The three survey lines that traversed the hill played host to a multitude of geophysical meth-ods, including hammer seismic, gravity and time-domain electromagnetic field methods. Asdissimilar rocks and buried structures exhibit different values of electrical conductivity, themapping of these variances was used to identify possible anomalous portions of the hill with theanticipation of finding intrusive features that could be tied to the indication of fluid movement.

The significance of the Chromo dikes and previously suggested faulting on Reservoir Hill areof paramount importance when addressing the camps fundamental questions: Can the regionalgeology be seen by geophysical methods? If so, can that aid in our understanding of the geother-mal system at Reservoir Hill and Chromo?

Lastly, an additional objective involved the careful collection and processing of the informa-

tion in order to produce the most accurate interpretations; this geophysical data was integratedwith all other relevant site statistics. The individual approaches were compiled into a joint inter-pretation of the collective results. The elucidation revealed discrete geological layers and faultsin the subsurface. The multiparty analysis was applied to develop of the final conclusion.

Contents

1 Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

1.1 Introduction 21

1.2 Background Information 21

1.3 Conclusion 31

2 Geophysical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.1 Hammer Seismic 35

2.2 Magnetics 40

2.3 Electromagnetics 42

2.4 DC Resistivity and Self Potential 47

2.5 Magnetotellurics 51

2.6 Gravity 53

2.7 Deep Seismic 56

2.8 GPS 61

2.9 Passive Seismic 65

2.10 Rock Physics 73

2.11 Well Logging 78

3 Chromo Main Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

3.1 Overarching Objectives 85

3.2 Main Line Geology 86


3.4 Magnetics 96

12



3.7 Gravity 114

3.8 Deep Seismic 121

3.9 Results and Interpretation 147

3.10 Conclusion 152

3.11 Recomendations 152

4 Reservoir Hill Student Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154


4.2 Student Site Geology 154



4.5 Gravity 179



4.8 Conclusion 204

4.9 Recommendations 204

5 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

5.1 Report on Investigation of Local Wells in Pagosa Springs 211

List of Figures

1.1 Stratigraphic column estimates the regional units in the Pagosa Springs area. Thick-nesses are rough estimates because the units vary so much over the region. Approx-imations were gathered from well log data gathered south of Pagosa Springs. Thecolumn also shows the Mesaverde Sandstones and Lewis Shales, two formationsthat are not present in the heart of Pagosa Springs [1]. . . . . . . . . . . . . . . . . . . . 22

1.2 Map of the inland seaways which existed in North America from the Mid to UpperCretaceous period. The area of interest, Pagosa Springs, is highlighted on the map [6]. 24

1.3 The cross-section of the San Juan Basin with the geologic time-scale of events [7].The map view of the cross-section can be seen in the bottom-left corner of the figure.Pagosa Springs lies approximately on the northeast corner of the cross section. . . . . . 26

1.4 Cross section of the Archuleta Anticlinorium with respect to the San Juan Basinand the San Juan Sag. This cross section comes from the surface geology map asseen in Figure 1.5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

1.5 Structural map around Pagosa Springs. The figure was obtained from Andrei Revil’spaper on Pagosa Springs geology [1]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

1.6 Generalized model of major features along Salado-Cumbres structural discontinuityincluding the Archuleta Anticlinorium, San Juan Basin, and the San Juan Sag [10]. . . 29

1.7 Surface geology map of around Pagosa Springs area. The northwest side of the mapis dominated by the Mancos Shale (Km, Kmu, Kml) at the surface. Meanwhile, theMesaverde Formation (Kmv) and the Lewis Shale (Kl) is more prevalent in the west,south, and north side of the map. The southwest side of the map around ArchuletaMesa has more Dakota Sandstone (Kd) with some intrusive rocks represented in red(Tki) [1]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

1.8 Interpretive geothermal heat flow around Pagosa Springs. There is increased heatflow directly under Pagosa Springs with a region of lower heat flow to the northeast [11].31

1.9 Tributaries and River System of San Juan Basin [11]. . . . . . . . . . . . . . . . . . . . . 321.10Approximate well locations as provided by Ford. Most of these wells are near to the

town of Pagosa Springs. The main line is south of the region displayed in this image,and is not viewable at this scale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1.11Map of wells relevant to the 2017 main line. . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2.1 Depiction of different wavefronts from a seismic source.[17] . . . . . . . . . . . . . . . . 37

14 LIST OF FIGURES

2.2 Figure depicting different ray paths and their behavior according to Fermat’s Prin-ciple.[18] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

2.3 Seismic shot gather with labeled wave moveout for reflected, refracted and directwaves.[19] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.4 Example of velocity analysis using hammer seismic data. The green line representsthe velocity above the interface, or V1. The red line represents the velocity below theinterface, or V2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

2.5 Diagram of induced electromagnetic fields and eddy currents in a FDEM survey [26]. . 442.6 Graph of how current varies in the transmitter loop in a TEM survey [29]. . . . . . . . . 442.7 This diagram shows how smoke rings induced from varying electromagnetic fields

travel in the subsurface during a TEM survey [31]. . . . . . . . . . . . . . . . . . . . . . 452.8 Electrode Spacing by Array Type: Dipole-Dipole (left), Wenner (right) . . . . . . . . . . . 482.9 The six types of anomalies commonly produced during an SP survey. These anoma-

lies are characterized by physical properties present in the subsurface. [36] . . . . . . . 492.10This map was created by A. Revil et al. for the Pagosa Springs study completed in

2014. Shows the local topography and the position of Electrical Resistivity ProfilesFrom 2012-2013 Surveys Around Pagosa Springs, CO [37] . . . . . . . . . . . . . . . . . 50

2.11The CG-5 uses a mass and zero length spring to measure the varying force of gravityon the mass. From this, the acceleration is calculated [39]. . . . . . . . . . . . . . . . . . 53

2.12CG-5 Autograv Gravity Meter with leveling plate. Two different CG-5 instrumentswere used along both survey lines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

2.13Illustration of Snell’s Law. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572.14Reflection traveltime in uniform halfspace model. . . . . . . . . . . . . . . . . . . . . . . . 582.15A GTI Wireless seismic node [41]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602.16DGPS Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 632.17Overview map of all the passive seismic sites in Reservoir Hill. The black dots

represent the approximate center of each of the surveys. Site 1 (Star) had thetriangular survey setup, while Site 1 (Line)’s setup was linear from North to South,spanning about 100 m. Site 2 had a star-shaped setup with the nodes going radiallyoutward of the center node. Site 3 had a square-like setup with varying position ofthe nodes inside of it. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

2.18Receiver array for the first site. The UTM coordinates are UTM 13S WGS 84 whereSouth point is (0321930, 4127166), North point is (0321956 ,4126204), and thecenter midpoint coordinates is (0321956 ,4126204). These points were acquiredusing hand-held GPS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

2.19Receiver array for the second site. The UTM coordinates are UTM 13S WGS 84for the far-east point is (0321772,4127031), and the west point has coordinates of(0321765,4126957). These points were acquired using hand-held GPS. . . . . . . . . . . 68

2.20Receiver array for the third site. The UTM coordinate is UTM 13S WGS 84 with thecenter point as (0321869,4126353). These points were acquired using hand-heldGPS. Black dots are the geophones arranged by the specified spacings. . . . . . . . . . . 69

2.21Passive seismic data from Node 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692.22Passive seismic data from Node 1 for time 10:00am and 6:00pm at May 18th, within

the noisiest time frame of the 38 hour survey. Because this node had much loweramplitudes overall than Node 2, the scale has been adjusted to better display the trends.70

2.23Passive seismic data from Node 1 for time 3:00 a.m. and 4:00 a.m. at May 19th,within the quietest time frame of the 38 hour survey. Because this node had muchlower amplitudes overall than Node 2, the scale has been adjusted to better displaythe trends. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

2.24Passive seismic data from Node 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 712.25Passive seismic data from Node 2 for time 10:00 a.m. and 6:00 p.m. at May 18th,

which is within the noisy time frame. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 712.26Passive seismic data from Node 2 for time 3:00 a.m. and 4:00 a.m. at May 19th,

which is within the quiet time frame. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

LIST OF FIGURES 15

2.27Archimedes’ Principle of Buoyancy Method [43] . . . . . . . . . . . . . . . . . . . . . . . . 742.28Core sample which densities were measured for this method . . . . . . . . . . . . . . . . 752.29Well locations in the Pagosa Springs area provided by Pagosa Verde, LLC . . . . . . . . . 792.30Relevant well locations for the 2017 main line. Information and locations found in

COGCC database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 802.31Part of the TG-1 Sonic Log . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 822.32Part of the TG-1 Gamma Ray Log that has been geologically annotated using pro-

vided well cores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 832.33Figure taken from the Mink report showing well temperatures in the Pagosa Springs

area [44]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 842.34Figure taken from the Mink report showing temperature gradient contours using

Galloway wells [44]. TG wells were posted on the map to show correlation. . . . . . . . 84

3.1 Preliminary cross section of the 2017 main line. The information was modeled basedon a combination of surface geology and local well data. Note that the sedimentsderived from the Ancestral Rockies is speculated to be the Entrada and WanakahFormation, in addition to the regional equivalent of the Fountain Formation. Sincethere is a lack of well information on this particular layer, the name was kept as is. . . 87

3.2 Preliminary cross section of the 2014 and 2015 seismic main line, determined fromseismic data. Note that the sediments derived from the Ancestral Rockies is specu-lated to be the Entrada and Wanakah Formation, in addition to the regional equiv-alent of the Fountain Formation. Since there is a lack of well information on thisparticular layer, the name was kept as is. . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

3.3 Map view of the 2014 and 2015 main line cross section found in Figure 3.2, com-bined with the 2017 main line cross section from Figure 3.1. The kink at the centerof the line marks the boundary between the 2014 seismic line to the east and the2015 seismic line to the west. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

3.4 Photograph of hammer seismic setup. The yellow box is the Geomatrix Geode Seis-mogrpah. Takeout cables (yellow) attach to the geode, allowing up to 48 channelsto record at once. The trigger sensor (black cable) connects the sledgehammer tothe geode for synchronized shot and acquisition . . . . . . . . . . . . . . . . . . . . . . . . 94

3.5 Example of refraction analysis. The red line depicts the shot location and is used asa reference line to find ti of the refracted wave. The black line represents the headwave or V1. The green line represents the refracted wave or V2. . . . . . . . . . . . . . . 94

3.6 Figure depicting the elevation and shale profile (top) and the calculated depth to thefirst layer, or difference between topography and the first interface (bottom), for theMain Line. The first layer is interpreted to be Lewis Shale. Note that this survey linewas taken off of the Main Line and contains no GPS data. A reference model wascreated from the closest main line station. Date of survey: May 23rd, 2017. . . . . . . . 95

3.7 Magnetics Survey Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963.8 Map of the DC Resistivity surveys conducted along the main line in Chromo, CO. . . . . 993.9 Map of the SP surveys conducted along the main line in Chromo, CO. . . . . . . . . . . 993.10Inverted DC Resistivity data from flags 1228-1360 along the main line. . . . . . . . . . 1043.11Inverted DC Resistivity data across the main dike, flags 1380-1400, along the main

line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1043.12SP surveys conducted along the main line. . . . . . . . . . . . . . . . . . . . . . . . . . . 1053.13All 8 sites are located in remote areas as to avoid cultural noise which will interfere

with the measurements. The overall site location geometry was chosen in order toget a cross section across the dike indicated by the black line. Site 8 is located on adike in order to obtain information about the dike itself. . . . . . . . . . . . . . . . . . . . 111

3.14Site 4 Apparent Resistivity Curve with a Stratigraphic Column next to it. . . . . . . . . . 1123.15Site 1 Apparent Resistivity Curve with a Stratigraphic Column next to it. . . . . . . . . . 1123.16Site 6 Apparent Resistivity Curve with a Stratigraphic Column next to it. . . . . . . . . . 1133.17Combined Gravity Survey Line. Includes the 2017, 2015, and 2014 survey lines. . . . . 114

16 LIST OF FIGURES

3.18Gravity Model based on seismic depth to basement. . . . . . . . . . . . . . . . . . . . . . 1163.19Main Line Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1183.20Main Line Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1193.21Legend for the following Gravity models for the main line. . . . . . . . . . . . . . . . . . . 1193.22Gravity Model based on matching the data. This model puts data lower than seismic,

but this is likely do to the orientation of the gravity line orthogonal to strike. Theseismic data is likely observing apparent dip and not true dip. Thus their depth tobasement is slightly shallower. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

3.23Preliminary Cross Section of the 2017 Seismic Line. . . . . . . . . . . . . . . . . . . . . . 1223.24Location of main survey line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1233.25Location of wireless nodes along main survey line. . . . . . . . . . . . . . . . . . . . . . . 1243.26Schematic illustration of acquisition geometry. . . . . . . . . . . . . . . . . . . . . . . . . 1253.27Schematic illustration of acquisition geometry. . . . . . . . . . . . . . . . . . . . . . . . . 1263.28This is the general flow we followed to process seismic data from the 2017 Field

Camp. The following steps are described in more details below. . . . . . . . . . . . . . . 1283.29An example of a four different shot records along the seismic line. The third record

down is an example of a record we decided to ”kill” due to the saturation of noisethroughout the entire record. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

3.30The layout of the main seismic line for the 2017 Field Camp. The colored linesignifies the surveying line (colors indicating fold) while the white area indicates thelocation of the CMP’s for all the records along the line. . . . . . . . . . . . . . . . . . . . . 133

3.31a) Uncorrected reflection; b) Proper Velocity; c) Velocity is too low; d) Velocity is toohigh. Yilmaz (2001). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

3.32The initial brute stack obtained using initial NMO velocities and elevation statics. . . . 1343.33The brute stacked image with refraction statics applied to it. . . . . . . . . . . . . . . . . 1353.34From left to right: (1) original shot, (2) 2000 m/s denoise, (3) 1500 m/s denoise and

(4) 1200 m/s denoise. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353.35Stack with a second iteration of NMO velocity analysis and a surface wave noise

attenuation filter applied to it. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.36Velocity Analysis of the Denoised CMP Gather. . . . . . . . . . . . . . . . . . . . . . . . . 1363.37Brute Stack Image After Applying Residual Statics Correction. . . . . . . . . . . . . . . . 1373.38(1) CMP gather of 2800, (2) Time variant spectrum whitened gather of 2800. . . . . . . . 1373.39The green curve corresponds to the energy spectrum of the original gather. The blue

curve corresponds to the energy spectrum of the whitened gather. . . . . . . . . . . . . . 1383.40Time migrated section of the main line using Kirchhoff migration. . . . . . . . . . . . . . 1383.41Depth Migrated section of the Main Line, from West to East, and the Elevation Profile

of the Line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1393.42The final product of depth migrated image (colored). . . . . . . . . . . . . . . . . . . . . . 1393.43Depth Migrated Cross-Section with the Interpreted Geology of the Subsurface along

the Main Line, from West to East. The x-direction of the cross-section is equivalentto the stations starting at Flag 1000 until Flag 2200. . . . . . . . . . . . . . . . . . . . . 140

3.44Shot at station 1560 recorded on wireless nodes. Station spacing 1.25 m. . . . . . . . . 1403.45Shot at station 1560 recorded on wired geophones. Station spacing 10 m. . . . . . . . . 1413.46Ground roll test experiment. Examples of shot records for every location indicated

on Figure Figure 3.47. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1423.47Shot locations for ground roll test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1433.48The survey line of 2014 and 2015 field camp. . . . . . . . . . . . . . . . . . . . . . . . . . 1443.49The survey line of 2017 field camp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1443.50The survey line of 2017 and 2015 field camp combined. . . . . . . . . . . . . . . . . . . . 1453.51The depth migrated seismic image of 2015 field camp. . . . . . . . . . . . . . . . . . . . . 1453.52The depth migrated seismic image of 2017 field camp. . . . . . . . . . . . . . . . . . . . . 1463.53Original geologic cross section of the main line in Chromo. . . . . . . . . . . . . . . . . . 1473.54Depth Migration of well velocities for deep seismic . . . . . . . . . . . . . . . . . . . . . . 148

LIST OF FIGURES 17

3.55Depth Migration of seismic with well velocities included. Velocities increasing fromsmall to large moving down the section. Note the section of interest near the base-ment layer in the bottom left. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

3.56Depth Migration of seismic using well velocities with geologic interpretation . . . . . . . 1503.57MT resistivity inversions overlaying geologic velocity section. . . . . . . . . . . . . . . . . 1503.58Final gravity model based on seismic, geology, and MT. This cross section includes

gravity conclusions from 2014 and 2015, at approximately 4.2 km onward. . . . . . . . 1513.59MT resistivity inversions overlaying geologic velocity section with gravity. . . . . . . . . . 1523.60MT resistivity inversions overlaying geologic velocity section with gravity. . . . . . . . . . 153

4.1 Map view of Student Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1554.2 Stratigraphic column from the P1 well drilled in Pagosa Springs [1]. . . . . . . . . . . . . 1574.3 Field Camp 2016 interpreted seismic line [11]. . . . . . . . . . . . . . . . . . . . . . . . . 1584.4 Map view of cross section line that is marked in red. . . . . . . . . . . . . . . . . . . . . . 1584.5 First preliminary cross section of Reservoir Hill using well, seismic, and geologic

determinants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1594.6 Second preliminary cross section of Reservoir Hill using well, seismic, and geologic

determinants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1604.7 Map depicting lines SO and S1 on Reservoir Hill. Black lines depict location of ham-

mer seismic surveys. The red line represents S0 while the turquoise line representsS1. The color bar on the right depicts elevation. . . . . . . . . . . . . . . . . . . . . . . . 162

4.8 Figure depicting the elevation and shale profile (top) and the calculated depth tothe first layer, or difference between topography and the first interface (bottom), forline S0 on Reservoir Hill. The first layer is interpreted to be Mancos Shale. The flagnumbers are oriented East-West with increasing flag number. Date of survey: May20th, 2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

4.9 Figure depicting the elevation and shale profile (top) and the calculated depth tothe first layer, or difference between topography and the first interface (bottom), forline S1 on Reservoir Hill. The first layer is interpreted to be Mancos Shale. The flagnumbers are oriented East-West with increasing flag number. Date of survey: May22nd, 2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

4.10Figure depicting the elevation and shale profile (top) and the calculated depth tothe first layer, or difference between topography and the first interface (bottom), forline S0 on Reservoir Hill. The first layer is interpreted to be Mancos Shale. The flagnumbers are oriented East-West with increasing flag number. Date of survey: May24th, 2017. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

4.11Map of the DC Resistivity surveys conducted along Reservoir Hill, S0, S1, and S2. . . . 1704.12Map of the SP surveys conducted along Reservoir Hill, S0 and S1. . . . . . . . . . . . . . 1714.13Inverted data for the DC Resistivity surveys conducted along S0, S1, and S2, re-

spectively, across Reservoir Hill. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1754.14The 3D inversion of the DC Resistivity data collected along student site lines S0, S1,

and S2 on Reservoir Hill. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1764.15Self potential data for the student site, line S0. . . . . . . . . . . . . . . . . . . . . . . . . 1774.16Self potential data for the student site, line S1. . . . . . . . . . . . . . . . . . . . . . . . . 1774.17SP data overlain DC data in order to see how the trends of the two methods correlate

along the student lines on Reservoir Hill. From top to bottom: SP for S0, DC for S0,SP for S1, and DC for S1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

4.18Student Site Gravity Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1794.19Student Site Quality Control - Line S0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1824.21Legend for the following Gravity models for the student site. . . . . . . . . . . . . . . . . 1824.20Student Site Processing - Line S0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1834.22Preliminary model of Reservoir Hill based on well data and geologic cross subsection.

Notice that the calculated model data does not trend well with the left side of the data. 183

18 LIST OF FIGURES

4.23Secondary model of Reservoir Hill which includes the fault speculated by DC meth-ods. Notice that the calculated model data now trends well with the left side of thedata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

4.24Map of EM survey locations around the student site of Reservoir Hill. Red crossesrepresent locations were TDEM surveys were completed and the colored lines repre-sent the survey lines in the area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

4.25TDEM setup from 5-19-2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1874.26TDEM setup from 5-23-2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1884.27TDEM setup from 5-21-2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1894.28Diagram of EM-31 Operation [47] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1904.29General picture describing the set up on a TDEM survey . . . . . . . . . . . . . . . . . . 1914.30The graph reveals different representative curves for the apparent conductivity data

obtained from EM-34 along S-1 line. The different curves in the graph representdifferent levels of model uncertainties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

4.31The L-curve that describes the relationship between model residuals and data resid-uals. The x − axis represents the difference between the predicted model m andthe reference model m. Alternatively, the y − axis represents the difference betweenthe predicted data Gm and the real data d. In addition, the variables WM and WD

represent the model and data uncertainties respectively. . . . . . . . . . . . . . . . . . . 1934.32EM-34 conductivities in the Easting direction along line S − 0 at the student site on

Reservoir Hill. Red circles represent exact conductivity values, the black curve repre-sents a specific smoothing curve, and the red curve represents a general smoothingcurve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

4.33EM-34 conductivities in the Easting direction along line S − 1 at the student site onReservoir Hill. Red circles represent exact conductivity values, the black curve repre-sents a specific smoothing curve, and the red curve represents a general smoothingcurve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

4.34EM-31 conductivities in the Easting direction along line S − 0 at the student site onReservoir Hill. Red circles represent exact conductivity values, the black curve repre-sents a specific smoothing curve, and the red curve represents a general smoothingcurve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

4.35EM-31 conductivities in the Easting direction along line S1 at the student site onReservoir Hill. Red circles represent exact conductivity values, the black curve repre-sents a specific smoothing curve, and the red curve represents a general smoothingcurve. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

4.36TEM site 1, west of Reservoir Hill, May 19th, 2017. Left: Voltage decay throughreceiver loop over time on a log-log scale. Blue data points represent high fre-quency, purple points represent low frequency. Right: Inverted earth resistivitymodel. Shows resistivity as a function of depth on a log-log scale . . . . . . . . . . . . . 199

4.37TEM site 2, on Reservoir Hill, May 23rd, 2017. Left: Voltage decay through receiverloop over time on a log-log scale. Blue data points represent high frequency, purplepoints represent low frequency. Right: Inverted earth resistivity model. Showsresistivity as a function of depth on a log-log scale . . . . . . . . . . . . . . . . . . . . . . 199

4.38TEM site 3, east of Reservoir Hill. May 21st, 2017. Left: Voltage decay throughreceiver loop over time on a log-log scale. Blue data points represent high fre-quency, purple points represent low frequency. Right: Inverted earth resistivitymodel. Shows resistivity as a function of depth on a log-log scale . . . . . . . . . . . . . 200

4.39Inverted resistivity models showing interpreted geologic formations at depth. Thesemodels are plotted on a log-log scale. Right: Site 1, May 19th, 2017. Center: Site 2,May 23st, 2017, Left: Site 3, May 21st, 2017 . . . . . . . . . . . . . . . . . . . . . . . . . 200

4.40Model showing relative location and depth of interpreted geologic layers taken at thevarious survey sites with the geologic cross section across Reservoir Hill underneath . . 201

LIST OF FIGURES 19

4.41Figure 4.41c and Figure 4.41b show the survey parameters of the historic DC resis-tivity surveys collected on Reservoir Hill. Figure 4.41a shows the survey parametersfor each of the methods conducted the 2017 field camp. . . . . . . . . . . . . . . . . . . . 205

4.42A conventional inversion of DC Resistivity Line PAGO-02. Data for this line wascollected at the 2012 field camp, and was later processed and published in Dr.Revil’s paper. Note the relatively conductive anomaly at distance x=1000m. Thisanomaly appears in almost the exact same section as the anomaly found in the S1DC profile. DC line S1 can be found as a part of Figure 4.44 . . . . . . . . . . . . . . . . 206

4.43Plots comparing the SP, EM31, EM34, and DC resistivity results on line S0. Thered dashed lines show correlations between the SP and EM methods. These corre-lations line up nicely with the conductive and resistive bodies in the DC Resistivityinversion. Specifically, the spike in the data seen in SP, EM-34, and EM-31 line upwith the conductive body on the west side of Reservoir Hill. This is followed by adrop in the data which lines up with the beginning of the resistive body a little bitfarther to the east. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

4.44Plots comparing the SP, EM31, EM34, and DC resistivity results on line S1. Thered dashed lines show correlations between the SP and EM methods. These corre-lations line up nicely with the conductive and resistive bodies in the DC Resistivityinversion. Specifically, there is a clear drop in data in the SP, EM-34, and EM-31where the resistive body begins, a spike where this body ends and transitions intoa more conductive zone, and another drop where the second resistive body begins. . . . 208

4.45A plot showing the DC inversion from line S1 overlayed with the vertical profile takenat TDEM site two. The sounding was taken in the middle of the conductive anomalyand fits the resulting DC inversion perfectly, further confirming the credibility ofthe DC results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

4.46A comparison between the gravity model and line S0. The model aligns a fault tothe anomaly of the DC survey. Positioning the fault here decreases the associatederror with the gravity data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

4.47A comparison of the DC data to the finalized geologic cross section. The area of theDC survey is outlined with a blue rectangle. . . . . . . . . . . . . . . . . . . . . . . . . . . 210

List of Tables

1.1 Formation depths obtained from local well data. The location of the wells can beseen on the map in Figure 1.11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.1 Measured mass of the core samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762.2 Calculated bulk density and volume of the core samples . . . . . . . . . . . . . . . . . . 77

3.1 Description of columns in exported ASCII files from ipi2win.MT . . . . . . . . . . . . . . 1093.2 Main Line Gravity Corrections: X indicates a completed correction. . . . . . . . . . . . 1163.3 Density Values of Geologic Layers: These values were used when modeling the

gravity response of the main line and S0. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173.4 2D Seismic Line Survey Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

4.1 Student Site Gravity Corrections: X indicates a completed correction. . . . . . . . . . 1804.2 Density Values of Geologic Layers: These values were used when modeling the

gravity response of the main line and S0. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1814.3 Site 1 EM-47 parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1894.4 Site 1 EM-57 parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1904.5 Site 2 EM-57 parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1904.6 Site 3 EM-57 parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

1. Geology

Contents

1.1 Introduction 21


1.3 Conclusion 31

1.1 IntroductionThe town of Pagosa Springs is located in southwest Colorado. Its complex geologic history has

shaped the hydrological and hydrothermal systems of the region, and understanding its historyis key to answering the geologic questions posed at the outset of this geophysical investigation ofPagosa Springs. Such questions include:

Main Line Questions:1. Do the dikes on the main line control water flow and quality?2. Is there any potential relation of the Chromo area to the northward Pagosa Springs geother-

mal system?Reservoir Hill Questions:

1. Is there a fault on Reservoir Hill?2. Is Reservoir Hill related to the Mother Spring?3. How does water flow to the Mother Spring?

1.2 Background InformationPagosa Springs rests on the western slope of the continental divide on the northeast edge of

the San Juan Basin. Chromo, the other site of investigation, is located South of Pagosa Springsjust north of the New Mexico border and lies within a complicated compressional system of foldsand faults. It is suspected that the faults and dikes of the region could be the reason that hotsprings exist by creating a pathway for springs to emerge. In order to properly characterize thiscomplex hydro-geothermal system, it is necessary to understand the regional geology in whichthe Pagosa Springs system resides.

22 Chapter 1. Geology

1.2.1 Regional Geology1.2.1.0.1 History and Timeline

A brief synopsis of deposition events that created the geology of southwestern Colorado start-ing in the Precambrian has been included and listed in chronological order. These formations areall examples of the geologic units that are expected to be present in the Pagosa Springs region.For the formation thicknesses, please see Table 1.1 for details. The stratigraphic column belowin Figure 1.1 depicts the same sequence of strata that was presented and described in the historysection above.

Figure 1.1: Stratigraphic column estimates the regional units in the Pagosa Springs area. Thick-nesses are rough estimates because the units vary so much over the region. Approximationswere gathered from well log data gathered south of Pagosa Springs. The column also showsthe Mesaverde Sandstones and Lewis Shales, two formations that are not present in the heartof Pagosa Springs [1].

1.2.1.0.2 Precambrian BasementThe basement rock present beneath Colorado dates from the middle Precambrian Era. The

rock is a mixture of metamorphic and igneous materials aged 980-1540 million years old. ThePrecambrian is the primary geothermal reservoir. There was a regional uplift of the Precambriancrystalline basement which resulted in the Ancestral Rockies approximately 300 million yearsago [2].


1.2.1.0.3 Cutler GroupThe debris from the erosion of the uplift created the Culter Group, which is observed in the

basins surrounding the former mountain range. The uplift and subsequent erosion also causedthe contact unconformity between the basement and the Jurassic-aged formation above it. Theformation is deposited during the Early Permian and contains sandstone, medium-grained mixedclastic with a hint of limestone.

1.2.1.0.4 Entrada SandstoneOverlying the basement, there is a unconformity between the Precambrian and Jurassic pe-

riods which denotes a period of erosion. During the Jurassic period, an eolian sandstone layerwas deposited on top of this unconformity. These sandstones were deposited in a near shoreenvironment approximately 160 million years ago. These sandstones, known as the EntradaSandstones, are composed of medium to fine grained quartz. In the Pagosa Springs area theformation is estimated to be about 160 feet (49 m) thick.

1.2.1.0.5 Wanakah FormationThe Wanakah Formation was deposited on top of the Entrada Formation in the middle of the

Jurassic period. The Wanakah Formation is not continuous throughout throughout the basinand perhaps through the study area. The Wanakah consists mainly of eolian sandstones withsome shale and limestone layers. The presence of limestone and crystalline anhydrite inter-spersed within the Wanakah Formation indicates that parts formation formed in a lacustrineenvironment [3]. The sandstone itself is soft and friable, with some clay layers closer to the top[4]. The formation is approximately 100 feet (30 m) thick in the Pagosa Springs Area.

1.2.1.0.6 Morrison FormationThe Jurassic aged Morrison Formation is deposited over the Wanakah Formation. The thick-

ness of the bed is approximated to be around 680 to 733 feet (207 to 223 m) around PagosaSprings. The formation was deposited during the Jurassic time period, marking the transgres-sion of the Western Interior Seaway. The Morrison formation is divided into 3 major sectionsbased on sedimentation. The base layer of the Morrison is fluvial fan sandstone. The mid-dle section consists of light-colored and thick sandstone ledges intermixed with siltstone andshale layers. The upper-most section is mostly gray, green, and red siltstone, containing minoramounts of limestone and conglomerate sandstone [4]. This sandstone was deposited during thetransgression of the Western Interior Seaway.

1.2.1.0.7 Dakota GroupThere was a 50 million year unconformity following the Morrison Formation. This 50 million

years unconformity is overlain by the Dakota Group, marking the beginning of the Late Creta-ceous Period. The formation also coincides with the coming of the Western Interior Seaway, aninland sea separating the Western and the Eastern side of North America. It is approximately110 to 270 feet (33 to 82 m) thick around the Pagosa Springs area. The Dakota Group is oneof main aquifers in the Pagosa Springs area with the groundwater confined and results fromfracture porosity [1]. A map of the seaway based on current outline of North America as seenin Figure 1.2. The Dakota indicates a transgression event following the creation of the Morri-son, and a regression event occurring during the formation of the upper layer of the DakotaSandstones. The formation largely contains sandstone, with conglomerate and shale inclusions[5]. This formation is relatively close to the surface and outcrops in several locations in PagosaSprings.

1.2.1.0.8 Mancos ShaleThe Mancos Shale is located on top of the Dakota Sandstone. Its presence indicates marine

environment in the region during the Upper Cretaceous. It is approximately 100 million yearsold and averages 400 feet (122 m) thick in the Pagosa region. The formation is composed of fissiledark carbonaceous shales with the presence of fossil bearings. The Mancos Shale is considered


Figure 1.2: Map of the inland seaways which existed in North America from the Mid to UpperCretaceous period. The area of interest, Pagosa Springs, is highlighted on the map [6].

to be one of the two aquifers in the Pagosa Springs area with the ground water flow as a resultof fractures and inter-granular porosity [1].

1.2.1.0.9 Greenhorn Limestone

There are several limestone members of the Mancos Shale including the Greenhorn Lime-stone, Carlile Shale, and Graneros Formation. Th Greenhorn Limestone is noteworthy as it isvisible throughout Pagosa Springs. A thin layer of Greenhorn limestone exists within the MancosShale layer, located about 30 meters above the top of the Dakota Sandstone and was depositedduring the Cretaceous period. The presence of fossils and limestones give evidence of marinelifeforms indicating a shallow marine environment.

1.2.1.0.10 Mesaverde Sandstones

The Mesaverde sandstone is enclosed by the Mancos shale. This sandstone layer is approx-imately 300 feet (91 m) thick and consists of brittle sandstones and gray shales. This layerpinches out due to depositional regions of transgression and then regression of the shoreline.The presence of the Mesaverde sandstones indicates a fluvial depositional environment [7].


1.2.1.0.11 Lewis ShaleThe Lewis Shale is deposited on top of the Mesaverde Sandstones during the Cretaceious

period and is approximately 1000 feet (300 m) thick. The Lewis Shale is similar to the MancosShale, the main difference being varying abundance of calcium carbonate. In regions where theMesaverde has been pinched out, the Lewis Shale is not considered to be a seperate formationfrom the Mancos. After the Lewis Shale deposited, the Larmide Orogeny period occurred. To-wards the end of the Cretaceous period, the Colorado region underwent uplift. This uplift, calledthe Laramide Orogeny, formed the current Rocky Mountains. The uplift occurred due to com-pressional forces brought on by a continental collision west of Colorado between the Kula andFarallon Plates sliding under the North American Plate forming the Archuleta Anticlinorium andother faults and folds in the region. The Laramide Orogeny occurred 80 to 55 million years agoand ended approximately 35 to 55 million years ago [8].

1.2.1.0.12 Fruitland and Kirtland FormationThe Fruitland and Kirtland Formations were deposited after the Laramide Orogeny approxi-

mately 75 million years ago. However, these formations are not apparent in Pagosa Springs dueto erosional forces [9]. The stratigraphic column for the main line ends here.

1.2.1.0.13 San Juan VolcanicsThe most recent geologic deposition event are the San Juan Volcanics. These volcanic moun-

tains were formed in the Oligocene and Miocene period. The presence of andesitic dikes [1], aswell as the thick volcanic tuff present near Wolf Creek Pass, shows a combination of intrusiveand extrusive volcanic events that occurred across the region. These volcanic remnants formthe points of highest relief for the region, predominantly making up the mountain ranges thatsurround the San Juan basin. Erosional debris from the San Juan Volcanics contributes toalluvium and fluvial deposits at the survey sites.

1.2.2 Structural Geology1.2.2.0.1 Basins

The majority of the San Juan basin is located in New Mexico, with Pagosa Springs locatedjust north of the Colorado-New Mexico border along the northeastern rim edge of the San JuanBasin. The area of the basin is approximately 12,000 km2. The basin is asymmetric with a maindrainage from the San Juan River [9]. The basin is at the foot of the San Juan Sag, separated bythe Archuleta Anticlinorium. The general cross section of the San Juan Basin-Sag system canbe seen in Figure 1.4.

1.2.2.0.2 AnticlinesThe Archuleta anticlinorium was formed from to the same compressional forces that resulted

in the Laramide Orogeny. The anticlinorium is located between the San Juan basin and theSan Juan Sag, as seen in Figure 1.3. The anticlinorium itself is nearly 120 km long. Thesmaller anticline examined in this report is the Chromo Anticline. That anticline lies within theArchuleta Anticlinorium system. The Chromo anticline behaves asymmetrically, with a northwestorientation. This orientation indicates there is a possibility that there is an underlying northwesttrend of the basement faults, which could have caused the northwest trend of the folds as seenin the Figure 1.4.

Additionally, Pagosa Springs is also located on the outer edge of the Stinking Springs Anti-cline. This is noted from the sedimentary layers flattened back out and becoming consistent withsurrounding geology. This anticline is believed to have been formed during the late Cretaceous,during same time frame as the Laramide Orogeny. It is characterized by dips approximately 5-10degrees on either side of the fold axis.


Figure 1.3: The cross-section of the San Juan Basin with the geologic time-scale of events [7]. Themap view of the cross-section can be seen in the bottom-left corner of the figure. Pagosa Springslies approximately on the northeast corner of the cross section.

1.2.2.0.3 FaultsThe anticline system underwent a series of extensional and compressional events, creating

a combination of normal and reverse faults throughout the Pagosa Springs region. The upliftevents, most notably the Laramide Orogeny, caused deformation and stress to the regional ge-ology. The presence of the reverse faults likely derives from this event, while the normal faultsare caused by a variety of other factors such as tensional stress where the rocks are pulled apartfrom each other. Like the folds, the faults trend northwest to southeast.

1.2.2.0.4 Volcanic DikesSeveral volcanic dikes are concentrated around the main line survey area in the south of

Pagosa Springs. The dikes are visible at the surface, forming long linear ridges that trend north-east to southwest [1]. The majority of these dikes were formed around the same time as volcanicsthat make up the San Juan mountains. The volcanic dikes provide a historic account of the re-gional stresses. The largest stress, σ1 is a result of gravitational forces, and with σ3 as the forcethat holds the rock units together laterally, that means that our secondary stress σ2 is in thesame direction as strike of the dikes. The orientation of the dikes, and by extension the stresses,provides a snapshot of the stress patterns of the region at the time the dikes were formed.

1.2.3 Surface GeologyThe surface geology of Pagosa Springs area varies widely, with several geological units that

make up the visible outcrops in both the Pagosa Springs and Chromo area. The northern edgeof Pagosa Springs is primarily Mancos Shale. To the south of Pagosa Springs and around theChromo area, there are outcrops of Lewis Shale and Mesaverde, along with dike outcrops vis-ible at the road cuts. This is seen in the surface geological map, available in Figure 1.7. TheMesaverde is present primarily south of Pagosa Springs, closer to Chromo. There is a gradualtransition between Mesaverde and Mancos approaching Pagosa Springs, indicated by the color


Figure 1.4: Cross section of the Archuleta Anticlinorium with respect to the San Juan Basin and theSan Juan Sag. This cross section comes from the surface geology map as seen in Figure 1.5.

change from gray Mesaverde to brown Mancos Shale. The Lewis Shale is seen at the surfacefurther south. Water from the San Juan river and snow melt reserves drive the rapid erosion ofthe Mancos and Lewis Shale, leaving behind the more resistive sandstone formations.

1.2.4 Geothermal SystemA hot spring is be produced by either localized volcanics, or by the Earth’s geothermal gradi-

ent in combination with a local aquifer within the system to allow the accumulation of geothermalwater. The history of the Pagosa Springs area is characterized by geologically recent volcanic ac-tivity, meaning a potential source of heat for the springs could be from an intrusive batholith.However, there is no current geologic information to support there being a batholith acting asa heat source as it may be too old and previous studies should have detected such a feature.Nonetheless, these volcanic events could have significantly altered the thermal gradient through-out southwestern Colorado [1]. The characterization of the current geothermal heat flow inPagosa Springs is a high amount of heat flow directly under the town, with decreasing heat flowto the north. The age of the thermal system is undetermined due to the large scale erosion of the


Figure 1.5: Structural map around Pagosa Springs. The figure was obtained from Andrei Revil’spaper on Pagosa Springs geology [1].

rocks in the area.The Dakota formation has produced the warmest water temperature values according to

well measurements. However, other wells drilled further from Pagosa Springs do not exhibit asimilar temperature spike in the Dakota. This implies that the Dakota is not the geothermalreservoir but rather a local aquifer. The Dakota makes a good aquifer because the comparativelyimpermeable Mancos shale overlies the Dakota and creates a cap for the water. In this case, therecharge source for the aquifer is known. In the Chromo area, it comes from snow melt west ofthe continental divide which drains into the Navajo river and into the groundwater. From this, itis possible to conclude that the geothermal system of Pagosa Springs is potentially disconnectedfrom that of Chromo.

1.2.5 HydrologyPagosa Springs is located on the eastern side of the San Juan Basin near the edge of the

watershed cut-off. The San Juan Mountain topography determines the regional fluid flow in the


Figure 1.6: Generalized model of major features along Salado-Cumbres structural discontinuityincluding the Archuleta Anticlinorium, San Juan Basin, and the San Juan Sag [10].

Pagosa Springs area. The surface water tends to flow northeast to southwest, following the riversystem of the San Juan Basin outlined below in Figure 1.9. The San Juan River runs throughColorado, New Mexico, Arizona and Utah. In its initial stages, it flows near Pagosa Springs eastto West fed by tributary rivers flowing in from the northeast towards the southwest [1].

The Mancos Shale and Dakota Sandstone are the two main aquifers in the Pagosa Springsarea. The Mancos Shale groundwater flow is a result of fractures and inter-granular porosity [1],allowing the interconnection of isolated porous units. The Dakota Sandstone groundwater flowis well confined and results from fracture porosity. The groundwater quality in the aquifers ishighly variable, exhibiting particular high concentrations of sodium, calcium, iron, sulfate, anddissolved hydrogen sulphide [7]. In addition to these elements, water from the hot springs alsocarries calcium carbonate which precipitates out at surface temperature and pressure conditions.The precipitates result in travertine deposits [12]. This may indicate a high degree of inter-connectivity between rock units.

The fault system in the area contains several known steeply dipping faults. It is possiblethat fluid flows through the Dakota sandstone and is fed vertically by upwelling through frac-tured rock and intersections of faults [1]. This assumption is supported by the presence of highgroundwater flow present in the Dakota Sandstone, potentially acting as a channel for water flowfrom the basement. Several other geologic factors such as lithology, fractures, stress field, dia-genesis, rock mechanics, fluid chemistry and geochemistry also impact fluid flow, permeability,porosity, and overall temperature distribution of the geothermal reservoir.

There are two types of geothermal system heat regimes: convection-dominated and conduction-dominated. Convection-dominated geothermal regimes are known as active geothermal systems,with high enthalpy resources of short-term fluid flow and dynamics occurring at plate tectonicmargins, active tectonics, or volcanic settings. In contrast, conduction-dominated geothermalsystems, speculated to be like the one in Pagosa Springs, have low to medium enthalpy re-sources and are located at passive tectonic plate settings with no significant recent tectonics orvolcanism. Conduction-dominated systems have geothermal activity is located in a low perme-ability layer such as tight sandstones, carbonates or crystalline rocks located at a greater depth[13].

Deformation and cementation within the fault zone in the system induces compartmentaliza-tion, creating low permeability zones and barriers to fluid flow. In another situation, fractures


Figure 1.7: Surface geology map of around Pagosa Springs area. The northwest side of the mapis dominated by the Mancos Shale (Km, Kmu, Kml) at the surface. Meanwhile, the MesaverdeFormation (Kmv) and the Lewis Shale (Kl) is more prevalent in the west, south, and north side ofthe map. The southwest side of the map around Archuleta Mesa has more Dakota Sandstone (Kd)with some intrusive rocks represented in red (Tki) [1].

in the fault zone acts as conduits for fluid flow [14]. Dikes are nearly vertical sheets of low per-meability rock intruding through existing rocks and extending vertically and laterally for a longdistances, impeding ground water flow in the regions around them. Dikes intersect and compart-mentalize the permeable rock where ground water is present, lowering the overall rock porosityand permeability. They also channel groundwater flow parallel to the general trend of the dikes.[15].

1.2.6 Local ReportsPagosa Verde is a renewable energy and sustainable agriculture company based in Pagosa

Springs, CO. The company was founded by Sally High and Jerome “Jerry” Smith. Smith intro-duced the Field Camp Outreach Crew to JR Ford, who works in a building near to the PagosaVerde offices.

According to Ford, an oil and gas company drilled several wells South of Pagosa Springs areawithout any regulations in 1951 and as a result, none of the wells were capped. Ford also men-tioned that locals north of downtown Pagosa Springs have had continued success drilling wellswith higher temperature and discharge. But those drilling south of town have experience lowersuccess, finding only lower temperature and lower discharge wells. The outreach crew hypothe-sized that this may be due to interference from San Juan river or the fault system through thearea.

Around the turn of the century, the town and retail spas started pumping more water for

1.3 Conclusion 31

Figure 1.8: Interpretive geothermal heat flow around Pagosa Springs. There is increased heat flowdirectly under Pagosa Springs with a region of lower heat flow to the northeast [11].

use. At the same time usage in town increased, local wells located directly south of the town,visualized in Figure 1.10, began to dry up starting from the northern wells and progressing to themost southern well. With one exception, all of these wells are currently dry and still uncapped.The exception, Well 6 in Figure 1.10, is not dry but is only producing at a rate of 2-3 gallons perminute and the temperature has cooled from 140 to 90-95 degrees Fahrenheit.

Another piece of information provided by Ford is that formerly capped wells near the SanJuan river are dumping geothermal waters into the river. One capped well, possibly locatedunder a drug store in Downtown Pagosa Springs, was close to flooding so an open pipe wasconstructed to funnel geothermal water from under the building and into the river. In 2013,Geophysics Field Camp came to Ford’s farm but no useful conclusion was made on the behaviorof the wells.

1.2.7 Well InformationMuch of the geological information used in this report was obtained from local well informa-

tion documented on the Colorado Oil and Gas Conservation Commission website. There were 4wells located close to the 2017 main line, providing information on the depths of the various for-mations. The most relevant wells used, from east to west, are Garcia 1 (05-007-06001), Richawk4 (05-007-06004), Griego 1-A (05-007-05010), and Valdez 1 (05-007-06113). There are otherwells near the main line that do not provide sufficient geological information, but have beenincluded for completeness. These wells are James Griego 1 (05-007-05009), and Talamante 1(05-007-05010).

1.3 ConclusionPagosa Springs has been developed through million years of deposition and erosions. Geo-

logical features like the faults, dikes, sedimentary layers, and anticlines are important factorscontrolling the groundwater flow and the creation of hot springs. The Pagosa Springs area is aconduction-dominated geothermal regime located at a passive tectonic plate setting with no sig-


Figure 1.9: Tributaries and River System of San Juan Basin [11].

nificant recent tectonics or volcanism. Faults, fractures, and dikes could be important conduitsfor direct flow of hot water. A greater understanding on the geology using geophysical meth-ods along the student site and main line will give a greater understanding on the geology andgroundwater flow of the geothermal system in Pagosa Springs.

1.3 Conclusion 33

Figure 1.10: Approximate well locations as provided by Ford. Most of these wells are near to thetown of Pagosa Springs. The main line is south of the region displayed in this image, and is notviewable at this scale.

Table 1.1: Formation depths obtained from local well data. The location of the wells can be seenon the map in Figure 1.11.

Garcia 1 well Richawk 4 wellFormation Log Top(ft) Formation Log Top (ft)Lewis 0 Lewis 0Mesaverde 1160 Mesaverde 1248Mancos 1457 Mancos 1533Elevation: 7409 ft Dakota 3303

Elevation: 7680 ft

Griego 1A well Valdez 1 wellFormation Log Top(ft) Formation Log Top(ft)Carlile 1937 Mesaverde 0Greenhorn 2349 Mancos 1253Dakota 2520 Greenhorn 2100Morrison 2712 Graneros 2132Elevation: 7275 ft Elevation: 7420 ft


Figure 1.11: Map of wells relevant to the 2017 main line.

2. Geophysical Methods

Contents


2.2 Magnetics 40




2.6 Gravity 53

2.7 Deep Seismic 56

2.8 GPS 61




2.1 Hammer Seismic2.1.1 Introduction

Hammer seismic is an active geophysical method using a hammer and a metal plate as theseismic source. The seismic waves generated from the contact between the hammer and the platepropagate radially into the subsurface. These waves are reflected, transmitted, refracted, anddiffracted depending on the material properties of different subsurface features and the angle ofincidence of the wave. These material properties include density, bulk/shear modulus, elasticity,and fluid content. Each of these properties affect the velocity of the wave, which can be computedusing the measured travel times of the first arrival waves and station spacing of the survey line.

36 Chapter 2. Geophysical Methods

Hammer seismic is generally a near surface method. The depth of investigation depends on thedifferent material properties, but a good rule of thumb is approximately 1/5th the length of thereceiver line. Ultimately, the maximum depth of observation is 20 meters.

Hammer seismic is preferable in remote areas that are not easily accessible by vehicles andprovides insight to near surface investigations. Hammer seismic is also a relatively simplemethod, making it an ideal preliminary method for subsurface investigation. Using this un-derstanding, students implemented hammer seismic surveys on two main sites. The first sitewas located on the main survey line along County Road 542 while the second site was located onReservoir Hill, East of the Mother Spring in Pagosa Springs. Both survey sites were selected withthe goals of identifying geologic structures (faults, dikes, etc.) as well as emphasizing understand-ing of the near surface geology. The main objective of the first site located along CR542 aimsat evaluating the dike outcropping. This dike is thought to have great significance in the sub-surface geothermal fluid flow. Reservoir Hill contained two different survey lines with the mainobjective of investigating the geologic structure to determine if previous years’ proposed faultindeed exists. The potential fault’s location has been interpolated by students using a numberof resources including previous resistivity, well log, and seismic data, gathered local information,as well as surficial geology. However, the identification of a fault and its exact location has yetto be completed, assuming it exists. The surveyed sites were chosen after preliminary geologicaldiscussions with Dr. Bob Reynolds who shared his insight of the area’s geology.

2.1.2 TheorySeismic waves that propagate in the subsurface are subject to a number of rules concern-

ing wave propagation. Waves travel radially through a uniform medium and are dependent ona number of physical properties including porosity, pore pressure, fluid content, and effectivepressure. Two different types of seismic waves exist: P-waves and S-waves. P-waves are com-pressional waves that have the same polarization and propagation direction, and they propagatefaster through the subsurface than S-waves. S-waves are shear waves that have perpendicu-lar polarizations to the propagation direction in the subsurface. The equation for velocities ofP-waves and S-waves are listed below in Equation 2.1 and Equation 2.2, respectively.

α =

√λ+ 2µ

ρ(2.1)

β =

√µ

ρ(2.2)

Here, µ and λ represent shear modulus and the bulk modulus (dynamic viscosity), respec-tively. These variables are referred to as Lames parameters. These parameters dictate the behav-ior of the different seismic waves. Because Earth is generally heterogeneous, different subsurfaceinterfaces contain different physical properties. For this reason, seismic waves will behave dif-ferently depending on their interaction with these different interfaces, causing the waves to bereflected, refracted, or diffracted. When a seismic wave encounters a sharp and continuous in-terface, a portion of the wave is reflected while the other portion is transmitted. The amountthat the seismic wave is either reflected or transmitted depends on the velocity contrast acrossthe interface, as well as the incidence angle of the seismic wave with the interface. When theincidence angle reaches the critical angle, the transmitted wave will travel along the interface(See Figure 2.1)[16].

Fermat’s principle states that waves propagate between two points along the path that re-quires the least amount of time as compared to other nearby paths. These extreme paths areinfluenced by a number of physical properties and explains whether that waves path will bereflected or refracted. This is otherwise known as the law of reflection and Snell’s Law whichexplains the behavior of the wave due to the angle of incidence with the subsurface interface.See Figure 2.2 below for a visualization.


Figure 2.1: Depiction of different wavefronts from a seismic source.[17]

Figure 2.2: Figure depicting different ray paths and their behavior according to Fermat’s Princi-ple.[18]

Huygens principle explains how a refracted wave traveling along an interface will generatewavefronts. These wavefronts propagate radially outward and arrive at the surface with thevelocity of the refracted wave traveling along the subsurface interface. These waves are referredto as head waves. Because the refracted waves must ultimately reach the surface before thedirect wave, the deeper seismic velocity must be greater than the first layer’s velocity. Due totopsoil’s inherently slow seismic velocity, this is typically not a problem with hammer seismic.The final interaction between a seismic wave and an interface are diffractions. Diffractions takeplace along a sharp and non-continuous interface, such as the corner of an interface. Thediffraction propagates radially outward, acting essentially as a secondary impulse source alongthe interface.

Reflected, refracted, diffracted and direct waves must be identified from the seismic datain order to give insight to the different subsurface velocities present. These different velocitiescorrelate to different interfaces. In order to differentiate between reflected and refracted wavesinvolves understanding how each wave behaves in the subsurface. Reflections will arrive first,but with slower velocity. Refractions will arrive later and at further offsets, but with a fastervelocity in comparison to the reflected waves. This occurs because the refraction can only occurat the critical angle as well as having a faster velocity below the interface. Lastly, one can identifythe differences between direct, refracted, and reflected waves based on the moveout of the seismicdata. Reflected waves have a hyperbolic moveout while direct and refracted waves have a linearmoveout. See Figure 2.3 for a visualization of these different moveout behaviors.

With this understanding of how different seismic waves interact with the subsurface, differentanalyses can be used to identify different velocities and compute thicknesses. These differentanalyses include reflection, refraction, and diffraction analysis. Reflection and refraction analysisestablish velocity and thickness to create a velocity model of the subsurface. This velocity modelcan be used to create an Earth model representing the changes in thickness and dip along


Figure 2.3: Seismic shot gather with labeled wave moveout for reflected, refracted and directwaves.[19]

different survey lines. Diffraction analysis is used to evaluate the constructed velocity model byanalyzing diffraction focusing.

In order to calculate the velocity of a specific interface, the following Equation 2.3 can be usedby implementing travel time (z-direction) and station spacing (x-direction).

v =∆x

∆z(2.3)

Using this equation, the inverse slope of the line is equivalent to velocity. See Figure 2.4 belowfor a visualization of implementing this equation on seismic data.

After calculating different velocities, thicknesses can be computed using Equation 2.4 below.

h =ti2

V1V2√V 22 − V 2

1

(2.4)

Here, V1 and V2 are velocities yielded from the seismic data. V1 is the slope of the directwave and V2 is the slope of the head wave and represents the velocity below the interface. ti isthe theoretical time at which the refracted wave would reach the surface at time 0. Typicallya heterogeneous subsurface exists, causing a number of potential earth models to fit the data.However, 2 major cases typically exist. Case 1 is a 2D layered Earth with horizontal interfaces.Case 2 is a 2D layered Earth with dipping interfaces. For Case 2, velocities traveling up-dip willbe faster than that of the velocities traveling down dip. In both cases, V1 is the slope of the directwave and V2 is the slope of the head wave and represents the velocity below the interface.

When considering seismic data acquisition, noise and error of the data collected should beconsidered. In order to reduce noise, one must take into account the signal to noise ratio (SNR).This is done by recording multiple shots at each trace and having data collected for the entirespread of geophones. Ideally, consistent geologic features will appear in each shot while noisewill be inconsistent between shots. This data can be averaged together to remove noise. Taking


Figure 2.4: Example of velocity analysis using hammer seismic data. The green line represents thevelocity above the interface, or V1. The red line represents the velocity below the interface, or V2.

shots at multiple locations allows for a better SNR and will identify geologic features from eachangle. This allows for the geometry of the geologic features to be identified from multiple shotswhilst eliminating inconsistent noise. This process is referred to as stacking and emphasizes thegeology while minimizing noise.

2.1.3 BackgroundStudents in previous years have used hammer seismic in the effort to image the subsurface

with the aim of identifying different geologic features. Last years students conducted reflectionand refraction analysis using Madagascar and Software Construction (Scons) to import, process,and visualize the acquired seismic data. Students were able to compare a forward model cre-ated by students in 2015 with the processed data from 2016. Students were able to identify adike in the data collected from the previous years student site as well as create an earth modelrepresenting the changes in subsurface depth to the first layer. The previous hammer seismicdata can be compared to the data collected on this year’s student site (lines S0 and S1). Thiscomparison allows for better interpretation of the subsurface geology and the depth to the firstlayer. By knowing the depth of the topsoil and alluvial deposits, hammer seismic can contributeto constraining other methods–most notably direct current resistivity. Looking forward, Reser-voir Hill could be a potential student site for the 2018 Field Camp if 2017’s data proves to beinteresting. See Figure 4.7 for a visualization of the location of the survey lines and hammerseismic lines on Reservoir hill.

2.1.3.1 Equipment• Geophones• Sledgehammer with trigger sensor• Metal plate• Car battery• Takeout cables• Geomatrix Geode Seismograph (Figure 3.4)• Laptop (record/view data)• Tape measure• Nails to secure ends of tape measure


2.2 Magnetics2.2.1 Introduction

Magnetic methods are used to determine variations of the magnetic susceptibility of rocksin the subsurface. This passive method relies on the magnetic field of the earth as a primarysource. This magnetic field, hypothesized to be driven by a dynamo effect in the core, can berepresented as a simple dipole model. The Earth’s magnetic field causes magnetic minerals toalign with this field which can cause a rock to become magnetized. This magnetization thenleads to a secondary magnetic field that can be measured.

The primary and secondary fields are both measured using a cesium vapor magnetometer.Afterwards, the primary field component is removed to provide a value for the induced secondaryfield. This secondary field yields information about the magnetic nature of the subsurface geologyhosting that field. Metallic objects and volcanic rocks have a high magnetic susceptibility andtherefore, are optimal targets for a magnetic survey. In the case of the 2017 field camp survey,the target is the dike along the Chromo main line at flag 1398.

2.2.2 TheoryMagnetics is a passive method used to measure a total magnetic field, which includes the

reference field (the earth’s magnetic field), and the induced magnetic field of magnetically sus-ceptible material in the subsurface. The two fields are superimposed, and thus we must monitorthe primary (inducing) field and remove its effect from the data to yield data strictly from thesecondary field. Subtracting the reference field from the total field leads to an understanding ofanomalous features in the subsurface. This secondary (anomalous) field that is obtained can beused to undertand the geologic structures of the subsurface. The reference field hits the surfaceof the earth as a plane wave at angles defined by inclination and declination. The inclination isthe angle between the magnetic field and its projection in a horizontal plane at the Earth’s sur-face, and declination is the angle made by the plane wave and geographic north. The referencefield then interacts with a magnetically susceptible object, creating a magnetic dipole out of themagnetically susceptible material so that it behaves similarly to a bar magnet.

Under the assumption that the materials under investigation have a small susceptibility (k <0.1), we can define the magnetization of a material to be the product of the susceptibility and theinducing field:

~M =k ~B0

µ0(2.5)

where ~B0 is the primary inducing field and µ0 is the permeability of free space with an exact valueof 4π × 10−7Hm−1.

Magnetic susceptibility, k, is a dimensionless proportionality constant that describes howeasily a material can me magnetized by an inducing magnetic field. The susceptibility of amaterial is dependent on its mineral content, saturation, and porosity. The anomalous secondaryfield produced by the magnetically susceptible material can be described as a superposition ofthat magnetization throughout the medium:

~Ba =µ0

4π

∫∫∫~M · ∇∇ 1

|~r − ~r′|dv (2.6)

~Ba =µ0

4π

∫∫∫T ~Mdv (2.7)

Equation 2.7 describes the anomalous magnetic field produced by a given magnetically sus-ceptible object under an inducing field. If we measure the ~Ba and the strength and direction ofthe inducing field, ~B0, then we can calculate the magnetic susceptibility, k, of the material inquestion through equations Equation 2.5 and Equation 2.7.

2.2 Magnetics 41

2.2.3 BackgroundThe goal in conducting a magnetics survey along the 2017 main line was to aid in under-

standing the exact location and nature of the dikes. This is because the volcanic rock of a dikeis expected to have a higher magnetic susceptibility. Furthermore, a magnetics survey can po-tentially aid in constraining the previously attained gravity survey adding statistical relevance toour interpreted geological models.

2.2.3.1 Equipment List• Cesium Vapor Magnetometer• Connector Cables• Base Station• DGPS System• Batteries


2.3 Electromagnetics2.3.1 Introduction

Electromagnetic (EM) geophysical methods take advantage of electric and magnetic fields thatare induced as a result of moving charges. These methods are useful for determining variabilityin the conductivity of the subsurface. There are two main branches of EM methods: time andfrequency domain. Both adhere to similar physical laws to induce a secondary electromagneticfield underground. The physical property these methods are interested in is the conductivity ofthe rock units underground. Water and other fluids are substances that might change the con-ductivity of the subsurface; thus, EM methods are useful for locating and mapping undergroundfluid flow, a topic of interest in the Pagosa Springs area [20].

Time domain EM (TEM), sometimes referred to as transient electromagnetics (TEM) is an EMmethod that generates a vertical conductivity sounding which provides information about theconductivity of the subsurface at various depths. This method is useful for seeing how conduc-tivity changes as a function of depth. The Geonics EM-47 and EM-57 are two instruments thatare commonly used to conduct TEM surveys. These instruments measure how the induced elec-tromagnetic field decays with time which is useful for determining the subsurface conductivityof an area because the field will decay at different rates in layers of differing conductivity.

Frequency domain EM (FDEM) involves inducing a secondary current in the subsurface byalternating the primary electromagnetic field in order to gain information about the ground con-ductivity. Depending on location specific survey parameters, FDEM can be useful in both verticaland lateral ground conductivity soundings. Two common instruments used to conduct FDEMsurveys are the Geonics EM-31 and EM-34. The EM-34 can acquire a vertical sounding of theearth using variable spacing and frequency settings. This can be useful in exploring faultedbedrock for groundwater applications. The EM-31 is more useful in acquiring lateral conductiv-ity soundings at depths near the surface. Both instruments are effective in that they can collectlarge quantities of data in relatively short periods of time. FDEM instruments also do not needto be in contact with the surface. This means they can be useful in numerous other applicationssuch as unexploded ordinance detection, and in locations that may be inaccessible by foot.

2.3.2 TheoryGeophysical electromagnetic methods measure an induced electromagnetic field underground

to gain a better understanding of subsurface conductivity. Both passive and active source EMmethods are used by geoscientists to better understand environmental, geotechnical, and explo-ration problems. Passive source EM methods deal with listening to the Earth’s natural variationsin its electromagnetic field caused by worldwide electrical storms and solar wind. In active sourceEM methods, we control the source by driving a time-varying current through a transmitter loop.The current in the transmitter loop induces a magnetic field perpendicular to current flow in thesubsurface due to Ampere’s Law, seen in Equation Equation 2.8 below [21].

Ampere’s Law (with Maxwell Correction):

−→∇ ×

−→B = µ0

(−→J + ε0

∂−→E

∂t

)(2.8)

This magnetic field thus induces an electric field because of Faraday’s Law, seen in EquationEquation 2.9 below.

Faraday’s Law:−→∇ ×

−→E = −∂B

∂t(2.9)


The change in flux of these secondary magnetic fields is measured through the receiver loop.The EM field passing through the receiver loop induces a current in the receiver loop which ismeasured as EMF. EMF is the electromotive force and is the voltage developed by a source ofelectrical energy (in this case, the secondary EM field induced in the subsurface) [22]. This canbe represented by Equation Equation 2.10 below.

− δ

δt

∫∫~B · ~δa = −δΦ

δt= EMF (2.10)

Frequency Domain EM Theory

In FDEM the current transmitted into the transmitter loop is a time varying sinusoidal currentthat follows Equation Equation 2.11 below.

I(t) = I0e−iωt (2.11)

The operating theory behind most of frequency domain geophysical surveys is explained byFaraday’s law which declares that any change in magnetic flux in a system induces a countereffect current to sustain the stability of the modified system, a phenomenon known as electro-magnetic induction. A primary magnetic field with an alternating angular frequency disturbsthe initial static state of the system and further induces an electric field. This primary electro-magnetic field induces eddy currents in conductive bodies underground according to Ohm’s law,seen in Equation Equation 2.12.

Ohm’s Law:

−→J = σ

−→E (2.12)

The eddy currents induce their own secondary electromagnetic fields which pass through thereceiver loop that measures the EMF produced by the electromagnetic fields. The magnitude ofthese eddy currents and the subsequent electromagnetic fields produced by them are dependenton the conductivity and permittivity of the body. This provides useful information regarding thesubsurface geology. The secondary EM field can be divided into two orthogonal components, theIn-phase (real component) and the Quadrature (imaginary component) [23]. This information isuseful in separating the secondary EM field from the primary field because the secondary field isreceived 90 degrees out of phase from the primary field. This helps to obtain the information thatis useful for analyzing and interpreting (the secondary EM field) [24]. The Quadrature componentis proportional to ground conductivity and therefore is useful in analyzing groundwater and fluidflow in the subsurface. The In-phase component is useful for finding metal objects underground[25]. A diagram of how FDEM generally works can be seen in Figure Figure 2.5 below.


Figure 2.5: Diagram of induced electromagnetic fields and eddy currents in a FDEM survey [26].

Geonics EM-31 is a ground conductivity meter commonly used in shallow FDEM surveys.It is useful for performing relatively quick geophysical EM surveys and measures the apparentconductivity in millisiemens per meter (mS/m). The effective depth of investigation of this instru-ment is about 6 meters [27].

Geonics EM-34 is a second conductivity meter with greater intercoil spacing and larger trans-mitter and receiver loop sizes. The depth of investigation depends on survey specific parameterssuch as intercoil spacing, however the EM-34 instrument is capable of seeing down to a depth of60 meters [28].

Time Domain EMThe principles of TEM are similar to those involved in FDEM in that both methods follow the

same physical laws (Ampere’s, Faraday’s, Ohm’s, etc...). However, the current source used forTEM is very different. For TEM, the current in the transmitter loop is a constant current that isshut off at a certain time. A visual of this current source can be seen in Figure Figure 2.6.

Figure 2.6: Graph of how current varies in the transmitter loop in a TEM survey [29].

When the current is shut off, the Earth resists this change by inducing a secondary current(and subsequent EM fields) directly below the current that was previously running through the


transmitter loop. This current is in the opposite direction of the change in magnetic flux causedby shutting off the transmitter current. This phenomenon is described by Lenz’s Law [30]. Loopsof current continue to be induced deeper in the subsurface, an effect colloquially known as”smoke rings”. The smoke ring effect can be seen in Figure Figure 2.7. These smoke rings travelfaster through more resistive bodies but decay faster in more conductive bodies. How the EMFmeasured in the receiver loop decays with time helps provide information about how the smokerings travel through the subsurface. This is useful in determining the physical properties suchas conductivity of the underground lithology.

Figure 2.7: This diagram shows how smoke rings induced from varying electromagnetic fieldstravel in the subsurface during a TEM survey [31].

Geonics EM-47 and EM-57 are two instruments that are used for TEM vertical conductivitysoundings. The EM-47 the transmitter loop can be up to 100m by 100m with a maximumtransmitter output of 3 Amps. The maximum depth of investigation for the EM-47 is 150 meters[32]. The EM-57 operates on lower frequencies and is capable of transmitting higher currents(up to about 24 Amps). This means that this instrument has a greater depth of investigation,capable of seeing up to 300 meters below the surface [33].

2.3.3 BackgroundBoth frequency and time domain EM methods were used in Pagosa Springs. Due to time and

space constraints only FDEM methods were deployed along the main line. Both methods wereable to be deployed at the student site at Reservoir Hill and in the surrounding area. These meth-ods proved extremely useful because they were able to acquire large quantities of data over thearea of interest in a relatively short period of time. The speed of acquisition with these methodsnot only aided in electromagnetic interpretation, but also helped to provide supplemental infor-mation useful for other methods designing surveys in the area and to evaluate the legitimacy ofother data throughout processing.

Regarding the main line, the primary objective of every method was to image the dike thatintersected the survey line on the Eastward side of the line, or about four kilometers from thestart of the line. It is suspected that this intrusion could be an important factor in identifyingthe pattern of groundwater flow in the area. Due to the logistics of surveying in this area andmany of the surrounding property owners withholding permits to conduct surveys on their land,


only FDEM was deployed along the main line. The lateral conductivity sounding provided by theEM-31 and EM-34 proved more practical for acquiring useful data while staying near CR502. Wewere hoping to be able to see a difference in ground conductivity on the East and West sides ofthe dike. This could be important information for understanding the varying properties of wateron opposing sides of the dike.

At the student site on Reservoir Hill, a much larger lateral area was made available, thus bothfrequency and time domain EM were used at this location. FDEM EM-31 and EM-34 instrumentswere run along lines S0 and S1 to gain a lateral conductivity sounding across Reservoir Hill.TEM was used only once on the hill itself, and was deployed in an area where it would intersectwith line S1 for the purpose of integrating the findings of this TEM survey with those of theFDEM survey as well as other methods used on this line. TEM was also deployed in two sitessurrounding Reservoir Hill, one site to the East and one site to the West of the hill. Thesesites were chosen in order to highlight any sudden local changes in the layering pattern ofsubsurface material caused by events such as faulting, which could provide insight into theorigin of Reservoir Hill as well as its connection to the underground hot springs in the area.

EquipmentBelow is the equipment that was used in Pagosa Springs to complete the FDEM and TEM

surveys.

Geonics EM-31• Transmitter Loop (End Tube)• Receiver Loop (End Tube)• Recorder Console (Middle Piece)• Datalogger 600• Shoulder Strap

Geonics EM-34• Transmitter(TX) Loop• Receiver(RX) Loop• Reference Cable (10m, 20m, 40m)• Transmitter(TX) Console• Receiver(RX) Console

Geonics EM-47• ProTEM Digital Receiver and battery• EM-47 Transmitter• Reference Cable

• Transmitter(TX) loop cable (400m)• EM-47 3D1 Receiver(RX) coil (x, y, and z

components)• Pre-amp for signal boost• Nonmetallic stakes and hammer (to hold

transmitter loop in place)• Handheld GPS

Geonics EM-57• ProTEM Digital Receiver and battery• EM-57 Transmitter• Reference Cable• Transmitter(TX) loop cable (400m)• Gasoline generator• EM-57 1D receiver coil• Nonmetallic stakes and hammer (to hold

transmitter loop in place)• Handheld GPS


2.4 DC Resistivity and Self Potential2.4.1 Introduction

DC resistivity is an active electrical method that measures the voltage between two electrodesas a direct current is injected into the ground from another pair of electrodes. The voltagesrecorded are calculated into apparent resistivity values, which are then transcribed into resistiv-ity values that can be mapped with respect to depth via data inversion. The ability to invert DCdata and interpret it with respect to depth is a notable advantage of the method. DC can be usedto find water, help delineate lithology, and map other conductive bodies (e.g. metallic ore bodies).

Self-potential (SP) is another electrical method, and is commonly used in conjunction withDC resistivity. Unlike DC, SP is a passive method that measures relative voltage to ultimatelydetermine the flow of telluric current through the subsurface. This is done by simply using non-polarizable electrodes to measure voltages at different points. Telluric currents flow naturally onthe surface and subsurface of the earth, and for our purposes, have been interpreted to monitorsubsurface fluid flow. Self potential surveying also has applications in mineral exploration andtemperature gradient monitoring.

SP and DC resistivity methods are often coupled together because they both fall under thejurisdiction of Ohm’s law. The control variable in DC, current, is an experimental variable inSP, making the methods complementary. Though DC surveys do not explicitly measure telluriccurrents, inverted DC data can provide a more in-depth interpretation of current flow outlinedby SP.

2.4.2 Theory2.4.2.1 Direct Current (DC) Resistivity

Electrical current flows between two terminals of a battery due to chemical reactions withinthe battery. Electron flow is used for the DC Resistivity survey method in geophysical exploration.In particular, the measurement describes the ease/difficulty of electrons to flow between twopoints. This ease/difficulty is a parameter of a material known as the conductivity (σ), andresistivity (ρ). The conductivity and resistivity of a material relate to eachother through:

ρ =1

σ. (2.13)

From the previous discussion, it is evident know that current and conductivity relate toeachother. Their relationship follows Ohm’s Law:

~J = σ ~E. (2.14)

Equation 2.14 introduces the terms ~J and ~E, which are current density and electric field re-spectively. The current density describes the flow of electrons in three dimensions. The electricalfield term describes the spatial distribution of forces on virtual charges in a three dimensionalspace.

Equation 2.14 needs to be taken one step further before evaluating the data by relatingthe electrical field to a measurable quantity, voltage. Voltage describes the three dimensionaldistribution of electric potentials. Now it is possible to relate potentials and forces throughderivatives in the following way:

~J = −σ∆V. (2.15)

Equation 2.15 describes the relationship between the current density, the measured voltage,and the desired conductivity/resistivity value. However, the term ~J is still the current densityand is not the input current (I). The current and current density relate to each other through ge-ometry. This gives rise to the geometric factor (k), which is dependent upon the survey geometry.The geometric factor is such that


R = kρ. (2.16)

Equation 2.16 provides a relationship between the material property (resistivity, ρ) and thephysical property (resistance, R). From this, it is possible to transition into the alternate form ofOhm’s Law:

V = IR, (2.17)

where the voltage (V) is measured and the input current (I) is controlled through the equip-ment. The next step is to convert back to the material conductivity and resistivity via Equa-tion 2.13 and Equation 2.16.

The only additional requirement to this computation is a known geometric factor (k). As weknow the geometric factors for Wenner and Dipole-Dipole arrays, two kinds of DC survey setups,the geometric factors can be calculated using the following:

kWenner =1

2πa

kDipole−Dipole =1

πan(n+ 1)(n+ 2),

(2.18)

where a and n are spacing parameters. In particular, a is the spacing of the voltage mea-suring electrodes and n is any natural number determined by the separation of the closestvoltage-measuring and current-injecting electrodes in the Dipole-Dipole array. These are de-picted visually in Figure 2.8.

Figure 2.8: Electrode Spacing by Array Type: Dipole-Dipole (left), Wenner (right)

2.4.2.2 Self Potential (SP)Self potential surveys aim to passively measure potentials flowing through the earth’s sub-

surface. Such potentials occur due to telluric current flow. Telluric currents are natural electriccurrents that flow through the earth and arise when displaced charges move in an attempt toattain a state of equilibrium. Telluric currents create natural potentials when fluid flows throughthe subsurface, due to subsurface medium disruption, and near areas with high concentrationsof electrolytic solutions [34]. Self potential surveys use such anomalies to map fluid flow acrossdikes, faults, and other geologic/physical boundaries.

The self potential method is valuable because it is the only geophysical tool capable of directlymeasuring fluid flow [35]. Fluid flow will present itself in the form of streaming potential. Stream-ing potentials occur when a fluid passes through a medium with different electrical propertiesthan that of the fluid. Such potentials flow in the same direction of the source fluid, with inflowproducing negative potentials, and outflow producing positive potentials [34]. Figure 2.9 showsthe characteristic anomalies of a Self Potential survey

SP surveys are taken by placing two non-polarizing electrodes in the ground and using avoltmeter to measure the potential difference across the electrode pathway. Relative changes in


Figure 2.9: The six types of anomalies commonly produced during an SP survey. These anomaliesare characterized by physical properties present in the subsurface. [36]

voltage may be indicative of streaming potential and thus subsurface fluid flow. Such changesoften present themselves on the scale of millivolts (mV), and are plotted relative to position thenobserved qualitatively. When conducting an SP survey, it is essential that good ground contact ismaintained, and that non-polarizable electrodes, like copper sulfate electrodes, are used. Suchpractices ensure that the small electric fields produced by telluric currents can be adequatelymeasured.

2.4.3 BackgroundColorado School of Mines (CSM) has explored Pagosa Springs and surrounding regions for its

geothermal exploration potential extensively since 2012. This year’s field camp had two differentsurvey sites: a main line just west of Chromo, CO, and a student site on reservoir hill. Thereservoir hill site was of particular interest to DC acquisition groups as several DC and ElectricalResistivity Tomography (ERT) were historically conducted by CSM students and affiliates. Fig-ure 2.10 below illustrates the previous electrical resistivity profiles completed by Colorado Schoolof Mines faculty. This year’s survey lines, specifically line S1, aim to serve as a continuation ofthe previous work done by extending the subsurface imagery of Reservoir Hill.

This area was of interest to this year’s field camp as well as previous CSM groups due to it’sproximity to the Big Springs. Known as the deepest hot springs in the world, the Big Springs isthe foremost source of geothermal energy for the town of Pagosa Springs [38]. Despite its titleof ”deepest hot springs in the world”, information on the springs’ plumbing and source reservoiris relatively unknown [37]. Findings from the surveys conducted in May 2012, October 2012,and May 2013 were ultimately published by Andre Revil et al in 2015. The paper suggestedthe potential presence of the Victore Fault, which was previously undocumented in the area.Furthermore, it suggests that the Victore Fault intersects with another fault, entitled ”Fault”,suggesting that the Big Spring originates from a fault controlled reservoir [37]. This conclusionwas ultimately drawn from the localization of low resistivity anomalies surrounding the springs.

The DC resistivity and SP surveys conducted on Reservoir Hill area this year aim to furtherexplore the findings published in 2015. DC and SP surveys were also conducted near ArchuletaMesa, just west of Chromo, CO, and approximately 28 miles south of downtown Pagosa Springs.


Figure 2.10: This map was created by A. Revil et al. for the Pagosa Springs study completed in2014. Shows the local topography and the position of Electrical Resistivity Profiles From 2012-2013Surveys Around Pagosa Springs, CO [37]

This region hosts many geologic dikes, which could possibly to interfere with subsurface waterflow. Thus, the surveys conducted further south aim to confirm this phenomena by observingresistivity and current flow anomalies.

2.4.3.1 DC Equipment• ABEM Cables (20 meter takeouts)• ABEM Switch Box• ABEM Terrameter• Cable Connectors• Connection Clips (64)• Electrodes (64)• Hammer• Saltwater• 12 Volt Battery

2.4.3.2 Equipment• Voltmeter• Non-polarizable copper sulfate electrodes (2)• Spool of wire• Rock Hammer• Watch


2.5 Magnetotellurics2.5.1 Introduction

Magnetotellurics is a geophysical method based on time variation-measurements of both themagnetic field B(t) and the induced electric field E(t).Using the measurements of the Earths time-varying, naturally occurring magnetic and electric fields, magnetotellurics (MT) survey teamscan derive resistivity (or conductivity) estimations of the subsurface. This geophysical methodis passive, i.e. the magnetotelluric source originates from solar-winds and lightning strikes.Charged, solar storm particles interacting with the Earths magnetic field account for the solarwind, while the ionosphere acts as a waveguide for lightning strikes. The periods of these passivesources go from 0.0001Hz up. As the magnetic fields from these sources pass over the ground,some of the energy is directed into the subsurface generating electric fields. The instrumentsused in the field then measure the corresponding horizontal E(t) and B(t) as well as the verticalB(t). We are able to take our surficial measurements of electromagnetic currents and use theratio of its components to estimate impedance, Z, as a function of frequency obtained through aFourier Transform. The Z is calculated using the ratio of E(w) and B(w), thus combining boththe electrical and magnetic components.

2.5.2 TheoryTo examine the physics that explicates the magnetotelluric method, the assumption must be

made that charges can accumulate along discontinuities in multi-layered Earth which behavesas an Ohmic conductor and obeying the equation,

~J = σ ~E (2.19)

where J is the electrical current density, σ is the conductivity and E is the electric fieldstrength.

Additionally, the assumption must be made that electromagnetic waves, produced by theaforementioned passive methods, vertically propagate into the earth in the form of plane waves.These planar waves are governed by Maxwell’s equations:

Faraday’s Law

∇× ~E = −∂~B

∂t(2.20)

Ampere’s Law

∇× ~B = µ0~J + µ0ε0

∂ ~E

∂t(2.21)

Gauss’ Law for magnetism∇ · ~B = ρv (2.22)

Gauss’ Law∇ · ~E =

ρ

ε0(2.23)

Based on these assumptions, we are then able to utilize Maxwell’s correction to Ampere’s law,denoting that a changing magnetic field creates an electric field and vice versa.

∇× ~B = µ0

(~J + ε0

∂ ~E

∂t

)(2.24)

where σ, ε and µ represent the intrinsic, material properties subject to electromagnetic wave-field propagation . ε is the electric permittivity and µ is the magnetic permeability.

The ratio of these fields, in the frequency domain, demonstrated through the above equationsis used to determine the apparent resistivity of the subsurface. The relationship is as follows:

Z(w) =µEi(0)

Bj(0)(2.25)


where i, j = {x, y}.Assuming E(w) and B(w) are consistent with the following equations

E(w) = Ei(0)exp(−√iwµσz) (2.26)

andB(w) = Bj(0)exp(−

√iwµσz) (2.27)

and knowing Faraday’s Law, one can arrive at the equation

Z(w) =

√iwµ

σ=µEi(0)

Bj(0)(2.28)

Using this relationship, the resistivity can be obtained by manipulating the equation into

ρa(w) =|Z(w)2|wµ

. (2.29)

The phase, angle between the imaginary and real parts of the impedance, can be found using

φ = tan−1(Im(Z)

Re(Z)

). (2.30)

All of the above equations are in terms of frequency, w, meaning any apparent resistivitycurve and phase will also be in terms of w. This is consistent with the idea of skin depth givenby

δ =

√2

ωµσ. (2.31)

As frequency increases, the depth-of-investigation (DOI) decreases. Similarly, as the con-ductivity increases, the DOI decreases. Using the apparent resistivity, phase, and skin depthequations the data obtained in the field can be transformed into apparent resistivity curves andeventually inverted into depth images.

2.5.3 BackgroundPrior Pagosa Springs field session MT methods (2014, 2015, and 2016) all shared the common

goal of assisting in the understanding of the towns unexplained geothermal activity. Dependentupon the sampling frequencies and the time span of a survey, the MT method can successfullyinvestigate depths from 0.30km to 10km. Due to its significant DOI assessment, the method wasagain chosen for the 2017 Geophysical field camp to ascertain, confirm and support the notionof the source-location of the hot springs and the existence of other significant, sub-surficialgeological features. Allowing the data to be tabulated over a significant time period, interpretationof the deep seismic line could be investigated at a depths to the Precambrian basement, beyondthe capabilities of ancillary methods.

2.5.3.1 EQUIPMENT LIST• MetronixTMADU-07(e) Magnetotelluric System• Part Number: 258,286• Field laptop• 6 induction coil magnetometers• 4 Cu/CuSO4 non-polarizing electrodes• 2 stainless steel- electrodes (for grounding)• 8 Self Potential Cables (to connect electrodes to ADU)• 4 50 meter Magnetometer Cables (to connect magnetometers to the ADU)• 2 12 volt Car batteries with appropriate battery cables• 2 mobile GPS units (that attach to the ADU, for time drift corrections)• 100 meter field measuring tape

2.6 Gravity 53

2.6 Gravity2.6.1 Introduction

The field of geophysics utilizes relative gravity measurements to find changes in density (∆ρ)throughout the subsurface. Changes in material density around instrumentation produce achange in the value of gravitational acceleration measured. During the 2017 field survey, a CG-5Autograv Gravity Meter was used to measure the vertical component of gravitational acceleration.

This gravimeter is very sensitive to gravity measurements on the order of microgals, whichmakes it susceptible to noise in the form of surrounding terrain, but allows for deep investigationinto the subsurface.

The changes in density observed by gravity can be paired with geological investigation andother geophysical methods to determine areas of interest such as changes in fluid or mineralcontent. This can be very useful for reservoir monitoring or mineral exploration. In the case ofthe 2017 field survey, the area of interest was fluid content differences west and east of the maindike on the Chromo main line and any sharp density contrasts that might occur on ReservoirHill.

2.6.2 TheoryThe CG-5 operates by monitoring the movement of a known mass attached to a spring. Den-

sity changes affect the force acting on the object according to Newton’s Law of Universal Gravita-tion:

~F = γm1m2

r2r (2.32)

where ~F is the force acting on mass m1 (which in this case is the mass in the gravimeter), γ isthe gravitational constant 6.67 × 10−11Nm2/kg2, m2 is the mass of the area of interest, r is thedistance between both masses, and r is the unit directional vector.

Figure 2.11: The CG-5 uses a mass and zero length spring to measure the varying force of gravityon the mass. From this, the acceleration is calculated [39].

For a given mass, m1, in the gravimeter, it is possible to measure the acceleration due togravity, ~g in m

s2 , through measuring the force pulling down on that mass vertically:

~F = m1~a (2.33)

~g ≡~F

m1= γ

m2

r2r (2.34)


The source of the changes in the gravity field observed and measured as the vertical compo-nent of ~g are the result of distributions of mass beneath the Earth’s surface. Mass, however, isan extrinsic property of a material that is a function of geometry and thus harder to remotelydetect. Because of this, density is more valuable in building an image of subsurface structures.Density is dependent on the mass and volume of an object, but independent of geometry:

ρ =∆m

∆V(2.35)

where m is a mass in Kg and V is volume in m3 such that ρ is Kg/m3. Different geologic units havedifferent densities and thus produce contrasts that can be detected by the gravimeter. Regions ofhigher density have more mass per unit area and thus produce a stronger gravitational responseat the surface. It is also crucial to note that the gravitational field decreases as a factor of distancesquared such that as the instrument moves further away from a dense unit, the measuredresponse of that anomaly will decrease greatly.

It is possible to take gravity measurements and build a constraining model to invert thedata to calculate different geological settings that are representative of the measured response.However, inversion was not used in the 2017 field camp. Instead, models were built that wereconstrained by an existing geological understanding, well data for depths to units with knowndensities, and seismic depth migrations. This constructed model is used for the interpretation ofthe geological structure of the survey line.

The gravity data itself is influenced by both tidal drift and instrument drift. Tidal drift is theresult of the varying positions of the moon and sun relative to the CG-5’s location on the Earthover time. Similarly, instrument drift is the result of variations within the instrumentation overtime, such as spring fatigue, electrical fluctuations, and internal temperature fluctuations.

2.6.3 Background

There have been no previous gravity surveys performed on the 2017 main line or the studentsite on Reservoir Hill. However, gravity surveys have been performed during previous geophysicsfield camps that were located near the 2017 main line and Reservoir Hill. These surveys providedinsight into the basic geology of both areas and will be used in tandem with our results to builda more complete geological understanding of the region beyond our survey locations alone.

The goal in implementing a gravity survey in the 2017 field camp was to improve upon thisbasic geologic understanding. The objective was to further clarify the location of a presumedfault beneath Reservoir Hill and to learn more about the structure of the dikes along the mainline in Chromo and how they might affect water flow in the region.

Gravity surveys can be sensitive to faulting that creates offset in deep basement layers. Thesedeep basement layers tend to be denser than overlying sediments, and when the depth of thatlayer suddenly changes, the gravimeters will be sensitive to the gradual physical response of thatinterface. Gravity can also help identify the fluid content of underlying geologic units, as thedensity of the fluids that fills its pores determines much of the density of those units. If fluidsare present in porous units on one side of an impermeable barrier like a vertical dike, and notpresent on the other side, it might be possible to detect this change.

2.6.3.1 Equipment List• CG-5 Autograv Gravity Meter• CG-5 leveling stand• CG-5 carrying case• Umbrella (to shelter from wind)

2.6 Gravity 55

Figure 2.12: CG-5 Autograv Gravity Meter with leveling plate. Two different CG-5 instruments wereused along both survey lines.


2.7 Deep Seismic2.7.1 Introduction

Seismic exploration is a useful geophysical methods for interpreting large scale geology andcharacterizing the subsurface at various depths. Seismic surveys image different geological fea-tures, such as lithological boundaries, folds, faults, and dikes. When imaging the subsurfaceusing the seismic method, one of the key concepts to first understand is the idea of acousticimpedance. Acoustic impedance is essentially how waves reflect or transmit through differentmedia in the subsurface, caused by contrasts in the physical properties of these media. Wecan identify acoustic impedance contrasts in the subsurface to analyze the geological featuresdescribed above. Because different rocks have different properties, sudden changes in lithology(seismic interfaces) lead to changes in seismic wave propagation through the subsurface. In thepresence of lithological boundaries and geological features, acoustic impedance contrasts arise(from differences in density and velocity) and allows one to determine how much energy will bereflected or transmitted by the seismic interface. Understanding these reflection events is thebasis for any type of seismic interpretation, and allows us to analyze what type of geologicalfeatures are present in the subsurface.

To analyze subsurface characteristics, seismic surveying first requires putting energy into theground and recording the travel time of the seismic waves to different locations along the surveyline. Instruments called geophones record the travel times and stores them as data. Preliminarydata comes in the form of ”shot records” for each location along the survey line. These shotrecords then undergo a series of processing techniques to reduce noise and to correct for geome-try in the subsurface. After completing the processing, the data will ultimately display a detailedimage of the subsurface geology. The goal of the 2017 Deep Seismic survey was to analyze thegeological characteristics of the Chromo, CO area to determine if any such characteristics areaffecting fluid flow in the subsurface.

2.7.2 Theory2.7.2.1 Introduction

Seismic methods rely on an active source to generate waves that propagate through the earthand reflect at seismic boundaries. An array of receivers, on the surface, records the reflectedenergy, and after extensive processing it is possible to create a depth image that is a detailedrepresentation of the subsurface.

2.7.2.2 Seismic SourcesIn land seismic acquisition, the most commonly used sources are vibrator trucks (vibroseis)

or explosives. In general, explosives are used in areas that are not accessible for vibrator trucks.We collected the 2017 seismic dataset using a vibroseis source. The truck has a rectangular platethat is pressed to the ground and shakes according to the programmed sweep parameters. Thefrequency of shaking changes over the duration of the sweep, going from low frequencies to highfrequencies (upsweep) or from high to low frequencies (downsweep). Before further processing,the sweep has to be autocorrelated to achieve an almost zero-phase signal.

2.7.2.3 Seismic WavesThere are two main types of waves: compressional waves (P-waves) and shear waves (S-waves).

P-waves are always faster than S-waves (common Vp/Vs ratio for crystalline rocks is 1.73). As aP-wave propagates, particles move in the direction of propagation. For S-waves, particle motion isperpendicular to the direction of propagation. As a result, S-waves can be polarized horizontally(SH-wave) or vertically (SV-wave). Furthermore, S-waves cannot propagate through fluids.

Wave propagation in a medium depends on the velocity and density of the medium wherecontrasts in these parameters creates a seismic boundary. Seismic boundaries often correspondto lithology changes, but that does not always have to be the case: for example they may be

2.7 Deep Seismic 57

a result of a change in the type of fluid filling the rock. As a wave comes to a boundary, partof its energy reflects back towards the surface and part is transmitted through the subsurface.The reflection coefficient (for incidence wave that is perpendicular to the surface) for a particularwave is defined by Equation 2.36:

Rc =ρ2V2 − ρ1V1ρ2V2 + ρ1V1

=Z2 − Z1

Z2 + Z1, (2.36)

where ρ is density, V is velocity, and their product Z is acoustic impedance. In reality, reflec-tion coefficients change with incidence angles, and that can be used to detect the presence ofhydrocarbons (for example AVO, AVA analysis).

2.7.2.4 Reflections and RefractionsSnells law from geometric optics helps describe the physics that occur at a seismic interface.

The angle of the transmitted wave depends on the velocity of two media and angle of the incidencewave:

sin(θi)

Vi=

sin(θt)

Vt, (2.37)

Where θi is the incidence angle, θt is the transmission angle, Vi is velocity of the medium in whichthe wave was propagating and Vt is velocity of the medium where the wave is transmitted. Thatconcept is illustrated on Figure 2.13.

Figure 2.13: Illustration of Snell’s Law.

If Vi is smaller than Vt and the incidence wave exceeds a certain critical angle (Equation 2.38),a phenomena called total internal reflection may occur. This means that all of the incidenceenergy is reflected back. The wave that is then recorded is referred to as a refracted wave or headwave.

θc = sin−1(ViVt

)(2.38)

Refractions provide information about the depth of shallow, weathered, layers and can be usedfor static corrections (discussed in more detail later). After all the information is extracted, re-fractions are removed from the seismic record.

2.7.2.5 Seismic VelocityMost seismic data processing aims to make the best use of reflections because they indicate

seismic boundaries. In shot records, reflections show up as coherent signals along hyperbolic-looking trajectories. Let’s consider the simple case of one a homogeneous layer with constantvelocity V and thickness h (Figure 2.14). According to Fermats principle, a wave will propagatealong the fastest path. Therefore, if ”t” is the travel time of the wave, from source to receiver, t0


Figure 2.14: Reflection traveltime in uniform halfspace model.

is zero-offset two-way travel time, and x is offset, we can calculate the moveout using the simpleformula:

t =

√t20 +

x2

V 2(2.39)

Equation 2.39 is hyperbolic. The velocity used in this equation is the interval velocity of a singlehomogeneous layer (in the case of a flat reflector). Lets now consider a multi-layered model withseveral flat reflectors. Now, the normal moveout equation has the following form:

t =

√t20 +

x2

V 2RMS

, (2.40)

where t is travel time of the wave from source to receiver, t0 is zero-offset two-way traveltime andx is offset. RMS velocity used in this expression is defined as:

V 2RMS =

1

t0

N∑i=1

V 2i ∆ti (2.41)

Equation 2.40 (also known as Dix formula) is widely used for velocity analysis. If no other datais available to constrain velocities, one often has to rely on NMO information to build a velocitymodel. A good velocity model is crucial for the migration procedure (discussed later). Errors invelocities translate to errors in spatial positioning of interfaces and wrong interpretations. Forthat reason, we often repeat the velocity analysis workflow multiple times. However, one has tokeep in mind that NMO assumes a simple, isotropic 1D model, and is only correct for near offsets.The more complex the geology is, the more difficult and less reliable a NMO analysis becomes.

2.7.2.6 Migrated Depth SectionData processing aims to reduce noise, enhance signal, and account for different factors that

have an impact on the data. This, combined with a well estimated velocity field, is necessaryto obtain a quality depth image - a final product for interpretation. The migration procedurecollapses diffractions and places reflectors in their correct depth positions; however, that positionis based on the estimated velocity field and ”inherits” all of its uncertainties. An interpreter mustbe aware of that uncertainty when making important decisions, for example when deciding whereto place a well.

2.7.3 BackgroundThere have been several surveys performed by the geophysics field camp around the area of

Pagosa Springs aiming to improve the overall geological understanding of the geothermal systemas well as local structures. Some of the surveys were focused in town, near Reservoir Hill, whileothers were centered in the Chromo area; as well as near the surrounding areas of county roads

2.7 Deep Seismic 59

542 and 359. The 2015 seismic line ties directly into the 2017 line; therefore, we are able tointegrate our data with 2015’s to bring continuity to the entire set of data.

The previous year’s focus was on surveying the area around the Mother Spring and east oftown to learn more about fluid flow and possible faulting. This year’s survey is more focused oninvestigating how known geological features such as dikes and faults affect fluid flow and wellquality. The school has obtained several sets of well data consisting of actual drilling logs, alongwith anecdotal evidence provided by locals in the area. The goal for the deep seismic crew is tocreate an image of the subsurface in order to better understand the local geology of the main line.After first understanding the local geology, the next goal is to model the fluid flow that mightexplain the differences observed in the well data.

This years’ line crosses two known dikes and a suspected fault. One of the goals was to imagethese structures in more detail.

Seismic methods are great tools for imaging the subsurface. As explained in the theory sec-tion, seismic waves are sensitive to changes in elastic properties of the rocks. As such, majorchanges in lithology should produce reflections (we assume that different rocks have differentdensities and velocities) whereas features like faults (sudden sharp change in parameters con-fined to a small area) act like point scatterers and result in diffractions. Furthermore, dependingon strength of the source, attenuation in the surveyed area, and noise in the data, we may beable to depict features at depths on order of several kilometers.

To further answer the question what we can see from the seismic method, one thing toconsider is the resolution. Conventionally we look at resolution of seismic data in terms oftemporal and spatial resolution, both of which depend on a wavelet. Temporal resolution refersto a minimum thickness of a layer that we can resolve. We often use the Rayleigh limit λ

2 (a half ofa dominant wavelength) to express achievable resolution in the data domain. A simple measureof horizontal resolution is Fresnel zone radius described by Equation 2.42, where v is averagevelocity, t is two way travel time and f is dominant frequency. First Fresnel zone for flat reflectoris elliptical or circular [40].

r =V

2

√t

f, (2.42)

Both temporal and spatial resolution are hard to quantify because the subsurface is much morecomplex than those simple definitions allow. It is important to remember there is a limit on thelayer thickness that we can resolve, and that horizontal features are never imaged perfectly. Forexample, reflectors may appear to be continuous when they are not.

2.7.3.1 Equipment ListA land seismic survey is conducted by generating a seismic wave on the surface of the ground

along our main survey line, in which the energy is picked up by receivers that are placed intothe ground at the desired survey area. The source and receiver types used are essential inreaching the goal of our survey, which is imaging and observing the subsurface water movementin Chromo. The equipments that were used to assemble the land seismic acquisition is as follows:

1. AHV-IV Commander (PLS-364) Vibroseis model• The vibrator truck is used as an impulsive source of energy used during the survey.

The vibrator model has a 61,800 lbs Peak Force and a frequency range of 3 - 250 Hz.2. Sercel SG-10 Geophones

• Geophones record signals for data acquisition. 6 Geophones were used between eachstation throughout the main line except between stations 1549 - 1598.

3. GTI NuSeismic NRU Wireless Seismic Nodes• The wireless seismic nodes record signals for data acquisition using natural sources.

The geophones were used between stations 1549 - 1598.4. Measuring Tape

• The measuring tape is used to measure 10m between stations before placing geo-phones into the ground.

5. Ground pounder


Figure 2.15: A GTI Wireless seismic node [41].

• The ground pounder is used to create a hole into the ground to insert the wirelessnodes.

6. Sercel FDU unit• The FDU unit converts analog signals from the geophones into digital signals which

can be processed by computers in the doghouse.7. Sercel 408 Cables (Geophone Cables)

• The geophone cables connect sets of geophones together to make a continuous line ofreceivers along the main line.

8. Land Acquisition Unit Cross (LAUX-428)• The LAUX is situated at the beginning of a line and acts as the interface between a

LCI-428 box and the rest of a seismic line. It allows for the active monitoring of a linethrough a graphical display.

9. Land Acquisition Unit Land (LAUL-428)• A LAUL is placed at every 40 meters along a seismic line to serve as a power source

for geophone cables and geophones, as well as acting as a communications hub.10. Line Control Interface (LCI-428)

• The LCI-428 serves as the management unit for a seismic line, and can handle 10,000channels at a time. It serves as the main interface connecting a seismic line to acomputer or server for seismic line management.

11. Doghouse• The doghouse was the base of operations for any CSM field camp deep seismic survey.

It served as the recording station for data collected along the seismic line and hostedservers and other geophysical equipment necessary to actively monitor a seismic line.It was also capable of performing preliminary QC checks on data collected in the field.

The programs used for data processing is as follows:

12. Sercel e-428• This software controlled the seismic line, and was also capable of computing stacks or

correlations before recording the data.13. SeisSpace 2D Software

• This software processes data acquired in the field to produce a stratigraphic model ofthe subsurface through series of noise assessment and quality checking.

14. SeiSee SEGD-2-SEGY• This software was used by Gene during the field camp to visualize the seismic data

obtained from the Doghouse.

2.8 GPS 61

2.8 GPS2.8.1 Introduction

The GPS method is used to take precise location and altitude measurements for flags alongboth the main line and the student site. The instrument used for the differential GPS measure-ments was the Trimble R10 GNSS system, which contains one base station unit and two roverunits. For the hand-held GPS system, the system was the Garmin-eTrex 30X Hand-held GPS.The GPS measurements are used to locate survey lines and to make terrain based corrections ofcollected geophysical data.

2.8.2 Theory2.8.2.0.1 Differential GPS (DGPS)

The network of GPS satellites orbiting above the Earth emits a signal that is then received bythe base station down at the surface. The satellites transmission includes the exact position andtime that the signal was transmitted. The base station receives this information and records thetime the signal was received. The difference between the time the signal was transmitted and thesignal was received is the travel time. Since the signal propagates at roughly the speed of light,the travel time can be used to calculate the distance between the satellite and the base station.However, this distance is unspecific, and the location of the base station could be at any pointon a sphere that distance away from the satellite. A second satellites transmission allows thebase station to create a second sphere, and the intersection of those spheres creates the circularregion of space where the coordinates of the base station is located. A third satellite reduces theregion down from a circle into two potential points where the base station could be located. Anyadditional satellites, with a minimum of 4 satellites [42], can exactly resolve the location of thebase station, specifically its altitude and UTM coordinates. The GPS satellite network is designedso that most locations on earth are always within view of 6 satellites [42] which allows the basestation to determine exact positions with a high degree of accuracy. Additional time allowingthe base station to triangulate improves accuracy, and also allows the system to correct for anyerrors accrued in distance calculation. Longer base station is allowed to obtain a more accuratecorrections sent to the rover, hence creating an even more accurate location. If the initial baseshot is recorded for over 4 hours, the file can be uploaded to the NOAA’s OPUS site and correctedagain against their CORRS stations for even greater accuracy.

Once the base station has a known location fix, the rovers can begin their survey. The basestation obtains corrections by taking a reading of the location and comparing that to its knownlocation from the long static shot for a more accurate shot, and then sends the difference betweenthose two locations to the rover.

2.8.2.0.2 Hand-held GPSThe hand-held GPS method consists of a GPS receiver that reads signals from GPS satellites

orbiting above the surface of the globe. This method trigulates its location similarly to the DGPS,but without the base station to use as a reference, any errors in in the traveltime caused byparticulates in the atmosphere are greatly magnified compared to the DGPS readings.

2.8.3 Background2.8.3.0.1 Equipment List

• 3 Trimble R10 GNSS receivers• 2 Trimble TSC3 handheld data collectors• 1 Trimble TDL450 radio• 1 Radio antenna• 2 Rover survey poles• 1 Base tripod• 1 Base tribrach


• 1 Car battery• Multiple Garmin eTrex handheld GPS’

2.8.4 Objectives

The objective of the GPS is to get as accurate position and elevation measurements as possiblefor use in other methods. The most important portion of what is accomplished by the GPS teamis accurate altitude measurements for the construction of a topography profile along both themain line and the student site. This information is used to perform gravity data corrections,both the free-air correction and a complete Bouguer terrain correction. The data is also used toperform DC resistivity terrain corrections for accurate data inversions.

2.8.5 Survey Locations

GPS points were measured along the main line and student site for each flag on each line.DGPS was the primary method used to obtain the GPS coordinates, since it has the higher degreeof accuracy required for terrain corrections. Hand-held GPS devices were used as back up for thesame locations. They were also used for the DC line S3 and gravity tie-in stations where DGPSwas not practical to set up. It is important to note that the MT surveys used GPS equipmentattached to the survey device, and used neither DGPS nor hand-held GPS.

2.8.6 Data Acquisition

For the hand-held GPS method, the multiple eTrex handheld GPS is positioned on top ofthe flag location and the GPS measurement is acquired. For the DGPS method, the DGPS basestation is assembled and setup up to take recordings throughout throughout the day with tworovers set up to collect the GPS coordinate points. The tripod is setup over a marked base stationlocation using a laser for accuracy. The tripod is leveled and the distance between the flag andthe lever is measured. The R10 is attached to quick release to the top of a black base and abattery is attached. An external radio is hooked onto the tripod and the a mount for the antennais attached with a disk. A cable is then attached from the antenna to the external radio andfrom the antenna to R10 attached to a tripod. Each of the two rovers has a TSC3 computerassociated with it. The TSC3 is then set up to take general surveys as an open job with an inputof the job name allowing us to take a set an auto store points, number of measurements, andoccupation time. As the rover is leveled directly over the flag, the ”measure” button is clicked inthe TSC3 computer to acquire the GPS coordinates. Typically, to increase efficiency, each of therovers leap frog over each other as one rover does even-numbered flags and the other rover doesodd-numbered flags.

2.8 GPS 63

(a) Leveled Rover and TSC3 Computer (b) Base Station

Figure 2.16: DGPS Data Acquisition

2.8.7 ProcessingAt the end of the day the DGPS data is downloaded into the main database by using a USB

to export the acquired data points. During processing, the data spreadsheet is translated fromlatitude and longitude to UTM coordinates. The spreadsheets are then fragmented to encompassthe locations that were covered by a particular method, like inversion sections for DC or mid-points for all the TEM sites. As a secondary measure used for data correction, all the job filesfrom the TSC3 computer are re-downloaded and checked with the files that had been uploadedto the Google Drive in the field. Any duplicate locations or QC plots are removed to determine thecorrect duplicate points. The missing flag locations are interpolated by creating flag coordinatesthat are directly in between the known locations right before and right after the missing flag. Asthere is no static base station shot taken, the OPUS corrections were not submitted.

There are a few processing stages implemented specifically for seismic processing. The pro-cessing begins by moving the seismic node locations from the main line file into its proprietarylocation. The passive wireless node positions are then transferred from the data file that camefrom the node.

GPS processing for use in gravity processing is more intensive. Using GPSeismic, the verticalprecision is extracted for both Main Line and S0 Student Site lines for use in gravity processing.This was done in order to see the parts of the line that had less accurate elevation data. Thisallows the gravity inversion to be more skeptical of the free air correction at those positions. Thiscorrection is particularly important along the S0 Line at the Student Site as the trees causedissues in obtaining accurate GPS positions. Another stage of GPS processing for use in grav-ity was the gravity tie-in to the 2016 line. This was implemented by correlating the acquired2017 hand-held position measurements with the more accurate DGPS measurements from theprevious years data set that correspond to the same location.

2.8.8 Errors and UncertaintiesFor GPS, there is a difference between the actual location of the base and the reading location

when taking a RTK, or rapid-static shot. This is because of the small amount of error in thetime-distance calculation made by the satellites, which is caused by water and particles in theatmosphere and multi path errors. Overall, because the base and rover are assumed to be in thesame area, with the satellite signals traveling through the same section of atmosphere, the errorsseen at the base station are assumed to be the same errors seen at the rover. The base transmitsthese errors over the TDL450 data radio and the rover uses these to correct its position. Thisallows the rover to stay in a position for short periods of time instead of taking a static shot. The


base station never took a long static shot while out in the field. This means that the base stationwas sending corrections to the rover based on a slightly incorrect true location. This means thatall of the stations shot will be shifted by however far off the initial base shot was from its truelocation. Position measurements will all be accurate relative to each other throughout a day ofsurveying, however, with some variability from one day to the next. This error is minimal andwill therefore not be an issue for any of the methods relying on location information.

The main issues specifically with DGPS arose during the surveying of Line S0 on Reservoir hill.S0 was located on Reservoir Hill, which is heavily forested. According to the survey job logs, ataround station 42 on Line S0 the accuracy of the locations went down dramatically. Most of thepoints beyond this flag were stored with poor precision, excess tilt, or position compromised tags.These station locations were stored out of the tolerance range, causing the vertical precision toreach as high as 3 meters. For gravity, this forced the choice to either disregard these elevationsand not create a gravity model for Line S0, or to try to interpolate the elevation between stationsthat had good information. In the future more time might need to be committed to areas withheavy tree cover to ensure accurate shots are taken, or to use a total station in these areas.

2.8.9 Recommendations2.8.9.1 Field Recommendations

For future surveys, it is advised to take a long static base reading on the first day in Pagosa tocreate a baseline that can be OPUS corrected. A static-rapid base shot each subsequent morningcould then be corrected to that initial base shot to tie each day’s measurements together. Thismay be excessive, but it will create as high an accuracy as possible. Combining other methodssuch as using a total station will be useful in order to reduce error due to interference in locationswhere the DGPS cannot accurately take readings.

2.8.10 ConclusionsApart from minor issues such as a few missing stations, DGPS along the main line went well

and had acceptable accuracies for use by the methods located there. The most notable thing tomention is the issues along Line S0. The issues could have an effect on gravity processing, butthe errors accrued are understood. The hand-held GPS is not an ideal tool, but it was useful togive an approximate location for the survey sites.


2.9 Passive Seismic2.9.1 Introduction

The passive seismic method measures naturally occurring low frequency seismic waves suchas from wind, ambient noise, ocean waves, and other similar sources to detect movements withinthe subsurface. Subsurface fluid flow is the most relevant source to the survey. This method isuseful in evaluating earthquake hazards, managing reservoirs in terms of deformation and micro-tremor sensitivity, and detecting fluid leaks. Unfortunately, due to the absence of a controlledsource, this method has a poor signal to noise ratio.

2.9.2 TheoryPassive seismic methods aim to measure ground vibrations not directly induced for measure-

ment purposes as they are with active methods. These waves have long periods and are producedby a wide variety of passive sources. Wireless NuSeis NRU-1C geophones developed by Geophys-ical Technology Inc. were used to measure the analog amplitude and record them into digitalsignals as a function of time.

2.9.3 Background2.9.3.1 Equipment List

• NuSeis NRU- 1C, wireless geophone nodes with built-in GPS• Pile driver• 9V battery• GTI App on Android tablet

2.9.4 ObjectivesThe main objective of using the passive seismic method is to understand the natural ground

water flow in relation to the Mother Spring. Three passive seismic survey locations were chosenon the western edge of Reservoir Hill surrounding the Mother Spring and the San Juan River asseen in the Figure 2.17 below. In the scope of this study, we will look to identify a differencebetween noise sources during the day and night time.

2.9.5 ExpectationsWe expect to be able to see a great amount of noise from the daily activities of Pagosa Springs,

as the nodes were left near downtown. However, anthropogenic noise is expected to reduceduring the night time.

2.9.6 Survey MapsThe nodes were arraged in three different sites, as shown in Figure 2.17. All three sites were

located in downtown Pagosa Springs, near the San Juan River and the Mother Spring.

2.9.7 Data AcquisitionThe seismic data acquired during the survey is stored in the wireless NuSeis NRU-1C geo-

phones. These wireless geophones were used to measure the analog amplitude and digital signalsas a function of time. The data is retrieved remotely by accessing the geophones, The positionand the site information are created on the GTI Application. The wireless NuSeis NRU-1C geo-phones were acquiring data for a total of 38 hours. According to the headers in Seispace, thedata was gathered from May 17 2017 at 4:00pm Mountain Time until May 19 2017 at 6:00amMountain Time.


Figure 2.17: Overview map of all the passive seismic sites in Reservoir Hill. The black dots repre-sent the approximate center of each of the surveys. Site 1 (Star) had the triangular survey setup,while Site 1 (Line)’s setup was linear from North to South, spanning about 100 m. Site 2 had a star-shaped setup with the nodes going radially outward of the center node. Site 3 had a square-likesetup with varying position of the nodes inside of it.

2.9.7.1 Survey Parameters

There are several steps needed to setup the passive seismic survey. First, holes are neededto be dug using a pile driver. Then, the nodes need to be activated using the electrode battery.The NuSeis NRU-1C geophes are needed to be placed inside the holes. The nodes then need tobe synced with the GTI applications in order to begin recording, check the node functions, andrecord accurate GPS locations.

Site 1: This site used 21 nodes with 5 meter spacing. The nodes were arranged in a triangularpattern consisting of 3 triangles encompassing each other as seen in Figure 2.18. The dimensionsof each side of the inner-most triangle are 3.2m with midpoint distance of 1.9m. The side of themiddle triangle is 10.4m with midpoint distance of 6m. The side of outermost triangle has alength of 24.3 m and a midpoint distance of 14m. Additionally, the length of the line in the siteis 100 m.

Site 2: This site used 24 nodes with 1m spacing in a star-shaped array as seen in Figure 2.19.

Site 3 This site used 16 nodes with 4m spacing with the exception of the 4 nodes located12m away from the center node. The nodes were arranged in an orthogonal array as seen inFigure 2.20.


Figure 2.18: Receiver array for the first site. The UTM coordinates are UTM 13S WGS 84 whereSouth point is (0321930, 4127166), North point is (0321956 ,4126204), and the center midpointcoordinates is (0321956 ,4126204). These points were acquired using hand-held GPS.

2.9.8 ProcessingWe analyzed the data from two of the nodes, one seismic node from site 1 and one seismic

node from site 3. The data was arranged in one minute traces in .segy files, one for each node.For each node, the data is organized using the SegyMAT package for MATLAB. Using the

ReadSegy function the data was read into MATLAB. Afterwards, the data for each node wasstacked into a single time series. The data was then divided into one hour segments and madeinto separate figures for qualitative analysis.

2.9.9 Errors and Uncertainties2.9.9.1 Field Errors

The Passive Seismic method is sensitive to several errors and uncertainties. A major sourceof error was the downtown location of the survey. Measurements were taken close to areas withhigh traffic overwhelming the smaller signals coming from the geology. This error can be reducedby limiting the data to quiet times, such as between 2 a.m. and 3 a.m. where human activityand noise is minimum.

Additional errors are due to the orientation of the receiver array as the water flow withinthe subsurface of Pagosa Springs. In order to make the receiver array maximally sensitive tothe water flow, the survey should be perpendicular to the water flow. If the arrays were insteadparallel to the water flow, the time difference becomes negligible and make receivers less sensitiveto the flow. In order to address the directionality problem, the array structures were made into2-D shapes to be sensitive in all directions. In this case, this error might be higher as the data isonly extracted from two nodes.

2.9.10 Recommendations2.9.10.1 Processing Recommendations

It is also very important to know how to read and convert the data from the GTI application,increasing the amount of time available to fully process the data. We would recommend usingall of the data from the nodes laid out to start out processing a new survey with some basic


Figure 2.19: Receiver array for the second site. The UTM coordinates are UTM 13S WGS 84 for thefar-east point is (0321772,4127031), and the west point has coordinates of (0321765,4126957).These points were acquired using hand-held GPS.

expectations of how anthropogenic noise behaves to more easily remove it.

2.9.10.2 Field Recommendations

For future surveys, finding a less disruptive survey site would be ideal. Sites that are relativelyisolated from human activity produces the best results for the natural sources. A secondaryrecommendation is to go in with a general understanding of the fluid flow of the site, which isimportant to maximize the number of nodes perpendicular to the flow.

Since there were only two nodes worth of data rather than an array of nodes, there are notmuch information regarding the fluid flow around the Mother Springs area gained from thismethod. However. the correlation between the amount of data through time was still visible fromeach of the successful nodes. From this correlation, any possible anomalous trend through timeis inspected and interpreted.

2.9.11 Interpretation

Node 1 Data Interpretation: This node is from the first site of the survey line at station 7.The gathered data from this node is in Figure 2.21. In the first 4 hours, the data was moderatelynoisy. The noise continuously progressed and regressed with a peak in between 16 to 26 hours,corresponding to 8:00am to 6:00pm of May 18 2017 as seen Figure 2.22. This noise is mostlikely due to human noise such as traffic.


Figure 2.20: Receiver array for the third site. The UTM coordinate is UTM 13S WGS 84 with thecenter point as (0321869,4126353). These points were acquired using hand-held GPS. Black dotsare the geophones arranged by the specified spacings.

Figure 2.21: Passive seismic data from Node 1.


Figure 2.22: Passive seismic data from Node 1 for time 10:00am and 6:00pm at May 18th, withinthe noisiest time frame of the 38 hour survey. Because this node had much lower amplitudesoverall than Node 2, the scale has been adjusted to better display the trends.

The quietest times were in between 30 to 38 hours, corresponding to 10:00 pm to 6:00 am of18-19 May 2017. There are no anomalies or repeating spikes within this time frame as seen inFigure 2.23.

Figure 2.23: Passive seismic data from Node 1 for time 3:00 a.m. and 4:00 a.m. at May 19th,within the quietest time frame of the 38 hour survey. Because this node had much lower amplitudesoverall than Node 2, the scale has been adjusted to better display the trends.

Node 2 Data Interpretation: This node is located in station number five within the thirdsurvey line. The data from this node seems to have more noisy than the data from the othernode as it is located in the park. The data from this node, however, follows the same trend asthe previous node. The data begins relatively noisy in the beginning, not too noisy in between 5to 15 hours, very noisy in between 16 to 26 hours, and quiet with low noise in between 30 to 36hours. The Figure 2.24 and Figure 2.25 represent the noisy passive seismic data from node 2


and the Figure 2.26 represents the quiet times. There are no apparent anomaly throughout thedata.

Figure 2.24: Passive seismic data from Node 2.

Figure 2.25: Passive seismic data from Node 2 for time 10:00 a.m. and 6:00 p.m. at May 18th,which is within the noisy time frame.


Figure 2.26: Passive seismic data from Node 2 for time 3:00 a.m. and 4:00 a.m. at May 19th,which is within the quiet time frame.

2.9.12 ConclusionsFrom the preliminary analysis, we can see a difference in noise levels between quieter night

hours and the busier daytime hours. We cannot specifcally point out any broader trends, like asignal from the San Juan River or groundwater sources.


2.10 Rock Physics2.10.1 Introduction

The physical measurements of the rocks at depth give an approximation of the rock proper-ties. These measured values help make the inversions done by other geophysical methods asaccurate and realistic as possible. For this method, the core samples provided by Pagosa Verdewere obtained from a local well, TG-1, located southwest of Pagosa Springs. The well’s exact lo-cation is seen in Figure 2.29. While equipment for both velocity and bulk density measurementswere available for use the bulk density measurements have been included, but it was determinedthat the velocity measurements would not be useful unless measurements could be performedunder pressure similar to the rocks differential pressures in the ground. The density measure-ments were performed with air-filled pores, and the cores were not saturated before or duringmeasurements.

2.10.2 TheoryThe bulk density of a rock incorporates the fluids inside the rocks. Basically it is a density

measurement of an entire rock, combining grains and any fluids, such as air or water, in thepores. The bulk density of a rock is measured in different ways. One simple method to determinethis physical property is to incorporate Archimedes principle of buoyancy. The principle statesthat when an object is immersed in fluid, the weight of the object will be equal to the upwardbuoyant force acting on the object. A graphic representation of this principle is available forviewing in Figure 2.27.

Archimedes’ principle of buoyancy is related through Equation Equation 2.43. It states thatthe bulk density of an object is equal to the mass of the object divided by the object’s volume.

ρbulk =Mobject

Vobject(2.43)

In Equation Equation 2.43, ρbulk is the bulk density, Mobject is the mass of the object, andVobject is the volume of object. Archimedes principle also states that the volume of water displacedby the object is equal to the volume of the object itself. With this information, the volume of therock was obtained by measuring the mass of the rock suspended water. The rock volume isobtained by dividing this value with the density of water. Knowing the mass of the rock, the bulkdensity is then calculated using Equation Equation 2.43. It is is important to note that the bulkdensity also accounts for the pore space and the pore fluid density.

2.10.3 BackgroundWe performed the petrophysical measurements to provide more starting information to use in

interpreting the field data. The density measurements were used to make models for the gravitydata.

2.10.3.0.1 Equipment• 1 L beaker• Strings• Digital scale• Water• Core samples

2.10.3.0.2 SetupFirst, the beaker is filled with water, to between 750mL and 1L. The water-filled beaker is

placed on the digital scale and the scale is then zeroed. The core sample is placed inside of thebeaker. The mass is recorded in grams, recorded to the 100th of a gram. The sample is thenplaced outside of the beaker and tied with a string, and the digital scale is zeroed again. The


Figure 2.27: Archimedes’ Principle of Buoyancy Method [43]

sample is suspended by the string in the water-filled beaker. The mass of the suspended sampleis recorded to the 10th or 100th of a gram, depending on whether the scale settles or varieswithin a small range of values. This is repeated for each sample.

Because sample 10 showed great fragility and fridge-ability, it was not used in measurementsin an attempt to preserve the core.

2.10.4 ObjectivesThe objective of this method is to obtain the bulk density of the various rock units from the

core samples recovered in Pagosa Springs. This information is highly useful for the interpreta-tion of other geophysical methods, particularly gravity inversion and seismic modeling. Thesemeasurements were carried out twice, once in Pagosa Springs but with very high uncertainties,and then with greater precision at the Colorado School of Mines.

2.10.5 ExpectationsThe core samples that were measured all originate from sedimentary formations. These for-

mations are suspected to be the Wanakah Formation, the Morrison Formation, and the DakotaSandstones based on the depth information that came from the samples. With that in mind, itis expected that the core samples should be around 2.0− 2.6 g/cm3, which is the typical dry bulkdensity of sandstones.


2.10.6 Data Acquisition2.10.6.0.1 Core Samples

Figure 2.28: Core sample which densities were measured for this method

• Sample 1: This sample is from about 1300 ft down, around the boundary between theMorrison and the Wanakah formation. The sample has a white color with black lines allover it. The sample is a fine-grained, well-sorted sandstone. It is also made out of anhydritematerials, indicating evaporite deposits.

• Sample 2: This sample is from 1117 ft down, solidly in the Morrison formation. Thissample is a reddish fine-grained, well-sorted sandstone with ***. There is no carbonatepresent in the sample

• Sample 3: This sample’s depth is unknown. It is a fine-grained, moderately-sorted sand-stone with a gradient color from light to dark gray. Mineralization has occurred in fracturedparts of the rock. It creates bubbles when submerged in water, indicating high porosity.

• Sample 4: This sample’s depth is unknown, it is a fine-grained, moderately-sorted sand-stone which a gradient color from medium to dark gray.

• Sample 5: This sample’s depth is unknown, it is a very fine-grained, very well-sorted blackmassive shale. It is most likely from Mancos.

• Sample 6: This sample is from 839 ft down, in the Morrison formation. This sample is ina gray medium to fine grained sandstone

• Sample 7: This samples depth is 837 ft and is a medium soft medium gray very fine grainedsandstone. The hand sample is very continuous

• Sample 8: This samples depth is 832 and is a dark red siltstone with some quartz andother mineralization. The sample is brittle and falls apart after in contact with water.

• Sample 9: This samples depth is 831 ft and is has a light gray color. It is a very fine-grainedand well-sorted sandstone and overall very continuous.

• Sample 10: This samples depth is 880 ft and is a reddish fine-grained, well-sorted sand-stone, with bands of calcite. Due to its fragility, the measurement was not done for thissample.

• Sample 11: This samples depth is 877 ft and is reddish, very fine-grained, well-sortedsandstone. After submerging the sample in water, the sample fractured into two.

• Sample 12: This samples depth is 878 ft and is light gray, coarse-grained, moderately-sorted sandstone with fine grains of calcite throughout the rocks. It has fine horizontalfractures.

• Sample 13: This sample’s depth is 934 ft and has a gradient color of light gray to green. Itis siltstone is has vertical fractures filled with quartz.


• Sample 14: This samples depth is 932 ft and is a dark to light gray sandstone with verticalfractures filled with quartz. Potentially muscovite as flakey mineral is seen on side ofsample. Sample transitions from light to dark, increased carbon content?

• Sample 15: This sample depth is 932 ft and is a green medium grained sandstone withquartz mineralization

• Sample 16: This samples depth is 929 ft and is medium gray very fine sandstones. Subvertical fractures with granular quartz exposed on the surface.

• Sample 17: This samples depth is 927 ft and is medium gray fine sandstone with few fineto medium calcareous grains. Few sub vertical fractures filled with quartz.

2.10.7 ProcessingBefore solving for bulk density, the measurements were converted to kilograms for consis-

tency. The displacement mass was then converted into a volume using the value of 1000 kg/m3,based on the definition of SI Liters, to convert from water to volume. The mass of the rock wasthen divided by its volume to find the bulk density.

2.10.8 Interpretation2.10.8.0.1 Results

Table 2.1: Measured mass of the core samples.

Core Sample Number Rock Mass (g) Mass of Suspended Rock (g) Uncertainty (g)1 376.83 129.19 0.22 320.62 121.30 0.23 166.58 69.59 0.24 161.21 62.2 0.25 128.76 49.77 0.156 290.02 108.5 0.27 123.47 46.9 0.28 232.42 88.8 0.29 280.75 105.1 0.210 No Data No Data No Data11 145.11 52.2 0.212 83.26 32.54 0.1513 282.69 108.85 0.214 344.43 129.9 0.315 124.36 48.1 0.216 358.63 138.5 0.417 195.91 74.90 0.2

The bulk density of the various samples was calculated using the values above in TableTable 2.1 and using Equation Equation 2.43 in the theory section. The results as seen in TableTable 2.2.


Table 2.2: Calculated bulk density and volume of the core samples

Core Sample Rock Volume (m3) Rock Bulk Density (kg/m3) Uncertainty (kg/m3)1 1.292E-4 2916 4.5082 1.213E-4 2643 4.3513 6.960E-5 2393 6.8584 6.222E-5 2591 8.3075 4.97E-5 2587 7.7746 1.085E-4 2673 4.9187 4.690E-5 2632 11.188 8.880E-5 2617 5.8829 1.051E-4 2671 5.07410 No Data No Data No Data11 5.220E-5 2779 10.6112 3.254E-5 2558 11.74113 1.089E-4 2597 4.76314 1.2990E-4 2651 6.10915 4.809E-5 2585 10.7116 1.385E-4 2589 7.45717 7.490E-5 2615 6.966

2.10.8.0.2 Errors and UncertaintiesThe errors in the bulk density measurement come from movement of the suspended rock,

which was done by hand. Additional errors could come from the slight chipping of the rocksamples between measurements for a couple samples, specifically samples 8, 11 and 13. Anothererror source would be any water impurities affecting the density of the water. Uncertainties inthe measurements come from scale variations, possible excess water on the rocks, and possiblewater infiltration in the rock pores. Additionally, there is an error in between the saturated andunsaturated sample weight. In other words, if the sample is saturated or more damp, the weightwould be heavier.

2.10.8.0.3 RecommendationsThe rocks were not held firmly enough when suspended in water, altering the measured mass

of the displaced water. It is recommended that in the future stable equipment, such as a hangingdevice, should be used to suspend the rock sample in order to minimize errors associated withsuspending the rock sample. It is also recommended that either porosity measurements be com-pleted as well or that the bulk density measurements be performed on water-saturated samplesto better represent subsurface material.

Velocity and resistivity measurements on the rock samples were not performed. It is highlyrecommended that future field camps should dedicate a more focused section for rock physics inorder to obtain more detailed information about the physical properties of the rocks from each ofthe formations.

2.10.9 ConclusionsThe rock physics measurements on the core samples were done to obtain quantitative values

of each rock’s physical properties. These measurements were fairly basic, as only bulk densitywas measured. These measurements were used along with density estimations from past yearsto calibrate gravity inversion models.


2.11 Well Logging2.11.1 Introduction

Well data is immensely important to study in tandem with surface methods because it pro-vides quantitative information to help constrain inversions, direct cross-sections, and providedirection for future studies. Depending on the available well tools, well logs provide informationsuch as depths of formations, lithology types, resistivities, velocities, and more. Wells have beendrilled in Pagosa Springs starting in the early 1900s. While many of these wells are undocu-mented, many of the more recent wells contain valuable information relevant to this study. Wewill be using gamma ray, sonic, and resistivity logs, as well as temperature gradients in variouswells.

2.11.2 TheoryOne of the key well log types obtained for a few of the wells in Pagosa Springs was gamma

ray logs. Gamma ray tools can either measure radioactive sources in the rocks themselves or thelogging tool emits gamma rays. For the first method, it is typically used to determine intervalsof shales and sands. This is because in general shales have higher concentrations of Potassium40, Thorium, and Uranium. A highly feldspathic sandstone, however,can fool this method. Thesecond gamma ray method is used to calculate density. The gamma rays are emitted from thesource end of the tool and move through the formation. The rays collide with the electrons inthe formation and are reflected, while the receiver end of the tool records how many electronsare sent back into the tool. Denser materials will have more collisions, and therefore moreelectrons will make it back to the tool. Porosities can also be calculated from the gamma densitylogs if something is known about the rock matrix and fluids. This process is characterized inEquation 2.44.

ρbulk = ρfluid + ρmatrix(1− φ) (2.44)

In Equation 2.44 ρbulk is the bulk density, ρfluid is the density of the fluid in the pore space,ρmatrix is the matrix density of the rock, and φ is the porosity of the rock.

Sonic logs were also very important to our data processing and geologic characterization.These logs are key for seismic depth migrations. The velocities measured by the sonic tool canbe used in the seismic velocity model to obtain more accurate travel time and therefore a moreaccurate depth migration. Sonic well logging tools rely on Snell’s law. Snell’s law gives the rela-tionship between angles of incidence and refraction for a wave impinging on an interface betweentwo media with different indices of refraction. The law follows from the boundary condition thata wave be continuous across a boundary, which requires that the phase of the wave be constanton any given plane, resulting in Equation 2.45.

n1 sin(θ1) = n2 sin(θ2) (2.45)

where θ1 and θ2 are the angles from the normal of the incident and refracted waves, respec-tively.

Finally, we will be relying on water temperature information from as many wells as we canfind in order to gain a more complete understanding of the fluid flow patterns of the region.Much of this information has been provided to us by Pagosa Verde, LLC and Paul Morgan fromthe Colorado Geological Survey. Some of these wells were discussed previously in the hydrologysection.

2.11.3 ObjectivesA combination of useful well logs were obtained to best characterize the geology for inversion

and interpretation purposes. Most of the well data is located in the areas surrounding the studentsite and Pagosa Springs, rather than the main line down near Chromo. However, the wells willbe used to assist in providing a broad geological context, and characterize both survey sites. We


have also received anecdotal information from local residents of Pagosa Springs and Chromo.These pieces of information helped to guide our surveys and target locations by providing asmall amount of background information. We were unable to locate or verify most of the wellinformation we were given though the information is still important to note.

2.11.4 Expectations

We expect the well data to help us piece together how fluids are moving in the area of PagosaSprings. We are also expecting the sonic logs to aid our seismic depth migrations for a moreaccurate image of the subsurface of our main line. When all of this information is combined withthe data from our other survey methods, we believe we will be able to clarify several questionsposed in this report.

2.11.5 Survey Maps

Wells were anlyzed from both the Pagosa Springs and near the main line in Chromo. Thewells in Pagosa Springs were logged by Pagosa Verde, LLC, who provided their logs. The locationsof these wells is shown in Figure 2.29. The wells near the main line were drilled at various times.Their locations are shown in Figure 2.30.

Figure 2.29: Well locations in the Pagosa Springs area provided by Pagosa Verde, LLC


Figure 2.30: Relevant well locations for the 2017 main line. Information and locations found inCOGCC database

2.11.6 Data AcquisitionWell logs and information were collected from various sources and no wells were drilled or

measured by the Colorado School of Mines Geophysics Field Session. The most used well infor-mation was given to us by Pagosa Verde, LLC. Some data was gathered via word of mouth fromlocal residents of Pagosa Springs and Chromo, and the rest of the data was collected from theColorado Oil and Gas Conservation Commission online database.

These wells, and some of the conclusions drawn from them, were provided by Pagosa Verde,LLC. Their findings were reported by Dr. Leland Mink and other contributors. Many of theirfindings included work done by Galloway and will be used for further interpretation and will bereported in the Final Interpretation section of the report. One of the key findings from Mink’sreport, however, was an integration of Pagosa Verde data with data from the Galloway report [1]as seen in Figure 2.34

2.11.7 ProcessingProcessing of this data was relatively uninvolved when compared to other methods. Much

of the data was processed by others and given to us to inform our work. The logs had to beannotated so that the information was more relevant for the other methods. We annotated thelogs by labeling the intervals of each geologic unit based on the core data given to us for eachwell.

For integration, well data was vital to understanding and fine tuning all other methods. Theformation tops from available wells were used as controls for cross sections, sanity checks for


seismic picks, and various other conclusions. The temperature distributions for the availablePagosa Springs wells were used to make sense of the data seen by DC resistivity and Self-Potential. The vast amount of anecdotal well information also helped to direct the targets ofthe surveys and weed out useless approaches. Overall, the TG wells drilled by Pagosa Verdeshowed that hot water continues to spread east but not directly south of Pagosa Springs. It alsoconfirms that the hot water is still occurring at relatively shallow depths and isn’t rising from thebasement at those locations, which means it must be flowing from somewhere north or west ofthe TG wells.

2.11.8 Errors and UncertaintiesThe error and uncertainty in this data is largely a mystery to us because we were not part

of the acquisition stage. But there is some uncertainty involved in the interpretation of the welllogs. In order to help the inversion processes of some of the other methods, average values foreach formation were chosen for resistivity and velocity values. This means that a lot of the subtlechanges within the layers were smoothed out to keep our models more simplistic.

2.11.9 ConclusionsMost of the conclusions drawn in this report were largely tied to the well data. Specifically,

the geologic understanding of the region can be directly tied to the formations identified by wellcores from the region. Water quality and temperatures recorded in Pagosa Springs show that,whether there is faulting going on or not, fluids are moving in the area. The well data also showsthat there are fluids on both sides of our suspected Reservoir Hill fault. The presence of hotwater in the Galloway well on the West side of Pagosa Springs seems to suggest that the water isnot originating from the Reservoir Hill fault but it does appear as if the fault could be affectingthe flow.

2.11.10 RecommendationsOne of the greatest difficulties with the well data is the vast range and quality of the data.

There are a large number of wells that are only known through word of mouth, many wellsthat barely penetrate the subsurface, and many that do not have relevant documented data.Additionally, the wells range from the 1940s to 2016. Because of this, making sense of all ofthe data can be a challenge. We recommend creating a master file that lists every known welltied to its location, useful data, internet links, age, and any other useful information. Withoutthis kind of organization the data is very difficult to understand and integrate. We would alsorecommend drilling northwest of town to gather more information about that side of the MotherSpring. Without that information, it will be very difficult to more clearly characterize the fluidflow in the area.


Figure 2.31: Part of the TG-1 Sonic Log


Figure 2.32: Part of the TG-1 Gamma Ray Log that has been geologically annotated using providedwell cores


Figure 2.33: Figure taken from the Mink report showing well temperatures in the Pagosa Springsarea [44].

Figure 2.34: Figure taken from the Mink report showing temperature gradient contours using Gal-loway wells [44]. TG wells were posted on the map to show correlation.

3. Chromo Main Line

Contents




3.4 Magnetics 96



3.7 Gravity 114



3.10 Conclusion 152


3.1 Overarching ObjectivesThe primary objective at the main line was to characterize the fluid flow in the area. Specifi-

cally, we wanted to charaterize the fluid flow controls, which were hypothesized to be the dikescrossing the line. To accomplish this goal, we used various geophysical methods to image bothsides of the dikes.

86 Chapter 3. Chromo Main Line

3.2 Main Line Geology3.2.1 Background Information

The Chromo area is of particular interest due to its disparate geothermal and hydrologicalsystems. The Archuleta Anticlinorium and igneous dikes are the dominant geological structuresin the area. These structures are found commonly in other known geothermal systems.

The underground water flow patterns for Chromo area’s geothermal springs are currently un-known, as well as any potential relation to Pagosa Springs geothermal system. As seen in Figure1.7, it is known that the geologic formations present within the main line are Precambrian Base-ment, sediments from the Ancestral Rockies, the package of Jurassic-aged sandstones includingthe Morrison Formation, the Mancos shale, Mesaverde Sandstone, and the Lewis Shale.

3.2.2 Preliminary InterpretationField observations throughout the Chromo area identified several geologic structures, geolog-

ical features, stratigraphic units, and wells. This information was used to control the geology toproduce a preliminary cross section of the main acquisition line, seen in Figure 3.1 below. Thepreliminary cross section, Figure 3.1, includes the 2017 main line which is connected to the2014 and 2015 seismic line towards the east of the 2017 main line.

As seen in the preliminary cross section in Figure 3.1, the Mancos Shale and MesaverdeSandstone dominate the surface of the Chromo area. From there, the formations were orderedfrom oldest to newest. That order is: Precambrian crystalline Basement, sediments derivedfrom the Ancestral Rockies, Morrison Formation, Dakota Formation, Mancos Shale, MesaverdeSandstone, and Lewis Shale. The thicknesses of the various layers were determined from welllog samples near to the main line.

In the 2014 Seismic Line, located east of the preliminary cross section of Figure 3.2, there isa reverse fault. It was determined to be dipping west and located toward the eastern edge of theChromo anticline with the anticline crest, centered on the geothermal feature known as StinkingSprings. In the 2015 Seismic Line there is a normal fault dipping west, seen in Figure 3.2. Thegeologic layers east of the reverse fault dip west.

On the 2017 main acquisition line there is a large dike located towards the eastern end ofthe main line, located between flag numbers 1394 and 1397. As seen in the geologic map inFigure 3.3, this is the longest and largest dike that intersects the main line. The dike intrudesinto the Mancos shale visible at the surface in the region and is located within a linear ridge.The dike appears to be sub-vertical and sub-linear, following the direction orthogonal to theminimum stress direction of the region. The locals have reported that water wells located east ofthe dike generally have good water quality and flow, with higher temperatures, while the waterwells located West of the dike generally have reported poor water quality, temperature, and flow.There is also a smaller dike located West of the main dike, as seen on the geologic map above(Figure 3.3). Additionally, data from two well logs close to the main line exhibit offset depth ofsubsurface layers, potentially implying a normal fault located in the middle of the line and eastof the major dike. The orientation and length of this fault is still under investigation.


Figure 3.1: Preliminary cross section of the 2017 main line. The information was modeled basedon a combination of surface geology and local well data. Note that the sediments derived fromthe Ancestral Rockies is speculated to be the Entrada and Wanakah Formation, in addition to theregional equivalent of the Fountain Formation. Since there is a lack of well information on thisparticular layer, the name was kept as is.


Figure 3.2: Preliminary cross section of the 2014 and 2015 seismic main line, determined fromseismic data. Note that the sediments derived from the Ancestral Rockies is speculated to be theEntrada and Wanakah Formation, in addition to the regional equivalent of the Fountain Formation.Since there is a lack of well information on this particular layer, the name was kept as is.


Figure 3.3: Map view of the 2014 and 2015 main line cross section found in Figure 3.2, combinedwith the 2017 main line cross section from Figure 3.1. The kink at the center of the line marks theboundary between the 2014 seismic line to the east and the 2015 seismic line to the west.


3.3 Hammer Seismic3.3.1 Objectives

The overall objective of hammer seismic is to analyze and assess the near-surface geology.This includes identifying potential faults, identifying dikes, computing depths, and providing anEarth model to be used by DC. Hammer seismic has the potential to identify several differentgeologic features such as faults and dikes during processing. Hammer seismic data can becombined with the DC data inversion process to constrain the inversion model. This will helpreduce contact resistance signatures recorded in the near-surface.

3.3.2 ExpectationsAcquiring hammer seismic data aims to capture an image of the near-surface geologic struc-

ture. Data acquired yields different velocities both above and below a specific interface. Thesevelocities can be used to create basic earth models depicting the profile of the first interface. Mod-els created with the data from hammer seismic would show the interface unit with overburden.Any discrepancy in continuity of the interface unit indicates external deformation or structure.In the event of there being a relatively young fault, the unit would display a vertical offset visi-ble in calculated Earth models. This survey aims to produce a velocity model, identify the firstinterface profile, and interpret the presence of a dike.

3.3.3 Data Acquisition3.3.3.1 Hammer Seismic Setup

Hammer seismic data is acquired using the following procedure:1. Identify location of interest.2. Determine geophone spacing for depth of investigation.3. Measure line with tape measure, secure at each end with nail.4. Lay takeout cable for length of line. Multiple takeout cables may be needed to complete the

length of the survey line.5. Place geode seismograph at the middle of the survey line (See Figure 3.4).6. Stomp geophones at desired spacing within 10 degrees of vertical to reduce error. Be sure

to check direction of shear wave geophones to have correct polarity if running shear wavesurvey.

7. Connect takeout cables to geode.8. Connect geode to laptop.9. Connect battery to geode.

10. Open MGOS and check noise window to ensure that all geophones are working properly.11. Choose seismic source. (Sledgehammer, weight drop, etc.)12. Choose shot spacing and number of shots for stack.13. Run test shot to ensure that trigger sensor and geophones are working properly.

3.3.3.2 May 23rd, 20173.3.3.2.1 Location

• County Road 542, Main line, off line• Stations: approximately 1390-1403 due to survey being acquired off main line, no GPS

data

3.3.3.2.2 Survey Parameters• Length of Line: 100m• 3 total surveys along line• (1) P-wave shot spacing: 10m• (1) P-wave number of shots: 5• (1) P-wave geophone spacing: 2m


• (2) P-wave shot spacing: 10m, 1m spacing from 21m to 69m• (2) P-wave number of shots: 5• (2) P-wave geophone spacing: 2m• (3) P-wave shot spacing: 1m spacing from 31m to 77m• (3) P-wave number of shots: 5• (3) P-wave geophone spacing: 1m• S-wave survey not conducted

3.3.4 ProcessingWhen processing hammer seismic data, two main types of analysis typically occur–reflection

and refraction. Due to the shallow depth of investigation of the surveys conducted, only refractionanalysis was completed with the data acquired. Refraction analysis is a first-order interpretationmethod that allows for velocities of the overburden layer and the first layer to be computed. Usingthese velocities, the depth to the first interface can be found. Finding depths at multiple shotpoints along each line provides x (location of shot) and z (depth to first layer) coordinates that areused to create a basic earth model. This model can be used to constrain other methods’ inversionprocesses while also providing insight into the near-surface geology. Analysis of the earth modelallows for the interpretation of potential faults or dikes that may influence the hydrology of thearea.

3.3.4.1 Refraction AnalysisRefraction analysis is conducted using MGOS seismograph controller software as well as

Paint XP, however paper and colored pencil can be used. To begin, a shot is selected from oneend of a survey line. Using the knowledge of how waves propagate through the subsurface, thefirst arrival represents the head wave and refracted wave of the first interface. Extend the shotlocation vertically downward as a reference line to find ti for the velocity below the interface,or V2. Slope lines can be drawn to fit the different traces for each channel. The steeper slopeoriginating closest to the shot point is the velocity above the interface, or V1. The more gentleslope extending laterally towards the end of the survey line is the velocity below the interface, orV2. The V2 slope line should be drawn before the V1 slope line as it easier to constrain each inthis order. See Figure 3.5 below for a visualization.

It is apparent in Figure 3.5 that choosing the correct slopes to fit each channel is difficult.Slope lines may be modified to better fit different channels. Once satisfied with the slope linesfor V1 and V2, velocities can be calculated by using Equation 2.3. Computed velocities along withti can be used in Equation 2.4 to yield a depth to the first layer. This process is repeated foreach end of the line along with several shots along the line to create a basic earth model. Datacan be interpreted between points to create a model with reduced noise. It is important to onlyconstrain the model with a few data points. Over constraining the model with too many shotpoints can create a noisy model that is not a realistic representation of the subsurface geology.Results of refraction analysis can be seen below in the Results subsection.

3.3.5 Errors and Uncertainties3.3.5.1 Processing Errors

When picking slope velocities by hand, some inherent error exists. It is difficult to pick aperfect slope that accurately represents each data point due to some smaller fluctuations insubsurface velocity. Even with drastically magnified shot traces, this leads to velocities that arenot perfectly representative of the data collected. This was especially apparent while conductingrefraction analysis and choosing V1 slopes as the slope could only be constrained by 3 geophonechannels. Properly constrained velocity slopes will take more into account. However, the datacollected proves difficult to choose a correct slope. To mitigate this error, a constant V1 value wasused for P-waves and S-waves when calculating depth to the first layer. The V1 velocity used forP-waves is 400 m/s. The V1 velocity used for S-waves is 250 m/s.


One source of noise that was mitigated during processing stemmed from inaccurate z datapoints used to constrain the earth model. Interpretation of data can be done between referencepoints used for the earth model. Noisy data points can over constrain the earth model. Removingthese points and interpreting a depth by averaging the nearest neighboring points yields a morerealistic earth model.

3.3.5.2 Field ErrorsA few errors occurred during data acquisition and processing. To begin, during data acqui-

sition of shear wave data, horizontal geophones are directionally dependent. A few geophonesshow reversed polarity in several shots acquired on different survey lines. While this error isnot extremely detrimental to the overall acquisition, it can cause problems in refraction analysisprocessing when picking slopes to calculate velocity. Another error is the geophone coupling withthe ground. Reservoir hill contained a large layer of dead shrubbery, fallen branches, wood chips,and pine needles. All of this needs to be removed in order to make a proper contact between thegeophone and the ground. Bad coupling will result in poor data acquisition for that geophone.This appears on several shots acquired in the field for a number of surveys. Lastly, one error thatoccurred was equipment malfunction. The wire attached to the trigger sensor attached to thesledgehammer was cut for the survey on May 24th, causing an error in acquisition as randomsamples were taken by the sensor when not in contact with the plate.This was discovered aftertest shots and 1-2 regular shots along the line. The sledgehammer was switched out for a smallerhand held hammer and acquisition continued.

Noise collected during acquisition is due to a number of different sources. Foot traffic oranimals create small seismic sources that are picked up by the geophones. Another sourceof noise is acoustic waves present during acquisition, this includes the sound created by thecontact between the hammer and the plate as well as vehicles, including planes flying over andcars driving by. This noise typically appears before the first arrival. Lastly, strong wind willproduce noise on geophones that are not fully buried. Noisy data can be avoided by accountingfor these sources while acquiring data. However, as time is a constant constraint, some noise isunavoidable during acquisition.


Recommendations for processing include double-checking that the first arrival chosen is aresult of a refraction and not noise. Another recommendation would involve choosing slopesfrom the top of the trace for each channel as well as drawing the V2 slope line before the V1 slopeline. In addition, gaining permission by email to use the test version of the ReflexW softwarefrom Karl-Joseph Sandmeier before processing would prove to save many hours and producemore accurate plots. This allows for one to produce velocities for different shots more easily.The software also allows for an inversion of the data to take place to produce an earth model byfinding depth for every shot location. Using this software will allow for a more comprehensiveanalysis of the subsurface and will streamline the tedious process of creating an earth model byhand. Another recommendation is not to be afraid to start drawing slope lines for the differentvelocities and adjusting them to fit different traces. This will allow for some room for error inproducing different velocities for V1 and V2. Lastly, if V1 velocities are consistently low/high, usea constant V1 velocity to find depth for each shot point.

3.3.6.2 Field RecommendationsRecommendations for future hammer seismic acquisition and processing teams include a

number of suggestions that will make the entire process easier. For acquisition, they must besure to choose a survey area that is relatively flat with minimal shrubbery and ground coverage.This will allow for easier acquisition with reduced chance of error. Less ground coverage makeslaying cable easier and ensure stomped geophones will have a good coupling with the ground.Another recommendation includes checking that all equipment is in proper working condition


before attempting to acquire. This will result in a smoother survey and reduce troubleshootingin the field. Data will also be more accurate than that acquired with malfunctioning equipment.Lastly, be sure to communicate with other crews or people present to reduce noise during ac-quisition. Additionally, lines between SO and S1 should be surveyed in order to determine thereason for the discontinuity found in S1.

3.3.7 Main Line Results3.3.8 Interpretation

Examination of the main line hammer seismic data (Figure 3.6) leads to several main observa-tions. To begin, the elevation profile used was created from the elevation recorded for the neareststation (1390) as a basic reference model when comparing to the shale model. Field notes forthis area are recorded as mostly flat, so this is a realistic reference model. When viewing thecalculated depth, it is apparent that the subsurface thickness to the first interface is closest tothe surface around location 59 meters on the survey line. The depth at this point is 3.22 meterswhile the depth to the first interface increases on either side of this location. The survey wasconducted perpendicularly to the dike, suggesting that this shift in depth to the first interfaceresembles the dike.

Velocity analysis of the main survey line allows for the interpretation of geologic unit or unitsof the first interface. Checking the velocities along the line compared to the point of interestcould further confirm the potential presence of the dike in the data. The average value of V2P-wave velocities is 3076 m/s. In comparison, the V2 P-wave velocity found at the location of thesuspected dike is 4111 m/s. Drawing from the 2016 Field Camp’s findings, the P-wave velocityof the first interface is consistent with what they identified as Mancos Shale. [45] However, thelocation of this years main line survey suggests a different geologic unit, most likely Lewis Shalewhich has a P-wave velocity range of 3400-4600 m/s.[46] This is a more realistic claim as thegeology in this area has interspersed Mesaverde Sandstone units which is a main characteristicof the Lewis Shale. When evaluating the velocity of the proposed dike, the velocity yielded islower than the regular range of typical basaltic igneous rocks (5000-6000 m/s). However, thepossibility that the surrounding geology is altering the computed velocity should not be ruledout. Thus, this location can still be considered as a potential dike.

3.3.9 ConclusionsRefraction analysis of the main line hammer seismic survey yielded interesting results. To

begin, the average V2 P-wave velocity computed is 3076 m/s. This was interpreted to be LewisShale as the observable geology of the area contains Mesaverde Sandstone units interspersedwith a shaly unit and dike outcrop. While the velocity found is lower than that of the typicalLewis Shale range of 3400-4600 m/s, the geology of the area greatly influences the interpretation.The depth to the Lewis Shale interface was found to be about 3-5 meters along the survey line.When analyzing the point of interest of the earth model, Figure 3.6, one must take into accountthe velocity of this point in comparison to the surrounding velocities as well as the range ofvelocities typically expected for a basaltic dike. The depth to this point is 3.22 meters, withgreater depths to the first interface surrounding this location. This is the first suggestion of thepresence of a dike. The velocity computed at this point is 4111 m/s which is lower than thetypical range of basalt at 5000-6000 m/s. However, the velocity at the point of interest is mostlikely influenced by the surrounding subsurface geology.


Figure 3.4: Photograph of hammer seismic setup. The yellow box is the Geomatrix Geode Seismo-grpah. Takeout cables (yellow) attach to the geode, allowing up to 48 channels to record at once.The trigger sensor (black cable) connects the sledgehammer to the geode for synchronized shot andacquisition .

Figure 3.5: Example of refraction analysis. The red line depicts the shot location and is used as areference line to find ti of the refracted wave. The black line represents the head wave or V1. Thegreen line represents the refracted wave or V2.


Figure 3.6: Figure depicting the elevation and shale profile (top) and the calculated depth to thefirst layer, or difference between topography and the first interface (bottom), for the Main Line. Thefirst layer is interpreted to be Lewis Shale. Note that this survey line was taken off of the Main Lineand contains no GPS data. A reference model was created from the closest main line station. Dateof survey: May 23rd, 2017.


3.4 Magnetics

3.4.1 ExpectationsInitially, magnetics was used because the hope was that the dike would be imaged because

of its potential to have an anomalous magnetic susceptibility in comparison to the surroundingsedimentary rock. We hoped to be able to build a model of the vertical extent of the volcanicintrusions in an effort to constrain the regional model for gravity processing. Additionally, thehope was to gain insight into the dikes being possible fluid flow barriers in the underlying poroussediments.

3.4.2 Survey Maps

Figure 3.7: Magnetics Survey Line

3.4.3 Data AcquisitionThe cesium vapor magnetometer was used to perform a single magnetics survey across the

dikes in Chromo. To do so, the instrument was connected accordingly and worn by the surveyor.The data was then acquired by walking with the instrument between flags 1412 and 1140. Themagnetometer took recordings every 2 seconds but a mark was placed every 100 meters. Thesemarks were placed in the data to help determine flag locations during the processing stage.

3.4 Magnetics 97

3.4.4 RecommendationsThe magnetics data acquisition relied heavily on the location of the survey. Unfortunately, the

2017 field camp main line was restricted to a well-traveled road in Chromo that was bordered byman-made structures such as fences, cattle guards, culverts, and underground electrical wires.Ideally, the survey would be conducted in an area without any of these man-made structures asthey do interfere with the measurement of magnetic susceptibilities. It is recommended that themagnetics survey should be conducted across the two dikes in an environment away from themain road. Information gained from a survey not along the main line may provide greater insightinto the subsurface around the dike which is more valuable than the data that was acquired.

3.4.5 ConclusionsProcessing was completed on the magnetic data along the main line using MATLAB. Top diur-

nal measurements were post processed by filtering out anomalous points corresponding to cattleguards, buried electrical wires, and other large spikes due to metal objects and general artificialnoise sources. Magnetics is typically useful in detecting areas of high magnetic susceptibility.Unfortunately, the survey conducted over the main line didn’t yield any useful results as themagnetic susceptibility appeared to be relatively uniform across the line. No conclusions aboutthe local geology could be made from the magnetics data unfortunately. Errors that influencedthe data included metallic objects along the road: fences, cattle grates, power lines, cars, etc.Even when these errors were removed the data still didn’t provide any value as the uniformityonly became emphasized by noise filtering.


3.5 DC Resistivity and Self Potential

3.5.1 Objectives

The objective of the overall survey is to characterize the hydrology and lithology of the regionsurrounding the main survey line near Chromo, Colorado. We suspect that the dikes in thisregion are redirecting fluid flow, causing the eastern side of the dikes to have plentiful wells withhot, flowing water, while the wells on the western side of the dikes had cold water with very slowflow rates. The DC resistivity and self-potential surveys attempt to characterize the conductivityand telluric currents at this site to describe the fluid flow around the dikes. The conductivityprofile will enable an improved understanding of the layering of the geologic formation, aidingin the knowledge of the local lithology. Then, we apply self-potential data to model the telluriccurrents, which relate to the fluid flow in the subsurface. Fluid flow findings are significantbecause they improve the understanding of the local hydrology.

3.5.2 Expectations

3.5.2.1 DC Resistivity

We anticipate that the DC Resistivity survey will observe a conductivity anomaly around thedike in the DC resistivity survey inversion results. This anomaly derives from the aforementionedtelluric currents that give rise to an augmented apparent conductivity. Additionally, due to therelatively shallow dip of the local geology, we anticipate that the data will show near horizontallayering in the subsurface. This would produce a vertically stratified conductivity array withminimal horizontal variance, barring the relative conductive anomaly stemming from the volcanicdike.

3.5.2.2 Self Potential

SP characterizes the fluid flow in an area, and because dikes tend to interrupt the fluid flowit is expected that the SP data would show a spike around local dikes. It is likely that this spikewill occur on the eastern side of this dike because the behavior of the wells on either side of thedike.

3.5.3 Survey Maps


The segment of the main line we surveyed spread across flags 1200 and 1712. The surveyitself used 20 meter spacing between electrodes, which included every other flag, specificallythe even numbered flags. At each of these locations Wenner arrays were used with site specificDipole-Dipole arrays as necessary. The Dipole-Dipole arrays occurred at flag 1392 on the mainline (at the dike) and along line S2 at the student site.


For the main line and the student site, SP measurements were taken with 20 meter spacing.The SP survey started at flag 1000 and continued for the following sections: 1000-1102, 1152-1200, 1370-1420, 1760-1852.


Figure 3.8: Map of the DC Resistivity surveys conducted along the main line in Chromo, CO.

Figure 3.9: Map of the SP surveys conducted along the main line in Chromo, CO.


3.5.4 Data Acquisition3.5.4.1 DC Resistivity3.5.4.1.1 Survey Parameters

The DC Resistivity survey began by planting the electrodes in the ground at the respectivesurvey flags with electrodes placed approximately 2/3 of the way into the ground. Cables wererun alongside and attached to each electrode, with the cables themselves ultimately attaching tothe ABEM. The ABEM lies in the center of the survey, between electrodes 32 and 33.

The following selections were made within the ABEM Terrameter once the device was poweredand turned on:

1. Select Lund Imaging System2. Select Resistivity Mode (other modes include SP and IP)3. Enter Nameline:filename4. Enter Smallest Electrode Spacing: 20 meters5. Enter Powerline Frequency: 60 Hz (for United States measurements)6. Enter Protocol

• For the Wenner array we used WEN64XL• For the Dipole-Dipole array we used DDP464XL

7. Set the midpoint (numbering is arbitrary provided incrementation)8. Set current to 200 mA9. Set acquisition time to 0.5 seconds

10. Set acquisition delay to 0.3 seconds11. Set cycle time to 3.8 seconds12. Skip internal errors (collect data, retroactively assess)13. Set minimum stacks to 214. Set maximum stacks to 515. Set error limit to 5%

Subsequently the ABEM system will run an electrode check. At this time the ABEM systemwill output any failed electrodes. Potential electrode failures derive from improper coupling or im-proper electrical connections. To improve coupling, pour saltwater on the problematic electrode.To resolve errors derived from connection errors, verify proper setup and the integrity of thecables and electrodes. A rule of thumb for trouble shooting is that multiple adjacent electrodefailures tend to derive from a connection error, while isolated electrode failures tend to resultfrom coupling issues.

Once a sufficient number of electrodes test positively, the survey may begin. The WEN64XLprotocol is written such that a roll along may be conducted. Once the ABEM completes all testsregarding an electrodes as well as the desired survey, the electrode may be moved to the newsurvey site for the next measurement.

3.5.4.2 Self Potential3.5.4.2.1 Survey Parameters

SP is a very portable method. To start the survey, choose the first flag location and take a tipto tip measurement of the electrodes. Then bury one of the electrodes, the reference electrode,about 3/4 of its length at the flag and connect it to a spool of wire. Plug the other electrode,the roving electrode, into the green port of the voltmeter and set the voltmeter to voltage (mV).Next, take the roving electrode, voltmeter, and spool of wire to the first flag, 20 meters away,and dig three holes using the rock hammer (just a few inches to get better ground contactand moisture in the soil). Take three measurements in these holes by placing the tip of theroving electrode into the soil and holding it there until the voltage value being shown by thevoltmeter appears to be stable. Each of these measurements, along with the flag number andtime of measurement, should be recorded in the field notebooks. This process continues untilthe desired line is completed. After completing the measurements with the roving electrode,return to the buried reference electrode and take another set of tip to tip measurements. Thisis a very important step as these tip to tip measurements are the only way to account for the


instrument drift when processing the data. Once the loop has been closed out with the final tipto tip measurement, reel up the spool of wire and proceed to the next line.

3.5.5 Processing3.5.5.1 DC Resistivity

Data processing began with an initial period of quality control through Geotomo Software’sRes2dInv inversion program. This inversion program produces a 3 iteration computed apparentresistivity pseudo-section. This inversion focuses on identifying and removing faulty electrodesprior to implementing further inversion that produce depth inversions and incorporate topo-graphic effects.

To complete the rest of the DC resistivity processing, use DCIP2D, a code written by UBCGeophysical Inversion Facility. DCIP2D is a non-linear inversion program that allows its usersto calculate conductivity of the subsurface based on the measured voltage from the survey. Theedited ABEM Terrameter data files are manipulated and converted into a DCIP2D data file usinga conversion code.

There are many parameters that contribute to the quality of DCIP2D’s inversion process. Themain parameters examined for these datasets are topography, mesh size, target data misfit, chifactor, and reference model. The best way to find the correct parameters is to execute multipleinversion runs through the same datasets and find the parameters that best fit the model.

To calculate topography use the survey line DGPS coordinates. To find mesh size use thedefault mesh settings, and tweak them to allow for better resolution. Most of the datasets use3-5 meter mesh size horizontally, and 5 meter mesh size vertically. Target data misfit is foundusing the default settings, an inversion uncertainty within five percent of each measurement.

The next calculation is the chi factor, the ratio between the target data misfit and the numberof data. This was kept default (a factor of one) for almost all the surveys except S0 on the studentsite. For S0, the chi factor was changed to two in order to lessen the weight of the data. Thisdecision was made because the data was initially over-fit due to a bad electrode that couldn’tbe removed earlier in processing, creating non-geological resistive anomalies. Changing the chifactor to two produced a superior final product.

The last thing added to the inversion parameters was a reference model. We determinedthe reference model based on resitivity data gathered by previous School of Mines geophysicsstudents. Using this data, the chosen resitivity value for the reference model was 50 Ohm-m,the resitivity of Mancos Shale, the primary geologic bed on Reservoir Hill. The reference modelis one of the most important parameters because it gives the inversion a value for the geologicbackground, instead of creating a default one based on solely the ABEM measured data. We alsocreated a model based off of a geologic cross section. This model yielded similar results to thatof the initial reference model, further supporting our inversion results.

3.5.5.2 Self PotentialSP crews collected raw data in their field notebooks each day and then compiled their findings

into an Excel spreadsheet according to flag number. The data that was collected includes theflag number, voltage, and time of recording. As the instruments used tend to drift and shiftthroughout the day, it is important to apply a drift correction as well as a shift correction to theExcel data points. These corrections can be made using the tip to tip measurements that weretaken in the field. Shift corrections include shifting all of the data to a common point. In thiscase the common point was flag 1000 on the main line and flag 0001 on both of the studentlines, S0 and S1. Self potential is a relative measurement where the trend of the data is thefocus rather than specific values. The drift and shift correction to the data was applied using thefollowing equation:

Drift =

(FinalT ip− InitialT ipTotalSurveyT ime

× TimeSinceInitialT ip)− InitialT ip (3.1)


After applying corrections, find the average difference between the survey points and subtractthis number from the dataset to adequately correlate the data. From here, the general trend ofthe survey can be analyzed.

3.5.6 Errors and Uncertainties3.5.6.1 DC Resistivity3.5.6.1.1 Processing Errors

Processing introduces potential errors within the interpretation of the DC data. These po-tential errors primarily focus on the inversion process. The inversion procedure utilizes certainbackground parameters (such as a 0 voltage at an infinite distance, a boundary condition im-plemented in the DCINV2D code) which may not always be correct for the specific survey site.This produces potential artifacts within the inversion. Likewise, the quality control process andthe final inversion utilized two different inversion programs. This improves the quality of theresult by reducing the potential for inversion artifacts to skew the procedure. Additionally, theresults from this survey were compared against other localized surveys that occurred during theduration of this survey along with results from prior surveys.

3.5.6.1.2 Field ErrorsThe DC resistivity survey uses the transmission of current between metal electrodes through

the subsurface. This process introduces systematic errors when looking at the coupling of theelectrodes, polarization effects, variability in weather conditions, and processing artifacts.

The most significant error on the main line for the DC Resistivity survey was the non-linearityof the survey. The DC survey followed CR 542, a very curvy road. When conducting DC Resistivitysurveys it is important to keep the line as straight as possible as curvature causes a great dealof discrepancy. Because of this, the data collected along the main line is not very dependable,and must be broken into smaller sections to be adequately analyzed.

During the survey we attempted to minimize errors by using salt water to increase couplingand implementing low frequency alternating current to reduce the impact of polarization effects.The pouring of saltwater over electrodes increases the local conductivity, which improves theelectrical coupling between the metal electrode and the earth. The usage of low frequency al-ternating current instead of true direct current reduces the impact of polarization effects. Thepolarization of the electrodes leads to oxidation-reduction reactions that introduce error into thesurvey. Another potential source of error present at the time of the survey was the weather. Dueto variances in the subsurface saturation across the duration of the total survey we cannot com-pare exact values of resistivity at provided locations. Instead, this paper focuses on the presenceof relative changes within each survey line.

3.5.6.2 Self Potential3.5.6.2.1 Processing Errors

One common error that occurs when conducting SP surveys is that the surveyor does notwrite the data collection time down. Keeping time is vital when accounting for instrument drift,so if time is not recorded, the drift corrected data is less accurate.

3.5.6.2.2 Field ErrorsWhen using self potential there are many sources of errors and noise. This method works

the best if the ground is moist, because soil moisture establishes sound contact between theelectrodes and the earth. However, if the amount of water in the ground has changed due to rain,snow, or other events that give the ground uncharacteristic moisture, the SP survey may yieldfalsely high voltage readings. This occurs because the survey will report the voltage of the waterin the ground rather than the ground itself.

It is also important not to use self potential around any methods that inject currents in theearth, such as DC Resistivity, EM, or MT. This is because the electrodes used in self potential


will measure the voltage according to the currents being produced by the other methods ratherthan the voltage according to the actual earth itself.

When using the voltmeter there is an element of human error because it does not read out asteady value for the voltage; the values fluctuate so it is up to the surveyor to choose a reliablevalue for that particular location.

Finally, having other sources of current or resistive material in the area uncharacteristicof the earth causes noise in the SP data. In this case there were buried cables and possibleirrigation systems along the main line that were carrying current through the earth while SPmeasurements were being taken, causing large outliers in the data.

3.5.7 Recommendations3.5.7.1 DC Resistivity3.5.7.1.1 Processing Recommendations

Data processing of these surveys could provide valuable data to improve the interpretationof the subsurface geology. The inversion process could begin within a similar method to thosepreviously described. The UBC inversion code provides reasonable interpretations that do notimpose the same restrictions as linear inversion programs. However, the usage of additionalinversion programs could improve the final result by reducing the possibility of inversion artifactswithin the inverted model.

3.5.7.1.2 Field RecommendationsFrom this interpretation of the DC resistivity survey, we recommend conducting further local

surveys that focus on characterizing the observed conductivity anomaly at the student site. Inparticular, a survey that could delineate between ancient stream beds, anthropogenic sources,and faulting would provide useful information for further characterizing the subsurface anomaly.These surveys may include well boring to obtain various depth-dependent data, seismic surveyswith a deeper depth of investigation than the current shallow-depth seismic survey, and/orfurther Time-Domain Electromagnetic surveys that could further constrain the inversion of theDC resistivity data.

In the event of implementing further DC resistivity surveys at the same site, a tighter surveyspacing near the suspected conductive anomaly would likely produce interesting results. Futuresurveys may also implement a Schlumberger array geometry, which could potentially improvethe horizontal resolution of the data.

The main line DC resistivity survey did not provide significant information with regards tothe suspected subsurface fluid flow surrounding the volcanic dike. The DC survey across thedike did not provide information that extended beyond confirming geologic observations. Like-wise, it may be beneficial to utilize other survey methods that could provide a greater depth ofinvestigation (such as a seismic survey) or provide three dimensional characterization of the sub-surface. Likewise, further DC resistivity surveys across the dike could implement an array forthree-dimensional inversion that could provide more information than the aforementioned 2.5dimensional inversion.

3.5.7.2 Self Potential3.5.7.2.1 Processing Recommendations

In order to maximize the accuracy of the SP data, it is important to emphasize the need torecord the time of the recordings and to take tip to tip measurements as that information is vitalto correcting the data. Microsoft Excel seems to be a very useful program for processing SP data.Be sure to use the equations to relate the data rather than plugging in concrete values so thatsomeone else can look at the data and understand processing procedures.

3.5.7.2.2 Field RecommendationsIn the future, in order to maximize the usefulness of SP, choose a location that is not espe-

cially rocky, where the soil is accessible and has a little bit of moisture in it to ensure that the


readings are reliable. It would also be good to conduct these surveys away from other methodsand away from environmental factors that may be affecting the current in the subsurface, suchas the buried wires.

As far as survey locations, it would be good to collect SP data along the entirety of the mainline in Chromo, especially around the dike in order to try to monitor the fluid flow there as thatis an area of significant interest concerning the behavior of the water.

3.5.8 Results


Figure 3.10: Inverted DC Resistivity data from flags 1228-1360 along the main line.

Figure 3.11: Inverted DC Resistivity data across the main dike, flags 1380-1400, along the mainline.



Figure 3.12: SP surveys conducted along the main line.

3.5.9 Interpretation3.5.9.1 DC Resistivity

Figure 3.11 incorporates the suspected volcanic dike addressed by the local geology due toan outcrop of igneous rocks along Country Road 542. The suspected dike appears near thelocation of flag 1300 on the horizontal axis. At this location, an outcrop appears containingigneous rocks with an adjacent burn zone. The presence of the burn zone indicates the intrusivenature of the observed volcanics, which indicates the relatively young nature of the volcanicrocks compared to the background country rock. This volcanic intrusion appears to be a dikedue to the relatively short horizontal spread along the outcrop. The DC anomaly in Figure 3.11also appears to indicate a relatively narrow width. The DC Resistivity survey conducted alongthis line did not lie in a straight line, rather the final inversion is a compilation of two separatestraight lines running in opposite directions and coming together at an angle. DC data is mostuseful when the survey is conducted in a continuous, straight line, and is not very useful at allwhen the line is curved, which means that, as far as characterizing the fluid flow in this region,this particular DC survey does not give an accurate picture of the area.

3.5.9.2 Self PotentialThe SP data on the main line is pretty inconclusive due to the lack of data available, as

seen in Figure 3.12. While in the field only a few segments of the main line were available fordata collection so a general trend for the entire line is not clear. There were also buried cablesthroughout the main line, specifically at the two major outliers in the data at flags 1168 and1840. These are opposite in sign but comparable in value, however, due to the orientation of thecurrent when taking the measurements. Furthermore the ground along the main line was quiterocky and dry, causing a lot of problems with getting good contact between the electrodes andthe ground. Because of these errors not much can be interpreted from this data.

3.5.10 ConclusionsFigure 3.11 indicates a resistive anomaly near 1300 meters on the horizontal axis. This

resistive anomaly could relate to the volcanic rocks observed at the surface along County Road


542 at the same location. The anomaly concurs with the geologic interpretation of the outcropas a volcanic dike due to the relatively resistive nature of volcanic rocks. The fluid flow of thisarea could not be accurately measured due to the curvature of the survey lines. We predict thatthe volcanic dike is preventing the flow of water in the area from moving westward. Perhaps abetter way to characterize the behavior of the fluid flow here would be to conduct several, smallerscale DC Resistivity and SP surveys and then stitch that data together so that the lines could becontinuous and straight which would result in better data.


3.6 Magnetotellurics3.6.1 Objectives

The 2017 magnetotelluric team possessed similar objectives of prior MT field-session teamstasked with the duplicate goal of identifying geological resistivity contrasts along the main line.Since typical subsurface pathways of fluid flow contain charged particle-infused liquid, whichpossess relatively higher conductivity compared to neighboring lithology, successful MT surveyexecution was key to the overall analysis and identification of subsurface contact layering. Enjoy-ing supplemental awareness of intrusions in the targeted boundaries would be fundamental inquantifying the geothermal basis for the Mother spring and other significant geological anomaliesthat support the underlying goal of geothermal-source discrimination by the following steps:• Derive electrical impedance value from electric and magnetic field-data• Convert electrical impedance to apparent resistivity.• Use apparent resistivity values to promote the identification and location depth of subsur-

face contacts within stratigraphy column.• Endorse the anticipated location of geothermal source for the Mother Spring in the town of

Pagosa Springs, Colorado.

3.6.2 ExpectationsThe MT sites were placed in an orientation such that a cross section over the dike could be

obtained from the data. It was expected that Sites 2-4 on the West side of the dike will havesimilar resistivity values and depths. Likewise, the sites to the East of the dike (Sites 1, 5-6, and8) were expected to be have the same characteristics. From the geological interpretation provided,the depth to basement is around 1,600 meters. Knowing this, it is expected that the 16 and 4 Hzbands in the data would show the basement (calculated using the skin depth equation). Throughnoise removal and inversion, these expectations can be met.

3.6.3 Survey Maps3.6.4 Data Acquisition

At each MT survey site, 2 pairs of electrodes are aligned in an orthogonal, typically North-South and East-West, arrangement in addition to ortho-normally aligned magnetometers. Elec-trode spacing was established at 50 meters from center, which is the overlay-point, adjacent tothe ADU 07e. They are then attached to a grounding electrode in the center of the configuration.Magnetometers were placed in separate quadrants surrounding the grounding electrode, in thecenter, at least 5 meters apart.

For the E-field measurements, the electrodes will provide low-noise, low-resistive couplingwith the Earth. In contrast to 1-Dimensional MT, where σ is solely dependent on ~z orientation,its noted that in 2-Dimensional MT where σ = σ(x, y), conductivity changes horizontally as afunction of depth; The subsidiary horizontal component is commonly known as strike. This issignificant when examining the anisotropic/isotropic nature of the subsurface, since at eachpoint σ is possibly dependent on vectors of current flow, meaning, if σ is dependent on directionan anisotropic subsurface has been encountered.

Based on the attenuated nature of the EM plane wave propagation into the subsurface, lowerfrequencies penetrate more deeply, whereby apparent resistivity, ρa varies with frequency if con-ductivity changes with depth. Apparent resistivity can be tabulated (by approximation) for anycombination of these horizontal layering, whether anisotropic or isotropic, as long as E and B-field alignment is perpendicular.

Since the induction coils of the magnetometers are motion-sensitive to wind and other vibra-tional sources, the coils relating to the x-axis and y-axis were buried in 12 to 18 -inch trenchesand leveled within 1◦


In order to run the survey, all electrodes and magnetometers are attached to the ADU 07eunit as well as a computer. In the computer, Firefox needs to be opened and the IP address typedinto the address bar. The correct ADU number must be selected. The gain must be set to auto.A test must be run under ”Self Test Configuration” to determine the appropriate gain. Once thetests are finished, the survey can begin. onSURVEY PARAMETERS• 50-meter spacing from center point (True North)• Sampling frequency: 1024 Hz (2 sites)• Sampling frequency: 256 Hz (all sites)• Channel gain : varied per station• Record duration: 24 hours

3.6.5 ProcessingIn order to meet the expectations previously mentioned, the data needed to be turned into a

format suitable for inversion. This was done through the use of the Mapros Software. Data setsfor each location were imported into the program. The first task was to remove the noise. Thetime series was filtered into 256, 64, 16, 4, and 1 Hz frequency bands. Then, the time series foreach band was looked over by a member of the MT processing crew. Any noise seen in the timeseries was marked. The team identified noise as any irregularities in the data. The irregularitiesgenerally consisted of spikes, extremely choppy data (the measurements would instantly increaseby over 200 mV then move back down), changes in general sinusoidal trends.

Once this was done, the data was processed using the setting ”Remove all Marked Sections”.The result was an amplitude spectrum and phase spectrum. If the amplitude spectrum wasgenerally smooth for all bands, the data was exported as an ASCII file. If the data was notsmooth enough, the parzen radius and sampling length were adjusted or the processor revisitedthe time series to mark more data.

3.6.5.1 ipi2win.MTThe exported ASCII file was in the format as shown in the table below, the file was then

convert to an appropriate format for the ipi2win.MT program to read. This was done by takingthe frequency, the apparent resistivity, the apparent resistivity variance, the phase, and thephase variance and putting them into an Excel spreadsheet in that order. The first column wasthen turned in the square root of the period (square root of the inverse) and the third was turnedinto percent error. The file was then saved as a .mt file so it could be opened by ipi2win.MT.Once opened in ipi2win.MT the percent error had to be added by hand to the file for the apparentresistivity. Afterwards, the suspected number of layers were added and a line of best fit and acorresponding model were found for the graphs.

3.6.5.2 IX1Dv3The IX1Dv3 software uses the same columns as the previous software. The frequency column

could remain in terms of frequency and the resistivity variance was changed to percent error(resistivity variance divided by resistivity). The team chose the best points to ensure there is onlyone point was given per frequency as was d ipi2win.MT input file. In the IX1Dv3 software, a new”MT Sounding” was created and all values were input into the program. The team created theinversion model by adjusting the layer lines so the apparent resistivity line matches the data,then the model was exported. Once the inversion models were correct, the model was exportedand a resistivity logarithmic model was created.


In the processing portion of field camp, the way of cleaning out the noise from the data wasdone manually, in which the team found anomalies throughout the data at various frequencies.The errors were expunged in the software Mapros, these were highlighted and then ignored when


Table 3.1: Description of columns in exported ASCII files from ipi2win.MT

Column InformationA MT SiteB BandC Frequency (Hz)D Zxx,real

E Zxx,imaginary

F Zxx varianceG Zxy,real

H Zxy,imaginary

I Zxx varianceJ Zyx,real

K Zyx,imaginary

L Zyx varianceM Zyy,real

N Zyy,imaginary

O Zyy varianceP Tx,realQ Tx,imaginary

R Ty,realS Ty,imaginary

T ρxyU φxyV ρxy variancew φxy variancex ρyxY φyxZ ρyx variance

AA φyx variance

processing. These cleansing happened at the frequencies of 256, 64, 16, 4, and 1 Hz, in thisdata sets were choppy and anomalies were taken out at the best judgment of the team, the issueis that these files are so extensive that there was a fine line between noise and data, thus notall of the noise could be filtered from the survey. The noise reduced data sets were exportedfrom Mapros in the form of an ASCII file and reduced in Microsoft Excel to only the data requiredfor an inversion. The data then had to be smoothed and points were chosen by hand to bestshowcase the inversion. This was a judgment call made by the processing team and again wassubject to human error, therefore the best point was not chosen in every case which leads toanother uncertainty within the final product. From this the file was then saved as a .mt file andthen inverted in ipi2win.MT. The percent error of these files had to be adjusted by hand and theresulting inversion of the MT data shows the various processing errors with the large error barsfound within the majority of the phase graphs along with the peaks and valleys found within theapparent resistivity curves.

3.6.6.2 Field ErrorsIn the field most of the sites were along the forest road with the focus being along the dike. An

issue with this is that there are people who drive on these roads and there cars and movementcreate serious issues with the quality of data acquisition. In the survey setup the magnetometers


must be perfectly aligned with North-South, and East-West, these can easily be knocked off ofalignment causing issues within the data. The magnetometer must be level, this is difficult todo with a hole as dirt is placed back in to bury the magnetometer the instrument may be movedand knocked off of alignment, thus giving rise to issues within data acquisition. Issues withacquisition of data results in issues that make processing more difficult, and will leave moreroom for uncertainty in the results and later analysis.


In order to smooth the processing portion, another noise removal program ProcMT could beutilized. It might be useful to understand how the program works and it could eventually makethe noise removal process easier. One recommendation for inversion would be to choose a singleinversion program. ipi2win.MT seemed to work better with the actual inversion. IX1Dv3 madeit easy to export the inverted model into a format that is easy to work with in excel. Choos-ing the right inversion and processing programs will lead to faster inversion and more time forintegration.

3.6.7.2 Field RecommendationsIn observation of data from this year and previous years the largest recommendation for each

MT survey would be to take a moment to observe and consider each survey site before charginginto the set-up. MT is extremely sensitive to the topography and surrounding environment, alittle bit of exploitative quality control in the field could provide a much cleaner data set. Thisquality control could include taking extra time to ensure Ex/Ey electrodes and Hx/Hy coils arealigned with proper directions, making sure survey location is generally flat, checking that thereare no power lines within a 1km radius, and even looking to make sure there are no nearbyroads. Another recommendation for next years MT acquisition crew would be that everyday inthe field Mapros should be used to try and remove noise from the acquired data. If the in fieldnoise removal was well documented and efficiently executed, then the processing portion of MTback in Golden would be finished much faster and the actual interpretation and inversion couldbe more meticulously done. This would allow for more time to really understand the methodand programs that encompass the MT method and give more freedom to make a more in depthinvestigation into the data acquired.

3.6.8 InterpretationWith the help of geology, the MT crew was able to interpret the data. Figure 3.14, Figure 3.15,

and Figure 3.16 all show the apparent resistivity log which describe the resistivity as a func-tion of depth. This log correlates with the geologic stratigraphy, which is demonstrated by thestratigraphic column next to it. As the data shows the log matches the contacts between differentgeologies. This is useful because it helps in interpreting and correlating seismic and gravity data.

3.6.9 ConclusionsThe MT surveys proved useful in the determination of how the geology of Pagosa Springs

affects the geothermal activity of the area. This is because of the fact that there were a coupleof surveys that produced great data showing the geologic contacts, which was confirmed by andhelped confirm other geophysical methods. The team then went on to combine this data with therest of the methods and help in the overall interpretations.


Figure 3.13: All 8 sites are located in remote areas as to avoid cultural noise which will interferewith the measurements. The overall site location geometry was chosen in order to get a crosssection across the dike indicated by the black line. Site 8 is located on a dike in order to obtaininformation about the dike itself.


Figure 3.14: Site 4 Apparent Resistivity Curve with a Stratigraphic Column next to it.



3.7 Gravity3.7.1 Objectives

1. Collect high quality gravity data with the CG-5 Autograv Gravity Meter.2. Perform data corrections (instrument/tidal drift, latitude, free air, simple Bouguer slab,

complete Bouguer terrain, and geometry corrections) on the collected data.3. Generate models.4. Integrate past field camp’s data with current data to produce a larger scale regional model.5. Provide useful recommendations for future field camps based on experience in the field and

processing.

3.7.2 ExpectationsAlong the main line the gravity survey was expected to assist in understanding the two dikes

that cross through the road. A large density contrast was expected due to the higher densityof the dike compared to the surrounding shales. The small dike at flag 1180 is relatively small,and thus is not expected to be seen in the gravity data. However, the large dike at flag 1398 wasexpected to be identified using the data. Furthermore, there was the expectation that water flowaround the dikes might be visible in the data, Figure 4.18.

3.7.3 Survey Maps

Figure 3.17: Combined Gravity Survey Line. Includes the 2017, 2015, and 2014 survey lines.

3.7.4 Data AcquisitionTwo Autograv CG-5 Gravity Meters were used to complete the gravity survey along the main

line. To begin data acquisition each day a gravity reading was first taken at a base station.Each CG-5 was then taken back to this base station every two-three hours to help account forinstrument and tidal drift. This information allows for the assumption that the rate of change

3.7 Gravity 115

due to drift in this time frame is linear, which can be easily corrected in the processing stage.The base station was located near the end of County Road 542. A leapfrog method was usedwith the two instruments which resulted in a reading being taken every 40m. A minimum ofthree 20 second readings were taken per station, with more taken if the standard deviationsexceeded 0.02mGal or the measurements seemed inaccurate due to outside errors such as windnoise, nearby crews, etc.

Proper field setup consisted of placing the gravity meter on the leveling stand to within ±2arcsec in X and Y tilt directions. Once the gravity meter was leveled, the gravity measurementcould be taken and the CG-5 would display all necessary information to be read off and savedon the instrument itself. Each measurement was stored on the instrument, and the operatoralso recorded values for station number, time (HH:MM:SS), value (mGal), and standard deviation(mGal).

3.7.5 Processing & IntegrationEach day, after field acquisition, the gravity data was taken off the CG-5 and quality checked

in Microsoft Office Excel. The best readings for each flag were noted for later use. These valueswere checked again and subjected to a number of corrections, the first being drift corrections.Correcting for drift requires correcting individual loops created by repeated measurements at abase station. The corrections were be done with a simple line equation:

gcorrected = gi − (g1 − g0t1 − t0

)(ti − t0)− g0 (3.2)

where gcorrected is the drift corrected gravity value, gi is the raw measurement needing correction,ti is the time of the raw measurement, g0 is the first base station measurement in the loop, g1is the final base station measurement in the loop, and t0 and t1 are the times those measure-ment were taken respectively. This equation was applied to each raw measurement in MicrosoftOffice Excel. Next, loops with differing base stations were shifted vertically to ensure that alloverlapping stations had the same value. The Chromo line also incorporated gravity data fromthe 2014 and 2015 field sessions, which required that those values were shifted to match the2017 measurements.

The shape of the Chromo line required that points be extrapolated to a straight line. This linehas an orientation orthogonal to the strike of the geology which was approximately 170◦. Usingthis, line values for elevation were acquired from a local digital elevation model (DEM) usingQGIS software. These values were then used for the latitude, free air, simple Bouguer correction,and complete Bouguer corrections. Geosoft’s geophysical exploration software, Oasis Montaj 8.5,was used to do this.

From the corrected data, a GYM-SYS model was created. The software allowed for the integra-tion of past year’s model images, which were used to reproduce the 2014/2015 model alongsidethe new model for the 2017 data. Theoretical models, created using the software, produced agravity response. This model was displayed alongside the measured gravity. By adjusting themodel to match, the theoretical response was made to match the real response Figure 3.18.

Preliminary models were made using the basic depth migrated seismic images, past models,geologic cross subsections, and well data. Because the orientation of the seismic line differedfrom the extrapolated line, the image could not be fully integrated with the Chromo main linedata. However, observation of depth to basement along with well data provided information aboutthe structure of the subsurface geology.

Integration of borehole information was used when creating the main line gravity models.Using core samples from the Pagosa Springs area, density values and thicknesses were measuredby the geology processing team. The core sample density and thicknesses were also averaged withvalues from previous field camps.

For the main line, the deep seismic processing and magnetotelluric teams also provided in-formation about layer depth, specifically the depth to the basement layer. Magnetotellurics and


geology also interpreted depths to specific layers in the subsurface. These methods were all usedto create an accurate model of the gravity response measured in the field along the main line.

Figure 3.18: Gravity Model based on seismic depth to basement.

Corrections Chromo Main LineDrift Correction XBase Station Tie-in Correction XLatitude Correction XFree Air Correction XSimple Bouguer Slab Correction XComplete Bouguer / Terrain Correction X

Table 3.2: Main Line Gravity Corrections: X indicates a completed correction.


The only processing errors that occurred during gravity post-processing were the use of ter-rain files for each region. The terrain files used had an uncertainty of 10 - 25m which introducedsome error into processing results. The modeling process also has some uncertainty in termsof the final result because the model itself is non-unique. The final model was fine tuned andaltered many times in order to fit the data from other sections.

3.7 Gravity 117

Formation Density (g/cc)Lewis Shale 2.4Mesaverde 2.57Mancos Shale 2.5Dakota Sandstone 2.34Morrison Formation 2.62Jurassic Sandstones 2.55Ancestral Rockies Sediment 2.5Low Density Zone 2.5Crystalline Basement 2.8

Table 3.3: Density Values of Geologic Layers: These values were used when modeling thegravity response of the main line and S0.

3.7.6.2 Field ErrorsThe main gravity survey line in Chromo moved along roads 542 and 359 of Archuleta County.

Unfortunately, these roads were not straight lines which introduced uncertainty into processingand interpretation. At some sections along the line, the road ran along the direction of strike inthe cross section; this created an overlap in the projection of the data points relative to the dip ofthe geology in the region. The overlap resulted in many data points being omitted as they wereunneeded in the projection. These omissions may have created errors in the values we used toconstruct the geological model as we could have chosen projection points that do not entirelywork for its location.

Uncertainty was also introduced at certain points along the line that were flagged in areasnot accessible by the CG-5. At these points, the CG-5 was placed as close as possible to the flagpoint with minimal change in elevation, but this introduced some uncertainty based on whereGPS measurements were previously taken.

There were multiple instances of both of the CG-5 instruments being mechanically shocked.This happened during transport of the instruments on the days when measurements were takenon line S0 because of the irregular terrain that the vehicles drove over. Normally after transport,the CG-5 should be left to rest so the internals of the instrument can normalize.

3.7.7 RecommendationsIt is the recommendation of the 2017 Geophysics Field Camp students that future classes

attempt to gather more permitting around the areas of interest. With more permitting, gravitywould ideally be able to create a straight gravity line that is perpendicular to geologic strike.There would also be a possibility for a 3D grid to be set up in order to employ 3D inversiontechniques. It would also be ideal if the flagging placement of the main line was done with CG-5placement in mind. Areas that are relatively flat, and a good distance away from fences andanthills are ideal when taking measurements. It is also highly recommended that the GPS pointsgathered are within a vertical accuracy of ±3cm.

3.7.8 Results3.7.9 Interpretation

Along the gravity survey’s main line there were no major anomalies displayed in the data. Alarge portion of the 2017 data was uninterpretable due to the geometry of the line and the strikedirection of local geology. The data trends upwards from left to right which agrees with previousfield camp years (2014-2015), see Figure 3.22. The upward trend likely relates to an anticlinalstructure causing the basement layer to be thrust upwards toward the surface.


Figure 3.19: Main Line Quality Control

3.7 Gravity 119

Figure 3.20: Main Line Processing

Figure 3.21: Legend for the following Gravity models for the main line.


Figure 3.22: Gravity Model based on matching the data. This model puts data lower than seismic,but this is likely do to the orientation of the gravity line orthogonal to strike. The seismic data islikely observing apparent dip and not true dip. Thus their depth to basement is slightly shallower.


3.8 Deep Seismic3.8.1 Objectives

The main geological question to answer is whether the dikes in Chromo control the waterquality in the area. As previously explained, seismic method is good at delineating local geologicstructures that are within its resolvable limits. We should be able to see reflections from litho-logical boundaries. Features like dikes or faults should also be interpretable on depth migratedsections. We can reduce uncertainty in the interpretation by integration with other methods:seismic data from previous years, geological knowledge of the area, well-log information andresults of other interpretations. Final interpretation should improve overall understanding ofgeological structures in the area and aid in building hydrogeological model for fluid flow.

3.8.2 ExpectationsBefore acquiring any data, we knew that there are two dikes in the area as well as a suspected

fault. Thanks to information from previous years, it is possible to create an initial geologicalmodel of the area. However, there is strong evidence that local geology is complex and for thatreason data extrapolation may not be the best choice. Deep seismic should aid in adding moredetail to what is already known. In particular, it can help with identifying geological boundaries,mapping faults and estimating depth to the basement. However, one thing to keep in mind isthe shape of the acquired 2 dimensional line. If there are 3 dimensional structures in the area,interpreting a 2 dimensional, crooked line is subject to many uncertainties.

Figure 3.23 is the preliminary cross section made for the 2017 seismic line before any pro-cessing has occurred. This figure shows the two outcropping dikes along the line as well as thesuspected fault. In addition, there are a number of wells that were located close to the line thathad information on the thickness of layers that have been plotted in their respective locations.

3.8.3 Maps3.8.4 Acquisition

Accounting for survey parameters is one of the most important steps in gathering seismicdata. It is important to understand the goals of the survey and optimize these parameters toobtain expected results. The parameters and configurations were carefully chosen throughoutthe survey in order to produce an accurate image of the geologic structures as well as the eventsin the subsurface.

3.8.4.1 Survey ParametersDue to constraints such as time and money, the 2017 seismic group decided to run a 2D

seismic line survey on county roads 542 and 359 in Chromo, Colorado. When acquiring a 2Dline, the ideal situation is to design a line as straight and flat as possible. The elevation for the2017 Seismic Line decreases going from West to East. Furthermore, the line geometry is alsocrooked because of road access. We account for these restrictions in data processing to achievesatisfactory results. Thanks to Dawson Geophysical, the seismic source used for the surveywas a 12 ft Vibroseis truck generating compressional waves (P-waves). The shot spacing was 10meters and the location coincided with the flags on the line. We acquired data along the entire12 kilometer seismic line (flags 1000 to 2200). In between each flag, there were 6 geophoneswith the first phone being 1 meter away from the flag and then 4 phones in between with 2meter spacing (two phones overlapping) and then another phone 1 meter away from the nextflag. Figure Figure 3.26 above is a visual representation of what the survey setup between flagslooked like. The survey group also decided to do a split spread survey with 120 channels activeon each side of the truck (Figure Figure 3.27). There were 60 shots that were skipped from flag1501 to flag 1560 because of issues with nearby homes. All parameters are summarized in TableTable 3.4.


Figure 3.23: Preliminary Cross Section of the 2017 Seismic Line.

3.8.4.2 Sweep ParametersIn order to acquire the best results from our source with the allotted time given, careful

inspection of the sweep parameters is important. Sweep parameters work on the principle ofuser specified band of frequencies into the Earth and and corresponding them with the recordeddata to define reflection events. These sweep parameters include the starting frequency, endingfrequency, type of sweep, sweep length, sweep time, and the number of sweeps recorded. Forthe 2017 field camp survey the parameters are as follows: starting frequency of 4Hz, endingfrequency of 140Hz, non-linear upsweep, 12 second sweep time, 2 - 4 sweeps at each point, and3 seconds of listening time. We chose these parameters in order to acquire the best data with theavailable time. The equation used to determine how long our total recording time is as follows:

TotalRecordingTime = #OfSweeps ∗ (SweepLength + ListeningTime) + MoveUpTime. (3.3)

Based on the chosen parameters, the total time needed to finish survey was approximately30 hours. During production, the number of sweeps changed multiple times due to concernsabout time constraints resulting from early equipment and communication failures. We used


Figure 3.24: Location of main survey line.

various methods of reducing the total survey time such as reducing the number of total shots byonly taking shots every 20 meters rather than every 10. This was not optimal because the fold ofthe survey would have dropped from 120 to 60, thus decreasing the quality of the data collected.We determined that reducing the number of sweeps was the best course of action, even thoughit reduced the signal-noise ratio of the data. Therefore, we reduced the number of sweeps from4 sweeps to 2 sweeps starting from Station 1402 to Station 1622. However, we increased thesweep to 3 sweeps at Station 1623 until the end of the line, which was at Station 2200.

3.8.5 Processing3.8.5.1 Processing Procedures

Seismic processing is the process by which seismic data in the form of shot records is trans-formed into a detailed image of the subsurface. To do this it requires meticulous attention todetail, a thorough understanding of the survey and sweep parameters used during data acqui-sition, and an understanding of the physics associated with seismic surveying. The techniquesthat we employed to the seismic data reduces noise and corrects for non-planar subsurface ge-ometry. Figure Figure 3.28 gives a general overview of the techniques that we used to processthe 2017 field camp seismic data. Listed below are detailed descriptions of each step we took toprocess this data, and a review, that specifically describes the procedure the 2017 seismic groupused to process the data.

1. Shot Record QC:We individually analyzed each shot record from the dataset before the beginning of ourprocessing flow. Then, we inspected every shot record along the line to see if there wereany shots that had extremely low signal-to-noise ratios. Each record was either kept for


Figure 3.25: Location of wireless nodes along main survey line.

further processing or discarded completely depending on its relative signal to noise ratio.The only records that we discarded (or ”killed”) were records that were completely saturatedwith extraneous noise. There were some records that had extraneous noise in them, likelycaused by cars moving along the surface. These gathers were not killed though, becausewe could filter out their noise using other techniques. Figure Figure 3.29 shows someexamples of good and bad field records.

2. Crooked Line Geometry:Do to the vast changes in direction along this year’s seismic line, geometry needed to beheavily QCd in the processing. A 2D seismic section generally has the best resolution whenthe survey is conducted along a straight path. This is because the assumption of the CMPgeometry is straight, and that all of the CMPs are directly on the line. However, since thisyear’s line was crooked, many of the CMPs were located away from our line of interest. Thisintroduces immense error in our ability to accurately image the subsurface. In order toreduce this error, we corrected for this using a Crooked Line Geometry flow to interpolatethe CMPs away from the line back onto the line. Figure Figure 3.30 shows the location ofthe CMP gathers (white area) along the main surveying line(black line). Ideally the CMP’swould fall directly on the surveying line, but as Figure Figure 3.30 shows, that is not thecase for this survey line.

3. Elevation Statics:Correction that compensates for changes in elevation which in turn reduces the effectsof topography and low-velocity zones near the Earth’s surface. The correction involvesestablishing a datum on which to locate source and receiver, and adding or subtracting theincremental time.

4. NMO Velocity Correction and Analysis:We applied normal moveout to the data set to flatten our CMP gathers and prepare themto become stacked. The initial step to doing this is to implement an NMO velocity analy-sis. In order to flatten the CMP gathers, we must acquire a correct velocity for the NMOprocess. When the velocity is too low, the reflection is over-corrected and the reflections


Figure 3.26: Schematic illustration of acquisition geometry.

curve upwards. When the velocity is too high, the reflections are under corrected and theevents curve downwards. When the correct velocity is chosen, your reflection is horizon-tal. We acquired these velocities by picking points on a semblance plot in order to findthe areas with the highest energies to create your stacking velocity profiles for a particularcommon midpoint. Figure Figure 3.31 shows how a reflection event is corrected by pickingan accurate velocity vs. an inaccurate velocity.

5. Brute Stack:We created a brute stack (Figure Figure 3.32) for an initial representation of what theimage of the subsurface is likely to look like. We did this step for quality control purposesafter correcting for initial sources of error to ensure the subsurface model is feasible. Thevelocity model estimated from the NMO correction analysis is used as the input velocity forthis brute stack. This velocity function taken from several ”supergathers” along the lineis generated to analyze the entire subsurface geology. The brute stack helps analyze thevelocity models moving forward to come up with a more accurate depiction of subsurfacevelocities.

6. First Break Picking and Refraction Statics:First break picking deals with identifying the earliest arrival of energy propagating from thesource at the surface. Essentially, we identify first breaks in order to acquire the velocitiesof the near surface in order to apply static corrections. Furthermore, picking first breaksare important in marking arrival times for refracted waves, except for very near offset areasin which refractions disappear Therefore, first break pickings are necessary in order toperform refraction statics on the stacks. On the other hand, refraction statics correctionare corrections that compensates for the delays in seismic reflection and refraction timesthat are induced by low-velocity layers such as highly heterogeneous weathering layer ofmaterial near the surface of the Earth. Refraction statics remove a significant part of thetravel-time positioning of shots and receivers to a datum along vertical ray paths whichamounts to static time corrections in a way that is surface consistent. Figure Figure 3.33shows the original brute stack an initial refraction statics correction applied to it.

7. SWNA De-noise:Surface Wave Noise Attenuation (SWNA) De-noising aims to eliminate the noise from theshot gathers, mainly the ground roll and air blast. The noise has certain amplitudes,


Figure 3.27: Schematic illustration of acquisition geometry.

dipping angles (velocities), and frequencies. In order to pick them up from the data, wecreated a window on the seismic shot gathers. In the window, there are three parametersthat the program can use to differentiate the noise and the signal. These include thedipping angle, the frequency, and the amplitude. The program first targets the data withthe preset dipping angle and frequency and then compares the selected amplitude to thebackground amplitude. If the ratio exceeds the programed value by a certain amount, thesegment of data inside the window is labeled as noise and is eliminated by the program.Notice, the program compares the selected amplitude to the background amplitude, so thesize of the window is very important. If the window is too big, or too small, it will not pickthe correct background amplitude. The ideal size is one that can cover the segment of noisewhile leaving a decent amount of space for the background.Figure Figure 3.34 shows the result of denoising for a single shot. In this case, only am-plitude and dipping angle were used as parameters. Inputting more parameters does givebetter accuracy, but it is not necessarily needed. The real application depends on the sit-uation. Since the ground roll is a package of waves instead of a wave of a single velocity,different velocities has been tested to kill the noise. The leftmost image is the original shot,and the other three are denoised using velocities of 2000 m/s, 1500 m/s and 1200 m/sfrom left to right. The 2000 m/s seems to yield the best results - we can see the hyper-bolas coming out from the data. So this velocity is the one used for all the shots. FigureFigure 3.35 shows the stack after denoising procedure.Velocity Analysis:The velocity analysis is done in the SeisSpace software which includes an interactive displaybetween the velocity spectrum of the subsurface, the gather, and stacks obtained fromthe previous steps. It also displays the velocity function stacks that correspond to thesubsurface spectrum. The display allows users to observe the range of velocities thatcorrelate with the NMO gathers. The program then automatically updates the gather andstack displays once the user picks a certain velocity, with respect to time. Furthermore,the velocity spectrum is obtained by understanding how well a hyperbolic event matchesthe real events on the CMP gather. ”supergathers” are then created from the CMP gathers.A supergather is the process of forming average common midpoints and rebinning the datato enhance the signal to noise ratio. This is done by colleting adjacent CMP’s and addingthem together. Once this is complete, a semblance is calculated using the central gatherof the newly combined gathers. In this years survey, we created supergathers using nine


Table 3.4: 2D Seismic Line Survey Parameters

Parameter ValueSource Vibroseis Truck

Sweep Type Non-Linear UpsweepSurvey Line Length 12 km (Stations 1000 - 2200)

Orientation 135◦

Station Spacing 10 metersSweep Length 12 secondsRecord Length 3 secondsShot Spacing 10 meters

Fold 120 Full FoldMinimum Frequency 4 HzMaximum Frequency 140 Hz

No. Of Sweeps 4 sweeps (Start - Station 1401), 2 Sweeps (Station 1402 - 1622),3 sweeps (Stations 1623 - 2200)

adjacent CMP gathers. Figure 3.36 displays the original semblance plot that was used topick the NMO velocities.Residual Statics:The NMO correction attempts to find a hyperbola that best fits the seismic data. However,in real cases, there are still bumps or discontinuities that may appear on the correctedCMP gather. Before forming the super-gathers, it is better to make the CMP gathers flatand smooth, which leads to the residual statics. The procedure to compute residual staticsbasically creates multiple horizontal gates, and compares each trace to find the position ofthe largest correlation, by slightly moving them, and then modifies the positions of sourcesand receivers. By doing this, the move-outs of the CMP gathers become smoother andflatter, thus enhancing the signal-noise ratio most efficiently after stacking. Figure 3.37shows the result after residual statics correction.Time Variant Spectrum Whitening:Earth attenuates high frequency waves much faster than low frequency waves. The loss ofenergy in high frequencies leads to an unbalanced and irregular distribution in the ampli-tude spectrum, which makes the image more difficult to interpret. In order to compensatefor the loss of energy, the whole range of the frequency has been divided into multiplebands, and for each band, an artificial gain has been applied to flatten out the energy spec-trum. Figure Figure 3.38 shows the comparison between the original CMP gather and thespectrum whitened gather. Figure Figure 3.39 shows the corresponding energy spectrum.Notice that there is an energy drop in the spectrum after spectrum whitening. This hap-pens because the model used implements last years frequency range: 4 to 128 Hz whilethis year’s ranges from 4 to 140 Hz.Time Migrated Section:Migration is the process in which seismic events are geometrically relocated in either spaceor time to the correct location of the event in the subsurface. Post stack time migrationinvolves processing a seismic section that has already had its CMP gathers stacked. Wethen apply this correction to the seismic data in time coordinates. There are various waysto do time migration; in this case we applied a Kirchhoff Migration. This method involves”swinging” each sample along a semicircular path described by the velocity field. The theorybehind this method is to put a sample at every possible location, eventually finding thecorrect location for a sample. This occurs where the sample builds constructively withother samples.Depth Migrated section:


Figure 3.28: This is the general flow we followed to process seismic data from the 2017 Field Camp.The following steps are described in more details below.

Depth migration deals with a seismic section that has already had its CMP gathers stackedand applied in depth (Regular Cartesian) coordinates which is calculated from seismic datain time. Because of this, the post stack depth migration requires a velocity model basedon interval velocities. Once we create the velocity model and test for accuracy, there is asignificant advantage of the depth migrated image:it can be used successfully in areas withlater velocity variations.

3.8.5.2 Processing OverviewAfter implementing the steps described in the processing work flow above, it is evident that

there are many steps that we must take to go from a raw seismic shot record to a final image thatis capable of being interpreted. With that being said, there were steps in the processing workflow that took priority over others to ensure a satisfactory image. Some of the most importantsteps in the work flow included: elevation statics, normal moveout correction/velocity analysis,and residual statics. In order to create the best image, we had to implement these flows ina meticulous manner because these corrections had the largest impact on the quality of theoverall data. We used the other steps in the flow to enhance geologic features such as faults andreflectors in the dataset.

It is worth noting that during the processing work flow, we implemented certain steps morethan once, (velocity analysis) and in several cases, between 3 and 5 times. Also, we did not useevery step implemented during the processing stage to create the final image of the subsurface.For example, we implemented refraction statics early and often in our work flow, but we did notinclude it in our final image. This is because when the refraction statics were included in ourfinal time/depth Post-Stack migrations, it destroyed the continuity of our reflectors on the eastend of the line. Figures Figure 3.29 through Figure 3.41 show how important processing is, as ittransforms an initial shot record into a final depth migrated seismic section.


Figure 3.29: An example of a four different shot records along the seismic line. The third recorddown is an example of a record we decided to ”kill” due to the saturation of noise throughout theentire record.

3.8.6 Errors & UncertaintiesThe errors and uncertainties in the data are mostly emanated from the survey acquisition

done in the field. These errors include:1. Vehicles

• Vehicles such as pick-up trucks that pass by the Main Line while the Vibroseis isongoing may contribute to noise in the data. This is because the extra vibrationscreated by vehicles could be picked up by the geophones as they are recording data.

2. Windy and/or Rainy Weather• Bad weather throughout the day such as wind or rain during the acquisition could

affect the signal connections of the geophones as well as the cables. Since most of theequipments used are extremely sensitive to fluids, rain-water hitting the geophones orthe cables can contribute to unwanted noises in the data.

3. Equipment Error• This type of error occurred when the wireless seismic nodes were used on the field.

When these nodes are not properly checked before using them, they may be put intoa predeployed state, which is called an ’IMX error’. This impedes the nodes fromobtaining signal connection for it to work properly.

4. Geophones Spacings and Orientation• Geophones that were put in at an angle instead of vertical could provide error into

the data, and affect the signal connection of the geophones. Moreover, the spacingsbetween the geophones may not be exactly 2 meters from each other, which could alsolead to error in the data. However, these errors can be ignored during data processingas they do not affect the quality of the data.

5. Other Crews in the Area• Crews from other methods that are within close range could contribute to additional

vibrations to the data, such as Hammer Seismic or DC/SP methods. These errors


are created as these crews rely on using hammers which create additional vibrationswhich can be detected by the geophones.

Errors are also most likely to occur during data processing due to the students’ unfamiliar-ity with the software and the procedural flows that can be quite challenging to follow. Someof the processing errors include:

6. Crooked Line Geometry• Errors may have arisen in trying to produce an image of the subsurface due to the

crookedness of the survey line, in which many of the CMP’s are located away from theline of interest.

7. NMO Correction• The NMO Correction is dependent on the offset and the velocity of the stacks. How-

ever, the velocities are not relatively accurate which could have contributed to error inflattening the reflections.

8. First Break Picking• First break picking can lead to a lot of errors if not done correctly. For example,

selecting inaccurate traces as the first arrival can ultimately contribute to errors inthe steps following first break picking, as well as error in the total observation of thesubsurface.

9. Velocity Analysis• The stack obtained from this analysis is derived from supergathers with the assump-

tion that they do not change shape with the offset. Not only that, the stack fromthe velocity analysis is also affected by the preceding processes that may also containerrors, and overlapping reflections from previous steps.

10. Depth Migrated Section• Errors produced from preceding processes leading to the depth migrated section will

make it harder for geophysicists to make accurate velocity estimation, therefore lead-ing to inaccurate assumptions of structural targets in the subsurface.

Systematic errors may be obtained from the interpretation of assumed geology features.Other errors may derive from the experimental arrangement, such as altering the sweeplength from 4 sweeps, to 2 sweeps, and 3 sweeps which could affect the quality of the data.

3.8.7 InterpretationFigure Figure 3.42 is the final product of the depth migrated seismic image. After a thorough,

cross-method analysis, in addition to having a strong understanding of the geological featuresseen in the depth migrated stack, we could make a well informed interpretation of the subsur-face geology about the main survey line. The initial steps in making the interpretation includeobserving and hand-drawing the reflectors that show significant contrasts in the seismic lines.In order to make inferences about the geology of the subsurface, we must understand the re-gional geology as well as the depositional systems of the area. Collaboration with other groups(outside of deep seismic) is vital in making the interpretations accurate and synchronized withthe interpretations of other methods. Discussions with other method groups, specifically the MTand the Gravity groups, helped significantly when predicting the depths of the formations. Thecore data from the Geology group provided information on the velocity of these formations, whichfurther validates our interpretation.

Figure Figure 3.43 shows the Depth Migrated Cross-Section with the interpreted geology ofthe formations in the subsurface. We generated this cross-section from the final processed imageof the depth migrated section (processed earlier) as shown in Figure Figure 3.41. The reflectorsare characterized by different colors, which are described in the legend. There is also a faultin the lower section of the stack, which is denoted by a straight black line. We drew the faultin such a way that it can be characterized as a normal fault; however, previous studies in the


area suggest otherwise. The 2015 field camp report speculated that the fault is instead a thrustfault. This is due to the compressional forces in the basement layer which are applied in themaximum horizontal direction, hence initiating a thrust fault, instead of a normal fault. Despitethis assumption, we based our interpretation of the fault mainly on the placement of the hangingwall compared to the foot wall seen in the 2017 stack. Therefore, we cannot conclude what typeof fault exists at this location based on the information we have.

Essentially, the layers of the formations found in the stack show relatively horizontal strataswith no major dippings and comparatively minor foldings. The top layer of the stack, shown inlight green, is the Lewis Shale. The Lewis formation thickens consistently from West to Eastwith a depth of 150 meters below the surface. The Sonic Log interpretation given by the Geologygroup shows that the Lewis shale has the slowest velocity as opposed to the other formations. Thethin layer that comes underneath the Lewis Shale is the Mesaverde formation, which consistsof a mixture of shale- and sandstone- type rocks. The third layer is the Mancos Shale with asubstantial thickness of about 750 meters. Below the shale is the Dakota Sandstone formationbefore reaching the Morrison Formation and the Jurassic sandstones. The Morrison and theJurassic sandstone formations are easily recognized from the rest of the formations in the depthmigrated section due to the significant differences in the stacking patterns. Furthermore, theMorrison, which is composed of sandstones and mudstones, have a relatively high velocity, basedon the Sonic Log interpretation. The left side of the cross section, at about depth of 2000 metersresides an exclusive formation called the Ancestral Rockies. This formation has a thicknessof about 325 meters dipping West. One of the assumptions made about the Ancestral Rockiesis that most of its sediments have eroded due to faulting. Therefore, we can only observe asegment of the formation. The rest of the Ancestral Rockies formation has been replaced bynewer formations lithified above the Precambrian basement. The bottom layer as denoted inbrown, is the Precambrian basement which has a depth of relatively 2240 meters.

The primary goal of the deep seismic method in the 2017 field camp was to image the geologicsubsurface of the main line, in hopes of understanding the geothermal system of Pagosa Springs,as well as to observe possible intrusions of fluid flow in the area. There were known geologicalfeatures that could contribute to the diversion of fluid flow along the main line. However, thelocations of the known dikes happen to be located on the crooked segment of the main line. thisis problematic because during the crooked line processing step, these areas lost substantial reso-lution(CMP’s had to be interpolated in these areas). Therefore, we cannot confirm the location ofthe dikes or the orientation of the fault. Because of the loss of resolution and uncertainty regard-ing the fault, in addition to the interpretation of the formations described above, we concludethat there is not sufficient evidence to claim there is fluid flow at the main line.

3.8.8 Conclusions3.8.8.1 Summary

During the 2017 field camp we conducted a 12 kilometer long seismic survey along CountyRoad 542 and County Road 359 in Chromo, Colorado. The goal of the seismic survey was toimage subsurface features to better understand fluid flow in the area. This survey required agreat deal of time and effort from multiple parties and was the most labor intensive survey everconducted at field camp. In the field, students were tasked with stomping geophones into theground along the entire length of the survey line while the seismic source was generated by thevibroseis truck. The truck took a shot every 10 meters along the line and geophones were placedevery 2 meters along the line. Students, with the guidance of industry professionals, designedsurvey parameters that would yield high quality data but also be efficient enough to finish thesurvey in the alloted time. Preliminary QC work was done on the data in the field to correctfor geometry issues, and to ensure quality data was collected. The hard work of the students,coupled with the knowledge and experience of industry professionals and professors generated asmooth and successful deep seismic field survey.

The next step in the seismic work flow was to process the seismic data in the computer labback at Colorado School of Mines. With the help of two experienced seismic processors, students


with relatively no processing experience eventually generated a quality image of the subsurface.Both theory and practical applications of seismic processing were emphasized each day duringthe processing portion of field camp. Using SeisSpace, students learned the process of denoisingdata by performing static corrections, NMO velocity analysis, and surface wave noise attenuation.Students iterated these processes countless times, but ultimately a clean, final depth migratedsection was produced for interpretation.

Figure Figure 3.43 shows the geologic interpretation of the depth migrated section, in whichthe layers of the formations are relatively horizontal stratas with no major dippings or faultings.The fault found in the basement is uncertain and does not likely prove fluid flow in the area.Moreover, the existing dikes as shown in Figure Figure 3.24 are not present in the stacked imageas they are situated on top of the crooked segment of the Main Line. Hence, the results shownfrom the interpretation do not show sufficient evidence to indicate any geothermal activity alongthe main line, which is the underlying objective of the deep seismic survey.

3.8.8.2 Wireless nodes vs wired geophones3.8.8.3 Ground roll test3.8.8.4 Integration With Previous Years

The survey line of 2017 overlapped with the survey line of 2015 by approximately 500 me-ters, as shown in figure Figure 3.48 and Figure 3.49. The total survey line combined is figureFigure 3.50. Theoretically, the east side of the seismic image of 2017 should tie directly into thewest side of the seismic image of 2015. In figure Figure 3.52, the main reflectors on the east sidelook similar to those of figure Figure 3.51. The depth of reflectors of both seismic images matchwell, but not perfectly. There are two main reasons. First, the methods chosen for data process-ing are not exactly the same for these two years, so the results are not be perfectly consistent.Second, the overlapping segment is where the main line of the 2017 survey line becomes crooked,bending almost at a 90 degree angle. Because of this, the imaged location is not exactly wherethe 2015 survey began so the subsurface geology of the two images will not match precisely.

3.8.8.5 RecommendationsOne of the largest difficulties faced in processing and interpreting the data collected during

this year was the geometry of the main line. Compared to previous years, the geometry of themain line was far more complicated and bent multiple times. This led to CMP’s that deviatedmuch farther from the Main Line in comparison to previous years with the maximum deviationreaching 400 meters. One recommendation for future seismic lines is to work in locations thatallow for a slightly straighter seismic line. This is because during processing, the stratigraphiclayers weren’t necessarily representative of the geology directly along the main line, and couldhave either introduced or removed geological features in the final image.


Figure 3.30: The layout of the main seismic line for the 2017 Field Camp. The colored line signifiesthe surveying line (colors indicating fold) while the white area indicates the location of the CMP’sfor all the records along the line.


Figure 3.31: a) Uncorrected reflection; b) Proper Velocity; c) Velocity is too low; d) Velocity is too high.Yilmaz (2001).

Figure 3.32: The initial brute stack obtained using initial NMO velocities and elevation statics.


Figure 3.33: The brute stacked image with refraction statics applied to it.

Figure 3.34: From left to right: (1) original shot, (2) 2000 m/s denoise, (3) 1500 m/s denoise and(4) 1200 m/s denoise.


Figure 3.35: Stack with a second iteration of NMO velocity analysis and a surface wave noiseattenuation filter applied to it.

Figure 3.36: Velocity Analysis of the Denoised CMP Gather.


Figure 3.37: Brute Stack Image After Applying Residual Statics Correction.

Figure 3.38: (1) CMP gather of 2800, (2) Time variant spectrum whitened gather of 2800.


Figure 3.39: The green curve corresponds to the energy spectrum of the original gather. The bluecurve corresponds to the energy spectrum of the whitened gather.

Figure 3.40: Time migrated section of the main line using Kirchhoff migration.


Figure 3.41: Depth Migrated section of the Main Line, from West to East, and the Elevation Profileof the Line.

Figure 3.42: The final product of depth migrated image (colored).


Figure 3.43: Depth Migrated Cross-Section with the Interpreted Geology of the Subsurface alongthe Main Line, from West to East. The x-direction of the cross-section is equivalent to the stationsstarting at Flag 1000 until Flag 2200.

Figure 3.44: Shot at station 1560 recorded on wireless nodes. Station spacing 1.25 m.


Figure 3.45: Shot at station 1560 recorded on wired geophones. Station spacing 10 m.


Figure 3.46: Ground roll test experiment. Examples of shot records for every location indicated onFigure Figure 3.47.


Figure 3.47: Shot locations for ground roll test.


Figure 3.48: The survey line of 2014 and 2015 field camp.

Figure 3.49: The survey line of 2017 field camp.


Figure 3.50: The survey line of 2017 and 2015 field camp combined.

Figure 3.51: The depth migrated seismic image of 2015 field camp.


Figure 3.52: The depth migrated seismic image of 2017 field camp.


3.9 Results and Interpretation3.9.1 Overview

Each method mentioned previously was used to investigate the main line subsurface. Unfor-tunately, not all of the methods were useful in imaging the subsurface along the main line dueto a lack of anomalous material or low depth of investigation. However, geophysical methods areinherently non-unique which meant that it is paramount to use all methods available to allow forconclusive results that agree to create a working model for the area of interest. The interpreta-tion of the main line in Chromo integrated results from Geology, Deep Seismic, Magnetotellurics,and Gravity to create a finalized model of the area. Other methods deployed along the main line,but were not used in interpretation included: self potential, DC resistivity, magnetics, passiveseismic, and frequency domain electromagnetics.

3.9.2 GeologyTo begin interpretation, the Geology field camp section created a preliminary cross section of

the 2017 main line, Figure 3.53, to provide other processing teams with expected thicknessesand depths. The geology team also measured formation densities in the Petrophysics lab atColorado School of Mines using borehole samples provided from a local well, TG-1.

Figure 3.53: Original geologic cross section of the main line in Chromo.


3.9.3 Deep SeismicUsing measured densities provided by geology, the deep seismic team was able to migrate

their acquired data from a time to depth, seen in Figure 3.54. This was done by calculatingacoustic velocities for each geologic layer in the subsurface which can be seen with a coloredoverlay in Figure 3.55. Geology was then able to use the depth migrated velocity model to sep-arate lithologic layers that trend similarly to previous field camp years (2014-2015) Figure 3.56.There is a point of interest on the western side of the main line, in which a wedge of new ma-terial appears. This has been denoted as ancestral rockies which created a small discrepancybetween seismic interpretation and the working gravity model, but agrees with the majority ofmagnetotelluric depth calculations.

Figure 3.54: Depth Migration of well velocities for deep seismic

Before the well logging data was available, the interpretation for the stacked image was mostlydone through a synthetic velocity model that was created based on the first-break pickings andthe velocity analysis. The velocity model was used to migrate time-migrated section to depth-migrated section. Then convert it back to the time-migrated section and replace the initial velocitymodel with a second velocity model that was created based on the well data provided by theGeology team. Although the well logging data was not taken directly on the Main Line, the datawas still used for the model with the understanding that velocities within the subsurface shouldnot change provided the formations are the same. Therefore, the velocity model with the welldata is then used to convert the time- migrated section into the final depth-migrated section for acomplete interpretation of the geologic features. The final stacked image shows promising resultsas the seismic data fits the velocity model with the well log data very well.

Figure 3.55 shows the depth migrated section with a color scheme implemented onto it todistinguish the reflectors and layers in the subsurface. Three distinctive reflectors are recognizedto separate the blue, green, yellow and red layers. These layers represent the velocity changesin the subsurface. Notice the lines that separate the layers are not right on where the reflectoris. The reason is because, as mentioned previously, the data was not collected right on thesurvey line, so the depth information may not be as accurate as the velocity information is,which makes thevelocity contrast a primary factor of picking the boundaries of layers. Furtherinterpretation suggests that there is a unique layer towards the left of the image denoted inorange. The seismic image does not show a distinct boundary between the yellow and the orangesections, but if those two layers are merged together, the thickness of this combined formationis not approximately constant, which is contradictory to the common sense that the thickness


Figure 3.55: Depth Migration of seismic with well velocities included. Velocities increasing fromsmall to large moving down the section. Note the section of interest near the basement layer in thebottom left.

of a formation is usually constant. The similar events were found in the 2014 and 2015 fieldcamps. One possible explanation is that before the yellow layer was deposited, there was a faultthat offset the orange layer, and the part on the other side of the fault plane got eroded, and thenthe yellow, green and blue layers were deposited, so that is why only a small chunk of the orangeformation left in the subsurface.

Essentially, generating a depth-migrated section using the well log data is a crucial step priorto interpreting the geologic features as it shows a more accurate positioning of the reflectors asopposed to just using geology information to locate the different formations in the subsurface.

The depth migration was created by the seismic crew and interpreted by the geology crew. Inorder to identify geologic layers, information from many sources had to be combined. First, thegeology team looked at well and core data from the Pagosa Springs area and along the main line.The well data from the main line never penetrated deeper than the Mancos Shale thus preliminaryinterpretation had to come from knowledge and trends of the surrounding area. Basement wasthe most obvious to identify, so geology started from there and worked up. The thickness ofthe layers based on the seismic reflectors were reasonable when compared to regional geologictrends. The fault in the basement did not seem to propagate through the overhead layers, thusthis is believed to be an old fault that happened prior to the deposits of the upper layers.

3.9.4 MTThe magnetotellurics team was able to calculate layer depths based on resistivity sections

that were taken on either side of the Eastern main dike near the main line. Of the thirteensurveys done around the dike only four(site 1, site 2,site 4, and site 6) provided feasible results.In the inversion of the data there were various models that fit the curve. Certain models wereomitted on account of illogical depths and unreasonable resistivity values. The remaining modelswere then compared to the stratigraphy of the region by the geology team. Resulting efforts werecombined to make Figure 3.57, which is the magnetotellurics data layered on the seismic depthmigration. This fits the lithology of the area and an anomaly is clearly seen in a change inresistivity along the western side of the dike. Attempts to integrate this years data with previousyears for MT was not very effective. Although 2015 data did show similar resistivity trends asdiscovered this current year and can help reinforce some of the current conclusions drawn. A


Figure 3.56: Depth Migration of seismic using well velocities with geologic interpretation

few examples seen in 2015 data include: a positive increase in resistivity between the Mancosshale and Dakota sandstone, as well as a positive increase in resistivity between the JurassicSandstones and the crystalline basement.

Figure 3.57: MT resistivity inversions overlaying geologic velocity section.

3.9.5 GravityThe gravity team was able to integrate data from two previous field camp years because

the two additional lines overlapped with the 2017 main line Figure 3.58. The gravity teamrelied heavily on the borehole information provided by geology, specifically, the density values oflithologic layers beneath the main line. These density values were averaged with previous year’s


measurements of the same layers in order to achieve a more accurate model. Deep seismic andMT also provided important information about depths to each of layers beneath the main line.The depth information was integrated into the preliminary model of gravity in order to fine tunethe model into a more accurate finalized model which agreed with the other methods.

Unfortunately, the geometry of the current main line caused problems for the modeling por-tion of gravity which did not allow for the Eastern portion of the line to be resolved in any greatdetail. The line geometry also made it difficult to compare results from gravity to seismic be-cause of the process in which each method creates their respective models. The gravity modelwas based on a projection of strike from regional geology information, but seismic was able tocreate their model by unravelling the data gathered along the line. However, gravity was ableto image the area of interest in the Western portion of the line. The data shows that there is adensity difference there which agrees with deep seismic and the geologic interpretations. Basedon geologic data, the low density area is assumed to be sediment from the ancestral Rockies. Thelayer appears to be a low density zone in relation to the basement material. This agrees with theresistivity change that is found in the MT data and the apparent velocity change found withinthe seismic data. It also relates very well with previous years gravity surveys which also saw anarea they classified as a low density zone on the hanging wall side of a reverse fault they found.

Figure 3.58: Final gravity model based on seismic, geology, and MT. This cross section includesgravity conclusions from 2014 and 2015, at approximately 4.2 km onward.


3.10 ConclusionThe interpretation of the various geophysical methods finds that there is an area of interest

on the western side of the main line, this formation has contrasts in density, velocity, andimpedance Figure 3.59. This is on the western side of the eastern dike found on the mainline.With the methods used to interpret the main line, the 2017 geophysics students were able toagree on the subsurface structure and geology in the region. Unfortunately, an area of watersaturation was not discovered using the methods integrated on the main line.

Figure 3.59: MT resistivity inversions overlaying geologic velocity section with gravity.

The additional layer in the subsurface on the western side of the eastern dike is indicativeof a fracture within the basement. This provides a crucial element to geothermal energy in thatthe fault is an opening for potential water flow in the subsurface. Additional surveys need tobe taken order the area in order to best understand the area and further prove evidence ofsignificant geothermal activity. This information has adjusted the geologic cross section and asa result a final cross section of the main line has been created Figure 3.60

3.11 RecomendationsIt is the recommendation of the 2017 Geophysics field camp class that subsequent years con-

tinue their line further west along county road 542. It is also recommended that more permittingbe acquired for better use of electrical methods near the line, but away from man made struc-tures. A problem that the 2017 group had was a lack of electrical methods near the main linewhich created difficulties in determining water content; further permitting would allow for possi-ble time domain electromagnetics surveys to be completed. These would give further insight intoresistivities at greater depths. Ideally, a time domain electromagnetic survey could be conductedabove the suspected basement fracturing on the west side of the line which would provide a bet-ter understanding of fluid content. The addition of various electric methods should overlay andprovide answers to the questions of geothermal events due to geologic formations in the subsur-face. This would also allow for other methods, specifically gravity, to create a straight survey linerather than one that curved backwards which introduced uncertainty in the processing portion.


Figure 3.60: MT resistivity inversions overlaying geologic velocity section with gravity.

4. Reservoir Hill Student Site

Contents





4.5 Gravity 179



4.8 Conclusion 204


4.1 Overarching ObjectivesThe primary objective at the student site was to identify the geologic structures of Reservoir

Hill and determine if the hill affects the geothermal system at the Mother Spring. Various geo-physical methods were used, with a focus on electrical methods to provide data that could easilybe combined with data that had been aquired in 2012 on Reservoir Hill.

4.2 Student Site Geology4.2.1 Background Information

The Pagosa Springs area is dominated by large geomorphologic features, including ReservoirHill which is this year’s student site. Reservoir Hill consists primarily of Mancos Shale cappedby fluvial deposits and is suspected to be a part of the geothermal system in Pagosa Springs.


As seen on the geologic map in the Figure 1.7, the dominant formations that characterize thearea are the Mancos Shale, Dakota Sandstone, and Tertiary Volcanics. The San Juan River cutsalong the north edge of the hill. Reservoir Hill is believed to be an ancient river channel of the SanJuan because of the fluvial deposits found on its top. These fluvial deposits are more resistive toweathering than the shale they cover so the topographic high has been formed. Reservoir Hill isparticularly interesting because it is located directly to the east of the Mother Spring. The MotherSpring is used by the city of Pagosa for industrial heating and recreation and is believed to be1000 ft deep. The source of this spring is unknown and makes up the main question we want toanswer. It is possible that the answer could be found by investigating Reservoir Hill.

Figure 4.1: Map view of Student Site

4.2.2 Preliminary InterpretationAs seen in the Figure 4.5 below, the Dakota sandstone outcrops west of Pagosa Springs, with

a measured dip of 10 degrees east. The rest of the Pagosa Springs area, including ReservoirHill, is made up of Mancos shale. The top of the hill is capped by fluvial deposits made up ofmostly eroded volcanic material. The P1 well, drilled just north of the Mother Spring, shows theDakota and subsequent layers at shallower depths than we would have expected eroded volcanicmaterial from the dipping outcrops. From these measurements, the dip gradually decreasesas the unit moves east. This is seen in the Mancos outcrops where the San Juan River cutsthrough the Mancos. At those locations the measured dip is near horizontal. We believe thischange in dip is caused by the nearby Stinking Springs Anticline, which has no affiliation withthe Stinking Springs geothermal source and site of investigation for the Field Camp 2014 mainline. Pagosa Springs is located on the outer edge of the anticline, evidenced by the sedimentarylayers flattening back out and becoming consistent with surrounding geology. This anticline is

156 Chapter 4. Reservoir Hill Student Site

believed to have been formed during the late Cretaceous, during same time frame as the LaramideOrogeny. It is characterized by dips approximately 5-10 degrees on either side of the fold axis.Another potential reason for this lateral change could be another fault that extends more to thenorth than previously identified in the 2013 Field Camp Seismic Line.

On the far east side of Pagosa Springs the 2016 Geophysics Field Camp ran a seismic linealong Mill Creek Road, which cuts through the Lewis shales, Mesaverde sandstones, and Mancosshales. The seismic data shows the Mancos-Dakota contact deeper than expected, based on thedata from the west side of Reservoir Hill. This contact occurs around 1600 ft below the surfaceand according to seismic data the layers appear to dip around 5 degrees. This matches veryclosely with the dips measured on the west.

Based on the data, there is a discrepancy between the depths to formation on both thewestern and eastern sides of Reservoir Hill. There could be several reasons for this discrepancybeside some sort of geologic structure. There are two interpretations made for this discrepancyas seen in Figure 4.5 and Figure 4.6 below. For the first preliminary cross section in Figure 4.5,to account for these different depths, there is likely a fault located near or on Reservoir Hillin between the two sides of the line. Further proof for this fault is seen from the previous DCResistivity data set collected by the 2012 Field Camp over the hill and seismic data set collectedby the 2013 Field Camp that located a fault just south of Reservoir Hill as seen in the Figure 4.3.For the second preliminary cross section in Figure 4.6, it is interpreted that there are potentiallya series of multiple faults in the area, which would account for the depth discrepancies.


Figure 4.2: Stratigraphic column from the P1 well drilled in Pagosa Springs [1].


Figure 4.3: Field Camp 2016 interpreted seismic line [11].

Figure 4.4: Map view of cross section line that is marked in red.


Figure 4.5: First preliminary cross section of Reservoir Hill using well, seismic, and geologic deter-minants.


Figure 4.6: Second preliminary cross section of Reservoir Hill using well, seismic, and geologicdeterminants.


4.3 Hammer Seismic4.3.1 Objectives

The overall objective of hammer seismic is to analyze and assess the near-surface geology.This includes identifying potential faults, identifying dikes, computing depths, and providing anEarth model to be used by DC. Hammer seismic has the potential to identify several differentgeologic features such as faults and dikes during processing. Hammer seismic data can be com-bined with the DC data inversion process to constrain the inversion model. This will help reducecontact resistance signatures recorded in the near-surface. Hammer seismic data provides valu-able insight into near-surface geology which can aid other methods in assessing the potentialhydrological system on Reservoir Hill.

4.3.2 ExpectationsAcquiring hammer seismic data aims to capture an image of the near-surface geologic struc-

ture. Data acquired yields different velocities both above and below a specific interface. Thesevelocities can be used to create basic earth models depicting the profile of the first interface. Mod-els created with the data from hammer seismic would show the interface unit with overburden.Any discrepancy in continuity of the interface unit indicates external deformation or structure.In the event of there being a relatively young fault, the unit would display a vertical offset visi-ble in calculated Earth models. This survey aims to produce a velocity model, identify the firstinterface profile, and interpret the presence of a fault.

4.3.3 Survey Maps4.3.3.1 May 20th, 20174.3.3.1.1 Location

• Student Site, Reservoir Hill• Line: S0• Stations: 0046-0061

4.3.3.2 May 22nd, 20174.3.3.2.1 Location


4.3.3.3 May 24th, 20174.3.3.3.1 Location


4.3.4 Data Acquisition4.3.4.1 Hammer Seismic Setup

Hammer seismic data is acquired using the following procedure:1. Identify location of interest.2. Determine geophone spacing for depth of investigation.3. Measure line with tape measure, secure at each end with nail.4. Lay takeout cable for length of line. Multiple takeout cables may be needed to complete the

length of the survey line.5. Place geode seismograph at the middle of the survey line (See Figure 3.4).


Figure 4.7: Map depicting lines SO and S1 on Reservoir Hill. Black lines depict location of hammerseismic surveys. The red line represents S0 while the turquoise line represents S1. The color baron the right depicts elevation.

6. Stomp geophones at desired spacing within 10 degrees of vertical to reduce error. Be sureto check direction of shear wave geophones to have correct polarity if running shear wavesurvey.

7. Connect takeout cables to geode.8. Connect geode to laptop.9. Connect battery to geode.

10. Open MGOS and check noise window to ensure that all geophones are working properly.11. Choose seismic source. (Sledgehammer, weight drop, etc.)12. Choose shot spacing and number of shots for stack.13. Run test shot to ensure that trigger sensor and geophones are working properly.

4.3.4.2 May 20th, 20174.3.4.2.1 Survey Parameters

• Geophone spacing: 3m• Length of Line: 150m


• P-wave shot spacing: 4m• P-wave number of shots: 5• S-wave shot spacing: 37.5m• S-wave number of shots: 6

4.3.4.3 May 22nd, 20174.3.4.3.1 Survey Parameters

• Geophone spacing: 2m• Length of Line: 100m• P-wave shot spacing: 4m• P-wave number of shots: 3• S-wave shot spacing: 8m• S-wave number of shots: 6

4.3.4.4 May 24th, 20174.3.4.4.1 Survey Parameters

• Geophone spacing: 1.5m• Length of Line: 75m• P-wave shot spacing: 4.5m• P-wave number of shots: 3• S-wave shot spacing: 18m• S-wave number of shots: 3

4.3.5 ProcessingWhen processing hammer seismic data, two main types of analysis typically occur–reflection

and refraction. Due to the shallow depth of investigation of the surveys conducted, only refractionanalysis was completed with the data acquired. Refraction analysis is a first-order interpretationmethod that allows for velocities of the overburden layer and the first layer to be computed. Usingthese velocities, the depth to the first interface can be found. Finding depths at multiple shotpoints along each line provides x (location of shot) and z (depth to first layer) coordinates that areused to create a basic earth model. This model can be used to constrain other methods’ inversionprocesses while also providing insight into the near-surface geology. Analysis of the earth modelallows for the interpretation of potential faults or dikes that may influence the hydrology of thearea.

4.3.5.1 Refraction AnalysisRefraction analysis is conducted using MGOS seismograph controller software as well as

Paint XP, however paper and colored pencil can be used. To begin, a shot is selected from oneend of a survey line. Using the knowledge of how waves propagate through the subsurface, thefirst arrival represents the head wave and refracted wave of the first interface. Extend the shotlocation vertically downward as a reference line to find ti for the velocity below the interface,or V2. Slope lines can be drawn to fit the different traces for each channel. The steeper slopeoriginating closest to the shot point is the velocity above the interface, or V1. The more gentleslope extending laterally towards the end of the survey line is the velocity below the interface, orV2. The V2 slope line should be drawn before the V1 slope line as it easier to constrain each inthis order. See Figure 3.5 below for a visualization.

It is apparent in Figure 3.5 that choosing the correct slopes to fit each channel is difficult.Slope lines may be modified to better fit different channels. Once satisfied with the slope linesfor V1 and V2, velocities can be calculated by using Equation 2.3. Computed velocities along withti can be used in Equation 2.4 to yield a depth to the first layer. This process is repeated foreach end of the line along with several shots along the line to create a basic earth model. Datacan be interpreted between points to create a model with reduced noise. It is important to onlyconstrain the model with a few data points. Over constraining the model with too many shot


points can create a noisy model that is not a realistic representation of the subsurface geology.Results of refraction analysis can be seen below in the Results subsection.


When picking slope velocities by hand, some inherent error exists. It is difficult to pick aperfect slope that accurately represents each data point due to some smaller fluctuations insubsurface velocity. Even with drastically magnified shot traces, this leads to velocities that arenot perfectly representative of the data collected. This was especially apparent while conductingrefraction analysis and choosing V1 slopes as the slope could only be constrained by 3 geophonechannels. Properly constrained velocity slopes will take more into account. However, the datacollected proves difficult to choose a correct slope. To mitigate this error, a constant V1 value wasused for P-waves and S-waves when calculating depth to the first layer. The V1 velocity used forP-waves is 400 m/s. The V1 velocity used for S-waves is 250 m/s.

One source of noise that was mitigated during processing stemmed from inaccurate z datapoints used to constrain the earth model. Interpretation of data can be done between referencepoints used for the earth model. Noisy data points can over constrain the earth model. Removingthese points and interpreting a depth by averaging the nearest neighboring points yields a morerealistic earth model.

4.3.6.2 Field ErrorsA few errors occurred during data acquisition and processing. To begin, during data acqui-

sition of shear wave data, horizontal geophones are directionally dependent. A few geophonesshow reversed polarity in several shots acquired on different survey lines. While this error isnot extremely detrimental to the overall acquisition, it can cause problems in refraction analysisprocessing when picking slopes to calculate velocity. Another error is the geophone coupling withthe ground. Reservoir hill contained a large layer of dead shrubbery, fallen branches, wood chips,and pine needles. All of this needs to be removed in order to make a proper contact between thegeophone and the ground. Bad coupling will result in poor data acquisition for that geophone.This appears on several shots acquired in the field for a number of surveys. Lastly, one error thatoccurred was equipment malfunction. The wire attached to the trigger sensor attached to thesledgehammer was cut for the survey on May 24th, causing an error in acquisition as randomsamples were taken by the sensor when not in contact with the plate.This was discovered aftertest shots and 1-2 regular shots along the line. The sledgehammer was switched out for a smallerhand held hammer and acquisition continued.

Noise collected during acquisition is due to a number of different sources. Foot traffic oranimals create small seismic sources that are picked up by the geophones. Another sourceof noise is acoustic waves present during acquisition, this includes the sound created by thecontact between the hammer and the plate as well as planes flying overhead. Vehicles producenoise, including planes flying over and cars driving by. This noise typically appears before thefirst arrival. Lastly, strong wind will produce noise on geophones that are not fully buried. Noisydata can be avoided by accounting for these sources while acquiring data. However, as time is aconstant constraint, some noise is unavoidable during acquisition.


Recommendations for processing include double-checking that the first arrival chosen is aresult of a refraction and not noise. Another recommendation would involve choosing slopesfrom the top of the trace for each channel as well as drawing the V2 slope line before the V1 slopeline. In addition, gaining permission by email to use the test version of the ReflexW softwarefrom Karl-Joseph Sandmeier before processing would prove to save many hours and producemore accurate plots. This allows for one to produce velocities for different shots more easily.


The software also allows for an inversion of the data to take place to produce an earth model byfinding depth for every shot location. Using this software will allow for a more comprehensiveanalysis of the subsurface and will streamline the tedious process of creating an earth model byhand. Another recommendation is not to be afraid to start drawing slope lines for the differentvelocities and adjusting them to fit different traces. This will allow for some room for error inproducing different velocities for V1 and V2. Lastly, if V1 velocities are consistently low/high, usea constant V1 velocity to find depth for each shot point.

4.3.7.2 Field RecommendationsRecommendations for future hammer seismic acquisition and processing teams include a

number of suggestions that will make the entire process easier. For acquisition, they must besure to choose a survey area that is relatively flat with minimal shrubbery and ground coverage.This will allow for easier acquisition with reduced chance of error. Less ground coverage makeslaying cable easier and ensure stomped geophones will have a good coupling with the ground.Another recommendation includes checking that all equipment is in proper working conditionbefore attempting to acquire. This will result in a smoother survey and reduce troubleshootingin the field. Data will also be more accurate than that acquired with malfunctioning equipment.Lastly, be sure to communicate with other crews or people present to reduce noise during ac-quisition. Additionally, lines between SO and S1 should be surveyed in order to determine thereason for the discontinuity found in S1.

4.3.8 Student Site Results

Figure 4.8: Figure depicting the elevation and shale profile (top) and the calculated depth to the firstlayer, or difference between topography and the first interface (bottom), for line S0 on Reservoir Hill.The first layer is interpreted to be Mancos Shale. The flag numbers are oriented East-West withincreasing flag number. Date of survey: May 20th, 2017.


Figure 4.9: Figure depicting the elevation and shale profile (top) and the calculated depth to the firstlayer, or difference between topography and the first interface (bottom), for line S1 on Reservoir Hill.The first layer is interpreted to be Mancos Shale. The flag numbers are oriented East-West withincreasing flag number. Date of survey: May 22nd, 2017.

4.3.9 InterpretationAfter conducting refraction analysis on the hammer seismic data collected, some conclusions

may be drawn. When viewing Figure 4.8 and Figure 4.10, it is apparent that the earth modelremains constant when progressing through greater flag numbers. This suggests that the over-burden thickness remains constant across the survey line. This conclusion is reinforced whenviewing the calculated depth, or difference, profiles in each figure. Both difference profiles remainbetween 4-6 meters for both Figure 4.8 and Figure 4.10. Due to this, it is difficult to interpretthe potential presence of a fault in this area. However, this consistent profile may be easily im-plemented as a boundary constraint, or reference model, for the DC inversion process. This willallow the contact resistance to be reduced in the inversion model.

When viewing Figure 4.9 in comparison to the previous figures, there is an observable dif-ference in the earth models as well as the calculated depth profiles. This could simply be theresult of the different locations of S0 (Figure 4.8 and Figure 4.10) and S1 (Figure 4.9). However,it is more likely that the different locations suggest changes in the subsurface geologic structure.One main observation is the relative distance between the shale profile and the elevation profileseen in Figure Figure 4.9 at flag number 59. As displayed in the calculated depth plot, the dif-ference profile between the ground and the first interface is found at a depth of only 1 meter.Subsequently at the following station, the depth to the first interface increases to 2.73 meters.This is an unclear indicator of a fault, but this suggests some geologic structural change over ashort distance (10m). This change could be indicative of a fault or the result of erosion. Furtheranalysis is needed to shed more insight on this area in order to conclude the presence of a fault.Due to the shallow depth of investigation involved with hammer seismic, it is difficult to discernthe exact geologic structure present in the subsurface.


Figure 4.10: Figure depicting the elevation and shale profile (top) and the calculated depth to thefirst layer, or difference between topography and the first interface (bottom), for line S0 on ReservoirHill. The first layer is interpreted to be Mancos Shale. The flag numbers are oriented East-Westwith increasing flag number. Date of survey: May 24th, 2017.

Velocity analysis of Reservoir Hill along both S0 and S1 allows for the interpretation of thegeologic unit of the first interface. Because both P-wave and S-wave data was collected, twoseparate comparisons can be made to see the accuracy of the velocities yielded to find depths.This is a good quality control check that can be implemented during processing. The averagevalue of all V2 P-wave and S-wave velocities represents the accuracy of the velocities computed.The average P-wave velocity is 2319 m/s while the average S-wave velocity is 1501 m/s. Thesevalues are less than 120 m/s off of standard velocities found for Mancos Shale (Vp = 2200 m/s, Vs =1400 m/s). This suggests that the first interface is the Mancos Shale. This data can be comparedwith sonic log data collected on field samples to check the accuracy of the velocities computed.Results using S-waves were omitted due to their redundancy and poor quality compared to P-waves.

4.3.10 ConclusionsAfter conducting refraction analysis on the four lines of hammer seismic data collected, some

basic conclusions can be drawn. To begin, the velocities yielded for both Reservoir Hill along lineS0 and S1 are consistent with each other. The average V2 P-wave velocity found is 2319 m/s whilethe average S − wave velocity found is 1501 m/s across both lines. These findings suggest thatthe first interface on Reservoir Hill is Mancos Shale. After finding velocities, different values fordepth to the first interface were computed. These results can be seen in Figure 4.8-Figure 4.10.The earth models produced suggest that the average depth to the Mancos Shale interface is about4-6 meters along line S0 while the average depth for line S1 is about 2-4 meters on the Westernhalf, and 5-7 meters on the Eastern half. An area of interest exists in Figure 4.9 at flag number59. This has been identified as a potential fault, however no definite conclusion can be drawn


without further investigation and integration with other methods. The earth models producedwere used by the DC resistivity inversion team to constrain their model. More data acquisitionand analysis over this area should take place in order to properly identify the fault and gaininsight on the near-surface geology.


4.4 DC Resistivity and Self Potential4.4.1 Objectives

The objective of the overall survey is to characterize the hydrology and lithology of the regionsurrounding Reservoir Hill. There is a suspected fault in this location which may be redirectingthe fluid flow in this region. With DC resistivity and SP surveys we attempt to characterize theconductivity and telluric currents at this site to either confirm or deny the presence of this fault.The conductivity profile will enable an improved understanding of the layering of the geologicformation, aiding in the knowledge of the local lithology. Then, apply the self-potential data tomodel the telluric currents, which relate to the fluid flow in the subsurface. The fluid flow thenimproves the understanding of the local hydrology.

4.4.2 Expectations4.4.2.1 DC Resistivity

The emphasis on hydrology stems from the nature of the anticipated DC and SP surveyresults. The DC survey measures apparent conductivity, which leads to information with regardsto saturation and lithology of the subsurface. Based on the geology of the area, we expect near-horizontal bedding. It is also likely that the majority of Reservoir Hill consists of Mancos Shaledue to the outcroppings seen in that region. If there is a fault on Reservoir Hill, the DC Resistivityis expected to show a truncation between conductive and resistive bodies, where the side of thefault that is saturated with fluid flow is more conductive, likely with a concentrated conductivebody at, or near the location of the fault.

4.4.2.2 Self PotentialSelf Potential characterizes the fluid flow of a region, so if there is actually a fault on Reservoir

Hill, we expect that the data would should something similar to Type I in ??, where there is anupward trend up to a peak and then a drop off of the voltage. This trend typically indicatesupwelling, which commonly occurs in the presence of a fault. This is because the fluid flowsalong the geologic boundaries and is then redirected upon reaching the discontinuity caused byfaulting. This results in a concentration of fluid at, or around, the location of the fault. ReservoirHill is made almost entirely of Mancos Shale, so there should be very little variance in the SPdata geologically, allowing the focus of the survey to be fluid flow indication of the area.

4.4.3 Survey Maps4.4.3.1 DC Resistivity

The DC Resistivity survey encompassed the three lines at the student site (S0, S1, and S2).The survey itself used 20 meter spacing between electrodes, which included every other flag,specifically the odd numbered flags on the student survey lines. At each of these locationsWenner arrays were used with site specific Dipole-Dipole arrays as necessary. The Dipole-Dipolearray occurred along line S2 at the student site.

4.4.3.2 Self PotentialFor the student site, SP measurements were taken at 20 meter spacing. SP was executed

across the entirety of student site lines S1 and S0.

4.4.4 Data Acquisition4.4.4.1 DC Resistivity4.4.4.1.1 Survey Parameters

The DC Resistivity survey began by planting the electrodes in the ground at the respectivesurvey flags with electrodes planted approximately 2/3 of the way into the ground. Cables were


Figure 4.11: Map of the DC Resistivity surveys conducted along Reservoir Hill, S0, S1, and S2.

run alongside and attached to each electrode, with the cables themselves ultimately attaching tothe ABEM. The ABEM lies in the center of the survey, between electrodes 32 and 33.

The following selections were made within the ABEM Terrameter once the device was poweredand turned on:

1. Select Lund Imaging System2. Select Resistivity Mode (other modes include SP and IP)3. Enter Nameline:filename4. Enter Smallest Electrode Spacing: 20 meters5. Enter Powerline Frequency: 60 Hz (for United States measurements)6. Enter Protocol

• For the Wenner array we used WEN64XL• For the Dipole-Dipole array we used DDP464XL

7. Set the midpoint (numbering is arbitrary provided incrementation)8. Set current to 200 mA9. Set acquisition time to 0.5 seconds

10. Set acquisition delay to 0.3 seconds11. Set cycle time to 3.8 seconds12. Skip internal errors (collect data, retroactively assess)13. Set minimum stacks to 214. Set maximum stacks to 515. Set error limit to 5%

Subsequently the ABEM system will run an electrode check. At this time the ABEM system willoutput any failed electrodes. Potential electrode failures derive from improper coupling or im-proper electrical connections. To improve coupling, pour saltwater on the problematic electrode.To resolve errors derived from connection errors, verify proper setup and the integrity of thecables and electrodes. A rule of thumb for trouble shooting is that multiple adjacent electrode


Figure 4.12: Map of the SP surveys conducted along Reservoir Hill, S0 and S1.

failures tend to derive from a connection error, while isolated electrode failures tend to resultfrom coupling issues.

Once a sufficient number of electrodes test as working the survey may begin. The WEN64XLprotocol is written such that a roll along may be conducted. Once the ABEM completes all testsregarding an electrodes as well as the desired survey, the electrode may be moved to the newsurvey site for the next measurement.


4.4.4.2.1 Survey Parameters

SP is a very portable method. To start the survey, choose the first flag location and take a tipto tip measurement of the electrodes. Then bury one of the electrodes, the reference electrode,about 3/4 of its length at the flag and connect it to a spool of wire. Plug the other electrode,the roving electrode, into the green port of the voltmeter and set the voltmeter to voltage (mV).Next, take the roving electrode, voltmeter, and spool of wire to the first flag, 20 meters away,and dig three holes using the rock hammer (just a few inches to get better ground contactand moisture in the soil). Take three measurements in these holes by placing the tip of theroving electrode into the soil and holding it there until the voltage value being shown by thevoltmeter appears to be stable. Each of these measurements, along with the flag number andtime of measurement, should be recorded in the field notebooks. This process continues untilthe desired line is completed. After completing the measurements with the roving electrode,return to the buried reference electrode and take another set of tip to tip measurements. Thisis a very important step as these tip to tip measurements are the only way to account for theinstrument drift when processing the data. Once the loop has been closed out with the final tipto tip measurement, reel up the spool of wire and proceed to the next line. .


4.4.5 Processing4.4.5.1 DC Resistivity

Data processing began with an initial period of quality control through Geotomo Software’sRes2dInv inversion program. This inversion program produces a 3 iteration computed apparentresistivity pseudo-section. This inversion focuses on identifying and removing faulty electrodesprior to implementing further inversion that produce depth inversions and incorporate topo-graphic effects.

To complete the rest of the DC resistivity processing, use DCIP2D, a code written by UBCGeophysical Inversion Facility. DCIP2D is a non-linear inversion program that allows its usersto calculate conductivity of the subsurface based on the measured voltage from the survey. Theedited ABEM Terrameter data files are manipulated and converted into a DCIP2D data file usinga conversion code.

There are many parameters that contribute to the quality of DCIP2D’s inversion process. Themain parameters examined for these datasets are topography, mesh size, target data misfit, chifactor, and reference model. The best way to find the correct parameters is to execute multipleinversion runs through the same datasets and find the parameters that best fit the model.

To calculate topography use the survey line DGPS coordinates. To find mesh size use thedefault mesh settings, and tweak them to allow for better resolution. Most of the datasets use3-5 meter mesh size horizontally, and 5 meter mesh size vertically. Target data misfit is foundusing the default settings, an inversion uncertainty within five percent of each measurement.

The next calculation is the chi factor, the ratio between the target data misfit and the numberof data. This was kept default (a factor of one) for almost all the surveys except S0 on the studentsite. For S0, the chi factor was changed to two in order to lessen the weight of the data. Thisdecision was made because the data was initially over-fit due to a bad electrode that couldn’tbe removed earlier in processing, creating non-geological resistive anomalies. Changing the chifactor to two produced a superior final product.

The last thing added to the inversion parameters was a reference model. We determined thereference model based on previous School of Mines geophysics students gathering local resistivitylog data. Using this data, the chosen conductivity value for the reference model was 0.02 S/mor 50 Ohm-m. This value is based on Reservoir Hill consisting primarily of Mancos Shale. Thereference model was one of the most important parameters because it gives the inversion avalue for the geologic background, instead of creating a default one based on solely the ABEMmeasured data. A reference model was also created from a geologic cross section. This modelproduced similar results to the strict utilize of our reference parameter for Mancos Shale, furthersupporting our inversion results.

4.4.5.2 Self PotentialEach day, collect the raw data in the field notebooks and then add it to an Excel spreadsheet

according to flag number. The data collected should include the flag number, voltage, and timeof recording. When using the instruments in the field they tend to drift, so once the data is in theexcel spreadsheet it is important to apply a drift correction as well as a shift correction. Thesecorrections are made using the tip to tip measurements taken in the field. Shift correctionsinclude shifting all of the data to a common point. In this case the common point was flag 1000on the main line and flag 0001 on both of the student lines, S0 and S1. Self potential is a relativemeasurement where the trend of the data is the focus rather than specific values. The drift andshift correction to the data was applied using the following equation:

Drift = (FinalT ip− InitialT ipTotalSurveyT ime

∗ TimeSinceInitialT ip)− InitialT ip (4.1)

Once all of these corrections are applied, the average difference between the overlying datapoints along the survey line is calculated and subtracted from the entire data set in order tocorrelate the data. Now all the data should be connected and the general trend of the survey canbe analyzed.


4.4.6 Errors and Uncertainties4.4.6.1 DC Resistivity4.4.6.1.1 Processing Errors

Processing introduces potential errors within the interpretation of the DC data. These po-tential errors primarily focus on the inversion process. The inversion procedure utilizes certainbackground parameters (such as a 0 voltage at an infinite distance, which is a boundary condi-tion implemented in the DCINV2D code) which may not always be correct for the specific surveysite. This produces potential artifacts within the inversion. Likewise, the quality control processand the final inversion utilized two different inversion programs. This improves the quality ofthe result by reducing the potential for inversion artifacts to skew the procedure. To ensure theaccuracy of the data, compare the results of this survey to the results of past surveys in thatsame region.

4.4.6.1.2 Field ErrorsThe DC resistivity survey uses the transmission of current between metal electrodes through

the subsurface. This process introduces systematic errors when looking at the coupling of theelectrodes, polarization effects, variability in weather conditions, and processing artifacts.

During the survey, attempts to minimize errors were made by using salt water to increasecoupling and implementing low frequency alternating current to reduce the impact of polariza-tion effects. Pouring saltwater over electrodes increases the local conductivity, which improvesthe electrical coupling between the metal electrode and the earth. The usage of low frequencyalternating current instead of true direct current reduces the impact of polarization effects. Thepolarization of the electrodes leads to oxidation-reduction reactions that introduce error into thesurvey. Another potential source of error present at the time of the survey was the weather. Dueto variances in the subsurface saturation across the duration of the total survey exact values ofresistivity cannot be compared at provided locations. Instead, this paper focuses on the presenceof relative changes within each survey line.

4.4.6.2 Self Potential4.4.6.2.1 Processing Errors

One common error that occurs when conducting SP surveys is that the surveyor does notwrite the data collection time down. Keeping time is vital when accounting for instrument drift,so if time is not recorded, the drift corrected data is less accurate.

4.4.6.2.2 Field ErrorsWhen using self potential there are many sources of errors and noise. This method works

the best if the ground is moist, because soil moisture establishes sound contact between theelectrodes and the earth. However, if the amount of water in the ground has changed due to rain,snow, or other events that give the ground uncharacteristic moisture, the SP survey may yieldfalsely high voltage readings. This occurs because the survey will report the voltage of the waterin the ground rather than the ground itself.

It is also important not to use self potential around any methods that inject currents in theearth, such as DC Resistivity, EM, or MT. This is because the electrodes used in self potentialwill measure the voltage according to the currents being produced by the other methods ratherthan the voltage according to the actual earth itself.

When using the voltmeter there is an element of human error because it does not read out asteady value for the voltage; the values fluctuate so it is up to the surveyor to choose a reliablevalue for that particular location.

Finally, having other sources of current or resistive material in the area uncharacteristicof the earth causes noise in the SP data. In this case there were buried cables and possibleirrigation systems along the main line that were carrying current through the earth while SPmeasurements were being taken, causing large outliers in the data.


4.4.7 Recommendations4.4.7.1 DC Resistivity4.4.7.1.1 Processing Recommendations

Data processing of these surveys could provide valuable data to improve the interpretationof the subsurface geology. The inversion process could begin within a similar method to thosepreviously described. The UBC inversion code provides reasonable interpretations that do notimpose the same restrictions as linear inversion programs. However, the usage of additionalinversion programs could improve the final result by reducing the possibility of inversion artifactswithin the inverted model.

4.4.7.1.2 Field RecommendationsFrom this interpretation of the DC resistivity survey we recommend conducting further local

surveys that focus on characterizing the observed conductivity anomaly at the student site. Inparticular, a survey that could delineate between ancient stream beds, anthropogenic sources,and faulting would provide useful information for further characterizing the subsurface anomaly.These surveys may include well boring to obtain various depth-dependent data, seismic surveyswith a deeper depth of investigation than the current shallow-depth seismic survey, and/orfurther Time-Domain Electromagnetic surveys that could further constrain the inversion of theDC resistivity data.

In the event of implementing further DC resistivity surveys at the same site, a tighter surveyspacing near the suspected conductive anomaly would likely lead to valuable results. Futuresurveys may also implement a Schlumberger array geometry, which could potentially improvethe horizontal resolution of the data.

4.4.7.2 Self Potential4.4.7.2.1 Processing Recommendations

In order to maximize the accuracy of the SP data, it is important to emphasize the need torecord the time of the recordings and to take tip to tip measurements as that information is vitalto correcting the data. Microsoft Excel seems to be a very useful program for processing SP data.Be sure to use the equations to relate the data rather than plugging in concrete values so thatsomeone else can look at the data and understand processing procedures.

4.4.7.2.2 Field RecommendationsIn the future, in order to maximize the usefulness of SP, choose a location that is not espe-

cially rocky, where the soil is accessible and has a little bit of moisture in it to ensure that thereadings are reliable. It would also be good to conduct these surveys away from other methodsand away from environmental factors that may be affecting the current in the subsurface, suchas the buried wires.

4.4.8 Results4.4.8.1 DC Resistivity

After processing the data on Reservoir Hill, we obtained the following results for DC resistivity:

4.4.8.2 Self PotentialAfter processing the data on Reservoir Hill, the following results were produced by SP:

4.4.9 Interpretation4.4.9.1 DC Resistivity

The student site contains an electrically resistive layer highlighted within the DC resistivitysections along survey lines S0, S1, and S2, seen in Figure 4.13. This resistive layer appears


Figure 4.13: Inverted data for the DC Resistivity surveys conducted along S0, S1, and S2, respec-tively, across Reservoir Hill.

to dip towards the east with a relatively shallow dip angle, which is consistent with the localgeologic interpretation.

There is an anomalous discontinuity present within this layer on the west side of ReservoirHill. This discontinuity indicates a comparatively conductive region, which may indicate theinfiltration of water or the presence of a different localized lithology. The potential upwellingobserved by the SP survey along S0 supports the claim that the anomaly genesis stems from theinfiltration of water.

When combining the S0, S1, and S2 inversion results in three-dimensions, in Figure 4.14,we observe a near linear continuity of the the apparent conductive anomaly. We suspect thatthe anomaly stems from a linear feature, which may come from a buried stream channel, a fault,or an anthropogenic source. Additionally, the local geology imposes further information withregards to these potential anomalous sources. The local geologic observations indicate a patternof regional extension, which is responsible for producing other nearby faults. Therefore, it ispossible that a normal fault occurs at this location. A normal fault would provide a conduit forsubsurface water flow, which would be consistent with both the SP and DC survey results.

4.4.9.2 Self PotentialThe student site had a good amount of moisture in the soil, allowing good contact between

the earth and the electrodes. Looking at the general trend of the data along S0, there is a clearincrease in voltage followed by a quick drop, as seen in Figure 4.15. The upward trend in the S0data indicates that fluid is flowing along the line, causing upwelling in the area. This means thatcations are moving upward, pushing the water up towards the surface and, in turn, pushing theearth up.

The second line, S1 on the student site was further south than S0. This trend, seen in


Figure 4.14: The 3D inversion of the DC Resistivity data collected along student site lines S0, S1,and S2 on Reservoir Hill.

Figure 4.16, just shows a general, linear increase in voltage along the line. It is important to notethe significant outlier appearing in the data around flag 62, which was due to the metal fence atthat location. There isn’t anything causing changes in this linear trend, which most likely meansthat there are no anomalous features along this line, indicating constant current flow. This lineis also showing signs of upwelling, but in a less concentrated manner which tells us that, if thereis a fault, it is likely farther north from this location.

In order to confirm this analysis the SP data needs to be overlain with the DC Resistivity datato see if both methods report similar fluid flow patterns. If they do, then it is likely that there issome sort of fault on Reservoir Hill directing the fluid flow in the area. Figure 4.17 above displaysthis idea. The DC data in this figure shows values characteristic of faulting on the eastern sideof the fault (around flag 40 on S0 and 60 on S1). At this location there clearly a conductive bodyresting below a resistive body interrupted by some kind of discontinuity. The remainder of theresistive body is offset from its original location, which indicates that this is likely the location offault.

4.4.10 ConclusionThe DC Resistivity and SP methods are often used together to verify the data sets collected

by each method. When overlaying the DC and SP data for the survey lines conducted acrossReservoir Hill, there is support for the idea that there is a fault in this area. Faults tend to redirectfluid flow, often causing upwelling at the location of the fault. On Reservoir Hill there is a cleartruncation of the conductive body on the eastern side of the hill as well as a break in the resistivelayer seen across the entirety of the hill. We suspect that the location of this fault is around Flag40 on S0 and Flag 60 on S1. The resistive body is constant through all of the DC survey sites,which could be an interesting location for further investigation to pinpoint the location of thefault more accurately. It would also be interesting to conduct surveys perpendicular to the threesurvey lines conducted on Reservoir Hill in order to create a more complete 3D inversion.


Figure 4.15: Self potential data for the student site, line S0.

Figure 4.16: Self potential data for the student site, line S1.


Figure 4.17: SP data overlain DC data in order to see how the trends of the two methods correlatealong the student lines on Reservoir Hill. From top to bottom: SP for S0, DC for S0, SP for S1, andDC for S1.

4.5 Gravity 179

4.5 Gravity

4.5.1 Objectives1. Collect high quality gravity data with the CG-5 Autograv Gravity Meter.2. Perform data corrections (instrument/tidal drift, latitude, free air, simple Bouguer slab,

complete Bouguer terrain, and geometry corrections) on the collected data.3. Generate models for the Reservoir Hill student site and the 2017 Main Line.4. Provide useful recommendations for future field camps based on experience in the field and

processing.

4.5.2 ExpectationsIt was hypothesized that there might be a fault running underneath reservoir hill. If the fault

was located within the sediments only there would be no anomaly present as there would notbe a significant offset for density contrasts in the bedding. However, if the fault extended tothe basement rock, an anomaly would be detectable as there would be a large density contrastbetween the underlying basement and overlying sediments. This large contrast would be theresult of offsets created by the fault, creating an anomaly that would be visible in the data.

4.5.3 Survey Maps

Figure 4.18: Student Site Gravity Survey


4.5.4 Data AcquisitionTwo Autograv CG-5 Gravity Meters were used to complete both gravity surveys. To begin each

data acquisition day a gravity reading was first taken at a base station. Each CG-5 was thentaken back to this base station every two-three hours to help account for instrument and tidaldrift. This information allows the assumption that the rate of change due to drift in this timeframe is linear, which is easily corrected in the processing stage. The base station was locatedat the first flag of the 2016 main line so that we could couple the 2016 and 2017 data sets inthis area if desired. A leapfrog method was used with the two instruments which resulted in areading being taken every 40m. A minimum of three 20 second readings were taken per station,with more taken if the standard deviations exceeded 0.02mGal or the measurements seemedinaccurate due to outside errors such as wind noise, nearby crews, etc.

Proper field setup consisted of placing the gravity meter on the leveling stand to within ±2arcsec. Once the gravity meter was leveled, the gravity measurement could be taken and the CG-5 would display all necessary information to be read off and saved on the instrument itself. Eachmeasurement was stored on the instrument, and the operator also recorded values for stationnumber, time (HH:MM:SS), value (mGal), and standard deviation (mGal).

4.5.5 ProcessingEach day, after field acquisition, the gravity data was taken off the CG-5 and quality checked

in Microsoft Office Excel. The best readings for each flag were noted for later use. These valueswere checked again and subjected to a number of corrections, the first being drift corrections.Correcting for drift requires correcting individual loops created by repeated measurements at abase station. The corrections were be done with a simple line equation:

gcorrected = gi − (g1 − g0t1 − t0

)(ti − t0)− g0 (4.2)

where gcorrected is the drift corrected gravity value, gi is the raw measurement needing correction,ti is the time of the raw measurement, g0 is the first base station measurement in the loop, g1is the final base station measurement in the loop, and t0 and t1 are the times those measure-ment were taken respectively. This equation was applied to each raw measurement in MicrosoftOffice Excel. Next, loops with differing base stations were shifted vertically to ensure that alloverlapping stations had the same value. Values for elevation were then acquired from a localdigital elevation model (DEM) using QGIS software. These values were then used for the latitude,free air, simple Bouguer correction, and complete Bouguer corrections. Geosoft’s geophysicalexploration software, Oasis Montaj 8.5, was used to do this.

From the corrected data, a GYM-SYS model was created. Theoretical models, created usingthe software, produced a gravity response. This model was displayed alongside the measuredgravity. By adjusting the model to match, the theoretical response was made to match the realresponse. Preliminary models were made using past models, geologic cross subsections, and welldata.

Corrections Reservoir Hill Student SiteDrift Correction XBase Station Tie-in CorrectionLatitude CorrectionFree Air Correction XSimple Bouguer Slab Correction XComplete Bouguer / Terrain Correction X

Table 4.1: Student Site Gravity Corrections: X indicates a completed correction.

4.5 Gravity 181

Formation Density (g/cc)Lewis Shale 2.4Mesaverde 2.57Mancos Shale 2.5Dakota Sandstone 2.34Morrison Formation 2.62Jurassic Sandstones 2.55Ancestral Rockies Sediment 2.5Low Density Zone 2.5Crystalline Basement 2.8

Table 4.2: Density Values of Geologic Layers: These values were used when modeling thegravity response of the main line and S0.

4.5.6 Errors and Uncertainties

4.5.6.1 Processing Errors

The only processing errors that occurred during gravity post-processing were the use of ter-rain files for each region. The terrain files used had an uncertainty of 10 - 25m which introducedsome error into processing results. The modeling process also has some uncertainty in termsof the final result because the model itself is non-unique. The final model was fine tuned andaltered many times in order to fit the data from other sections.

4.5.6.2 Field Errors

There were multiple instances of both of the CG-5 instruments being mechanically shocked.This happened during transport of the instruments on the days when measurements were takenon line S0 because of the irregular terrain that the vehicles drove over. Normally after transport,the CG-5 should be left to rest so the internals of the instrument can normalize.

As discussed for the S0 line at the Reservoir Hill student site, position inaccuracies cancreate significant errors in the processed gravity data. Any point with more than a ±3cm verticalaccuracy should be considered erroneous. Overall, on the main line, the positioning informationwas trusted and no points were omitted due to vertical inaccuracies. However, the S0 line at thestudent site still contains erroneous data points due to position inaccuracies.

4.5.7 Recommendations

4.5.7.1 Field Recommendations

It is the recommendation of the 2017 Geophysics Field Camp students that future classesattempt to gather more permitting around the areas of interest. With more permitting, gravitywould ideally be able to create a straight gravity line that is perpendicular to geologic strike.There would also be a possibility for a 3D grid to be set up in order to employ 3D inversiontechniques. It is also highly recommended that the GPS points gathered are within a verticalaccuracy of ±3cm.


Figure 4.19: Student Site Quality Control - Line S0.

4.5.8 Results

Figure 4.21: Legend for the following Gravity models for the student site.

4.5 Gravity 183

Figure 4.20: Student Site Processing - Line S0.

Figure 4.22: Preliminary model of Reservoir Hill based on well data and geologic cross subsection.Notice that the calculated model data does not trend well with the left side of the data.


’

Figure 4.23: Secondary model of Reservoir Hill which includes the fault speculated by DC methods.Notice that the calculated model data now trends well with the left side of the data.

4.5.9 InterpretationOverall, gravity results at Reservoir Hill were inconclusive on their own as approximately 20%

of the data had to be omitted due to vertical precision errors from the differential GPS. As shownin Figure 4.19, a vertical inaccuracy on the order of 5m creates an error of ±1.5mGal which iswhere the large error bars on the free air corrected data originate. To correct for the verticalprecision errors, we removed the data points that had larger than 1m uncertainty in elevation.Unfortunately, this removed about 20% of the data on the S0 line at the Reservoir Hill studentsite which required us to leave the data points with above a 10cm vertical precision. This is whythere are still errors on the order of ±0.5mGal seen in the final processed data in Figure 4.20.These ±0.5mGal errors in the final processed data coupled with significant differences in the realand forward modeling gravity responses hinders us from making any final decision about thepresence of a fault under Reservoir Hill. The model showing a fault present beneath ReservoirHill fits the data quite well, however we do not feel comfortable asserting that the model iscorrect. Instead, we will use the model to help support other models in their assertion that afault is present on the West side of the fault near flag 50.


4.6 Electromagnetics

4.6.1 Objectives

Our objectives and goals for TDEM surveys in Pagosa Springs are:

1) To image vertical conductivity changes and identify possible depths of different lithologicalunits (TDEM)

2) To image lateral conductivity changes running W/E across Reservoir Hill and along the mainsurvey line (FDEM)

3) To generate a depth-conductivity model useful in data processing for other methods

4) To locate high conductivity zones that might indicate the presence of fluids

5) To ascertain the existence of a potential fault on Reservoir Hill

4.6.2 Expectations

Data and information from previous DC resistivity surveys done in the Reservoir Hill area(PAGO02 and PAGO6 from the 2012 Field Camp) were viewed briefly prior to the completion ofthese EM surveys. The previous data shows a large conductivity contrast on the West side ofthe hill. Due to this prior information, students were expecting to find a change in conductivityvalues across the two survey lines S0 and S1 moving West to East. This information has pre-viously been interpreted as a fault running North-South along the Hill. This year’s goal was togain better insight as to if there is a fault in the area and where it is located. It was expectedthat the conductivity information obtained from the FDEM and TDEM surveys could be usedin conjunction with information obtained using other methods on the two Reservoir Hill surveylines to determine the location of, and characterize a potential fault and any groundwater flow inthe surrounding area.

4.6.3 Survey Maps

The majority of the EM surveys done in Pagosa Springs were performed on the student sitearound Reservoir Hill. A map of where the TDEM and FDEM surveys were done on the studentsite can be seen in Figure Figure 4.24.


Figure 4.24: Map of EM survey locations around the student site of Reservoir Hill. Red crossesrepresent locations were TDEM surveys were completed and the colored lines represent the surveylines in the area.

Above: TDEM surveys at the three different site locations. Red dots represent receiver looplocations and blue dots and lines represent the location of the transmitter loop.

Both EM-31 and EM-34 surveys were completed on the entirety of Line S0 and Line S1 onReservoir Hill. TDEM surveys were done at 3 separate locations, one on the hill and one to boththe East and the West of the hill. These locations were chosen because they were suitable fora 100m by 100m transmitter loop and because students wanted to see if there was a differencein the depth-resistivity model on different sides of the hill. TDEM Site 1 was located on thesoccer fields in Yamaguchi park, owned by the city of Pagosa Springs. This park is located westof Reservoir Hill and south of the Mother Spring. Both EM-47 and EM-57 instruments collecteddata at this location on May 19th 2017. An image of where measurements were taken at differentreceiver locations at this site can be seen in Figure Figure 4.25. TDEM Site 2 was located nearline S1 on public land on Reservoir Hill. This survey was completed using only the EM-57 onMay 23rd 2017. An image of the transmitter loop and receiver locations at this site can be seenin Figure Figure 4.26. TDEM Site 3 was located on private property to the East of Reservoir Hill.Only the EM-57 was used at this location. Data was gathered from this site on May 21st 2017.Receiver and transmitter locations for this site can be seen in Figure 4.27.

4.6.4 Data AcquisitionEM-31

The spacing between transmitter and receiver is fixed at 3.66m on the EM-31, making thisthe spacing for all of the surveys done using this instrument. After calibrating the EM-31, stu-


Figure 4.25: TDEM setup from 5-19-2017

dents selected between continuous measurements and measurements taken with each press ofthe orange button located on the transmitter tube. In the continuous setting, the orange but-ton still needed to be pressed at each station in order to mark the time and/or location of themeasurement. When progress down the line stopped, the ”Enter” button was used to pause andstart the survey again once progress continued.

EM-34After assembling the EM-34 and calibrating it using the quick-start guide, the transmitter coil

should be positioned at the first flag with the receiver coil positioned the length of the selectedreference cable away. In order to ensure that the coils were properly aligned prior to recordinga conductivity value, the ”Separation” value had to be somewhere between -300 and 300, prefer-ably as close to 0 as possible. Conductivity values were hand-recorded by the user. For all ofthe surveys done in Pagosa Springs, the transmitter and receiver coils were positioned in a ver-tical orientation in order to take advantage of the geometry of the horizontal dipole produced byvertical current loops. Horizontal dipole geometry provides for a shallower sounding whereas avertical dipole provides a deeper sounding and better detection of vertical anomalies.



TDEM

All three of the TDEM surveys completed in Pagosa Springs for this report were 100m x 100msquare loops. Initially, students ran the transmitter coil in one loop around a 100m x 100msquare, securing the corners by wrapping the cable once or twice around a plastic stake in theground. In a TDEM survey, the orientation of the square transmitter loop does not matter. Thetransmitter loop was then connected to the transmitter box for either the EM-47 or the EM-57depending on the survey being conducted. Next, the receiver loop for the respective survey (EM-47 or EM-57) was placed in the center of the square transmitter loop. For the EM-47 receiver, all3 components (x,y,and z) were used. A compass was used to line up the x component up in theNorthing direction and the y component up in the Easting direction. The receiver was leveled sothat the z component was parallel to the ground. The EM-47 receiver was then connected to thepre-amp to give the signal a boost and then to the ProTEM digital receiver. Similarly, for EM-57surveys the EM-57 receiver loop was leveled and connected to the ProTEM digital receiver. FigureFigure 4.29 shows the general set up for a TDEM survey using either the EM-47 or the EM-57.

The following tables provide information regarding the parameters used for the various TDEMsurveys done in Pagosa Springs.



Table 4.3: Site 1 EM-47 parameters

Parameter ValueGate selection 20

Rx area 31.4 m2 (The effective area of the receiver loop)Turn off time 8 µ sec (time it takes for current to shut off)

Repetition rate 285Hz (frequency), changed to 30HzIntegration time 15 sec (total time that the instrument stacks for)

Tx current 3 Amps (current sent through transmitter loop)

Parameters used for the EM-47 survey taken at Site 1. This survey was completed onMay 19th, 2017.


Figure 4.28: Diagram of EM-31 Operation [47]


Parameter ValueRx area 100 m2

Repetition rate 30Hz and 3HzTurn off time 100 µ sec

Tx current 20 Amps

Parameters used for the EM-57 survey taken at Site 1 (if different from previous surveys).This survey was completed on May 19th, 2017.


Parameter ValueTurn off time 113 µ sec and 50 µ sec

Tx current 30 Amps (with 113 µ sec T/O time) and 15 Amps (with 50 µ sec T/O time)

Parameters used for the EM-57 survey taken at Site 2 (if different from previous EM-57surveys). This survey was completed on May 23rd 2017.


Figure 4.29: General picture describing the set up on a TDEM survey


Parameter ValueTurn off time 86 µ sec and 50 µ sec

Tx current 20 Amps (with 86 µ sec T/O time) and 15 Amps (with 50 µ sec T/O time)

Parameters used for the EM-57 survey taken at Site 3 (if different from previous EM-57surveys). This survey was completed on May 21st, 2017.

4.6.5 ProcessingFDEM

Processing frequency-domain EM data involved dealing with EM-31 and EM-34 data sets.The initial process involved converting the G31 data files, obtained from EM-31 instrument, into.txt format. Furthermore, we converted the GPS and EM-34 data files, saved as .csv files, intoa .txt format. The ultimate goal from processing FDEM data was to acquire the best fittingcurve that minimized the residuals and describes the specific trend traced by the data points.In addition, it’s essential to realize that FDEM instruments only provide final measurements ofapparent resistivity and eliminate any intermediate calculations. FDEM only gives an averageresistivity value over the depth of investigation. This means that we are unable to see verticalconductivity changes through this method

Because FDEM instruments do not provide direct measurements about the Earth’s subsur-face as TDEM instruments do, a reliable geophysical inversion can not be carried out with FDEMdata to reflect a precise image of the subsurface physical properties. Nevertheless, we were ableto formulate the FDEM problem as if we were solving a horizon interpolation problem and intro-


duce different smoothing curves based on different model uncertainties. These fine curves areobtained from regularizing the problem by taking the second spatial derivative of the acquireddata. The underlaid assumptions here are that all data and model variables are independent,the physical operator that operates on the data is linear, and the uncertainty of the model anddata variables are independent.

A typical EM-31 data file contains information about the station number, conductivity mea-surements, in-phase measurements and the associated time. When EM-34 data was takenmanually, and it only contains the station number and conductivity measurements. In order toread and process the data, a mat-lab script was written that scans the data files with predefinedformat setup. For EM-34 data processing, filters were applied to all GPS data corresponding toodd flags along the survey line. Alternatively, for EM-31 data we needed to interpolate betweenevery two flags coordinates to fit the points taken at 5m intervals between flags.

The next step of processing involved creating the associated operator of the physical problemand introducing the least-square solution m = (GTWT

DWDG+WTMWM )−1(GTWT

DWDd) where G isthe constructed operator of 0s and 1s to match the data into their respective location, WM is theregularizing square matrix divided by the chosen model uncertainty, WD is the data uncertainty,d is the measured conductivity data, and m is the model that maps the smoothing curves. Theregularizing matrix R is just a diagonal matrix of the discrete-form of the second spatial derivativeto ensure a smooth curvature of the representative data points. In this regard, Figure Figure 4.30below reveals all possible representative curves, corresponding to different model uncertainties,for EM-34 conductivity data along S-1 line.

EM-34 Conductivity Data Representative Curves for S-1 Line

Figure 4.30: The graph reveals different representative curves for the apparent conductivity dataobtained from EM-34 along S-1 line. The different curves in the graph represent different levels ofmodel uncertainties.

After acquiring all representative curves, the last step required choosing a model that withdescribes the general trend with a moderate uncertainty value. To do this, we needed to introducethe notion of the L − curve which tries to represent the relationship between the data residualsand the model residuals. In general, residuals describe the difference between the actual andpredicted data points. In this case, each pair of data and model residuals corresponds to aknown uncertainty value. The ultimate goal is to choose the data point that achieves stabilitybetween residuals and extract the ultimate regularizing parameter, uncertainty value. In thisregard, Figure Figure 4.31 below reveals all residual pairs that correspond to the representativecurves in Figure Figure 4.30.


Constructed L-Curve

Figure 4.31: The L-curve that describes the relationship between model residuals and data residu-als. The x− axis represents the difference between the predicted model m and the reference modelm. Alternatively, the y − axis represents the difference between the predicted data Gm and thereal data d. In addition, the variables WM and WD represent the model and data uncertaintiesrespectively.

The previous process was repeated for each data set obtained from FDEM based surveys forthe instruments EM-31 and EM-34 along lines S-0 and S-1.

TDEMData collected in the field is initially stored in the ProTEM digital receiver. This data can

be downloaded in the form of a .Gx7 file and later converted to a .txt file to view the data.The first step in processing the data was to obtain a single value for each time stamp at eachreceiver location. Because the TDEM instrument stacks measurements for the duration of theset integration time, there are multiple measurements for each time stamp (often 10-20 for thesurveys taken in Pagosa Springs). Joe Capriotti created a Python code in order to calculate theaverage and standard deviation of the stacked measurements at each receiver location [48]. Thiswas done for both the low and high frequency measurements. This averaged data was convertedinto a .txt file and opened in Excel with tab delimiters.

In the Excel spread sheet, voltage values were multiplied by 104 and time values were multi-plied by 10−3 in order to scale the values to the appropriate format used in the inversion software.Also, percent error was calculated for each voltage value by dividing the standard deviation ofthe averaged measurements by the average voltage value obtained from Joe’s program.

After data at each specific receiver location was averaged and manipulated slightly in Excel, asoftware program named IX1D was used to invert the time domain data. Upon opening the IX1Dprogram, a new TDEM sounding was created. Next, the program asks the user to choose variousparameters such as array type, voltage units, and number of sweeps. For these parameters,”Fixed Loop TDEM Data”, ”nanoVolts per Amp-square meter”, and 1 were chosen, respectively.After entering this information a dialog opens up labeled TDEM time/Voltage Entry/Edit. In thisdialog, the user can enter coordinates for Easting, Northing, and elevation for the center of thetransmitter loop at the survey location. A value was also entered for Azimuth, or the angle thatthe transmitter loop is off from magnetic North. Next, a value was input for transmitter loop size(100m by 100m for our surveys). Directly below is the receiver coil position in reference to thecenter of the loop. It was assumed that East and North were in the positive x and y direction andthat West and South were in the negative x and y direction. After this, the survey parameters for


each sweep were entered. All of this information was normalized in the program written by JoeCapriotti [48]. The values that were entered for these parameters are Freq (Hz):0.4, Ramp (µ S):100, Receiver Coil Area (m2): 100, Current (A): 30, and Tx Turns: 1. Finally, the time, voltage,and % error data was entered in for each sweep. Any data points after the % error becomes largerthan 3 or before the ramp time were deleted and not included in the inversion.

After entering all the previous parameters, a screen opens up that includes a plot of receivedvoltage vs. time and the depth-resistivity model. First, the voltage-time plot was checked forany ”bad” data points and these points were deleted. To process the data the ”View VoltageCurve”, ”Smooth Model Display”, and ”Layered Model Display buttons had to be switched on.Next, the button with a red squiggly curve was pressed to estimate the smooth model. The boxeswere then checked for ”Occam’s Inversion” and ”Estimate models automatically This generatesa green computer-generated curve to fit the data. However, this curve was not good enough forinterpretation. Thus, the Edit Model tool was used to input how many layers we estimated therewas in the survey area subsurface. This generated a red curve which could be dragged around todifferent depths and resistivity values to best fit the model to the data. Once we had a model thatreasonably matched the data, the ”Multiple Iteration Inversion” or the ”Single Iteration Inversion”button was pressed to find a model based on our data points and the previously generated model.The idea behind this is to combine the model created based on the reliability of certain datapoints with any outside information used in the creation of the model, and the model that waspreviously computer-generated solely based off of the data. Before saving the model, the modelwas resized so that it displayed depths from 10 meters to 1000 meters. To save the models, theywere exported as an ASCII model file. This is in the format used by the DC resistivity group touse the TDEM data in their inversion. [49].

4.6.6 Errors and UncertaintiesThere are many sources of error and noise that could affect the inversion and interpretation

of the data seen in this report. This error can be associated with the survey design, instrumentsused, or with steps taken in processing.

Field Errors• Incorrectly oriented receiver loops may record data that does not represent the expected

component.• Transmitter current fluctuations during a measurement will change the amplitude of recorded

signals.• Poor connection in the cables may change the apparent resistivity in unexpected and un-

predictable ways.• Thermal variations from varying weather conditions can cause changes in equipment resis-

tance. This could potentially change the behavior of the receiver or transmitter loops.• Inaccurate or poorly designed transmitter loops can produce varying inversion results.• Equipment malfunctions such as reference cable delays or drifting crystal syncs may cause

the ProTEM Digital Receiver to record at the wrong times.

Field Noise• Man-made conductive bodies such as pipes, metal goal posts, etc... can induce currents

that interfere with desired data measurements.• Large metallic items may interact with electromagnetic fields and interfere with or mask

signal.• Natural solar storms and wind can induce magnetotelluric currents that may be recorded

by our receivers and impact the measured components of the magnetic field.• Strong magnetic rocks in the survey area can distort primary and induced EM fields.

Processing Errors• Incorrect interpretation of different resistivity layers seen in TDEM data.


• Incorrect number of expected layers in TDEM inversion• Over or under smoothing our FDEM data. Over-smoothing could result in aliasing and

missing important features in the data while under-smoothing may result in noisy datapoints which can lead to incorrect conclusions.

• TDEM data taken at site 1 was incredibly noisy. This means that fewer data points couldbe used to fit a model because much of the data had a high percent error associated with.This leads to a poorly constrained and less accurate model.

• Poor note taking in the field which leads to the incorrect input of parameters or certainsources of noise and error being left out.

• The IX1D inversion software produced multiple models that fit the data in the same way.This means it was up to our discretion to choose the model that we thought was the best.

• No presence of data uncertainty for FDEM data points so all points were weighted andtrusted the same.

4.6.7 Interpretation4.6.7.1 FDEM

EM-34 Conductivity Data Along S-0 Line

Figure 4.32: EM-34 conductivities in the Easting direction along line S − 0 at the student site onReservoir Hill. Red circles represent exact conductivity values, the black curve represents a specificsmoothing curve, and the red curve represents a general smoothing curve.

The interpretation of FDEM data mostly involved identifying and characterizing any trends thatcould be seen in the smoothing curves that were fitted to the data. The initial motivation fordisplaying two different curves was to see both general and specific trends in conductivitiesalong these lines. The ability to identify trends in the data on different orders of generalizationhelped to streamline the process of integrating EM with the results of other methods during theinterpretation process.

The data shows an overall downward trend in conductivity values, with the line S0 data having amore distinctive downward trend. The EM-34 data clearly supports this trend. The EM-31 datagenerally also supports the downward trend of the EM-34 data, however not as distinctly. Thiscould be due to noise or the fact that coil spacing on the EM-31 is fixed at 3.66m, making itsmaximum depth of investigation much shallower than that of the EM-34.

4.6.7.2 TDEMThese plots were interpreted to determine the depths of different geological layers in the area.

Figure Figure 4.39 shows how these layers were interpreted across the three different TEM sites.One important thing to note about this model is the less resistive Mancos Shale layer seen insites 2 and 3. This less resistive layer provides an indication of fluids at this depth that makesthis region of Mancos more conductive than above.




This interpretation produces the following model, seen in Figure Figure 4.40, adjusted to the rel-ative elevations and locations along a line at which the TEM surveys were taken. There are a fewimportant things to note about these models. First, for this model the TEM sites were assumedto be taken in a straight line. In reality, these surveys were taken in almost a triangular geome-try. This means that this model can not truly be interpreted in a line straight across ReservoirHill. Additionally, due to the depth of investigation of the EM-57 instrument, anything below500m depth for each individual survey is an inference of the resistivity and cannot considered asaccurate as measurements taken above this depth.One additional way that the TEM data was processed was in an attempt to see the depth dif-ference between the Mancos/Dakota boundary at different survey locations. Interface depthswere calculated given a dip of 5◦ NE and compared to the depth differences seen in the models.To calculate the expected difference, the apparent dip in the direction between the two surveylocations had to be determined. This was done using Equation Equation 4.3.

tanα = tanβ sin θ (4.3)

α = apparent dipβ = angle between line of interest and true dip direction

θ = true dip

Next, we used basic trigonometry laws and the difference between the two survey locations tocalculate an expected depth difference between the Mancos/Dakota boundary.

4.6.8 Conclusions4.6.8.1 FDEM

Geologic insight from a visit to Reservoir Hill showed that it primarily consists of MancosShale with a cap of alluvium at the top of the hill. Considering that the maximum investigationdepth for the EM-34 using 40m spacing is 60m, and that the depth of the Mancos Shale onReservoir Hill was determined to be around 200m, it is apparent that there is some change inthe lithology within this investigation depth. Shales tend to be more conductive than sandstonesand limestones, so it is possible that the Dakota Sandstone that lays beneath the Mancos Shaleor a layer of limestone concretions within the Mancos was shifted upward at some point on the




hill by a geologic event such as faulting or folding, resulting in a resistive anomaly closer to thesurface.

4.6.8.2 TDEMIt was found that between sites 1 and 3 there was an expected depth difference of 104 meters

and an actual difference of 107 meters, giving an offset of 3 meters. This amount of offset isnegligible due to survey and processing errors. Between sites 1 and 2 the expected differencewas 95 meters and the actual was 151 meters. This gives an offset of 57 meters. Last, betweensites 2 and 3 there was an expected depth difference of 0 meters, an actual difference of 44meters. These last two measurements are significant and could be indicative of a fault somewheresurrounding site 2. For both cases the actual depth difference is greater than expected. Thisprovides evidence of a normal fault with the West-most side as the hang wall. However, becauseonly 3 TEM surveys were completed these results are by no means conclusive. These resultsmust be integrated with other methods in order to gain a more thorough understanding of theReservoir Hill area.

4.6.9 RecommendationsOne thing that slowed the processing stage of this project was the lack of TEM data availabledue to minimal surveys being conducted in Pagosa Springs. The inversion software had a diffi-cult time working with data when the receiver loop was located outside the transmitter loop andthis was the case for the majority of the TEM measurements. Because of this, only 3 discretemeasurements had to be analyzed. This made interpretation significantly more difficult and lessreliable. We recommend taking more TEM surveys, perhaps in a more linear format. This will en-able students to better investigate how resistivity changes as a function of depth across ReservoirHill, giving a more 3 dimensional sense to the data. Another recommendation is to take moreTEM surveys on Reservoir Hill itself, as this is an area of interest in the Pagosa Springs area.For this project, only one TEM survey was completed on the hill. TEM is a useful tool becauseit has a high depth of investigation. More TEM surveys completed on the hill will hopefully givestudents a better understanding of the geologic structure in this area. Keeping the surveys awayfrom busy areas such as the town of Pagosa Springs will also help to improve data quality by mit-



Figure 4.35: EM-31 conductivities in the Easting direction along line S1 at the student site onReservoir Hill. Red circles represent exact conductivity values, the black curve represents a specificsmoothing curve, and the red curve represents a general smoothing curve.

igating sources of noise. Additionally, it would be extremely beneficial to perform TEM surveysin areas close to wells with wellbore data, or near MT survey locations. This will help correlatethe TEM data with other data sets to make integration and interpretation easier.

For future FDEM surveys it is recommended that surveys are completed in areas away fromsources of error such as power lines, magnetic fences, and other surveys. The main line datawas extremely noisy and was unable to be processed and interpreted because the survey wastaken near several conductive objects. A second recommendation is to be sure that the surveyortakes good notes on surrounding conductive anomalies such as water towers, vehicles, or piping.This will enable the processing team to correct for such errors and produce a result that is easierto interpret


Figure 4.36: TEM site 1, west of Reservoir Hill, May 19th, 2017. Left: Voltage decay throughreceiver loop over time on a log-log scale. Blue data points represent high frequency, purple pointsrepresent low frequency. Right: Inverted earth resistivity model. Shows resistivity as a function ofdepth on a log-log scale

Figure 4.37: TEM site 2, on Reservoir Hill, May 23rd, 2017. Left: Voltage decay through receiverloop over time on a log-log scale. Blue data points represent high frequency, purple points representlow frequency. Right: Inverted earth resistivity model. Shows resistivity as a function of depth ona log-log scale


Figure 4.38: TEM site 3, east of Reservoir Hill. May 21st, 2017. Left: Voltage decay throughreceiver loop over time on a log-log scale. Blue data points represent high frequency, purple pointsrepresent low frequency. Right: Inverted earth resistivity model. Shows resistivity as a function ofdepth on a log-log scale

Figure 4.39: Inverted resistivity models showing interpreted geologic formations at depth. Thesemodels are plotted on a log-log scale. Right: Site 1, May 19th, 2017. Center: Site 2, May 23st,2017, Left: Site 3, May 21st, 2017


Figure 4.40: Model showing relative location and depth of interpreted geologic layers taken at thevarious survey sites with the geologic cross section across Reservoir Hill underneath


4.7 Results and Interpretation4.7.1 Overview

Four methods–DC Resistivity, Self Potential, EM, and Gravity–conducted surveys on theReservoir Hill Student Site. The surveys conducted aim to re-explore the findings of the CSM2012 field camp and the results of the paper ”The Plumbing System of the Pagosa ThermalSprings, Colorado: Application of Geologically Constrained Geophysical Inversion and Data Fu-sion” by Dr. Andre Revil, Stephen Cuttler, and other CSM affiliates. Figure 4.41a, Figure 4.41b,and Figure 4.41c, shown below, give the locations of historic surveys conducted previously alongwith those conducted this year. These publications are of interest because they suggest the pres-ence of a normal fault on Reservoir Hill. Quantifying the faults in the area of Pagosa Springs isvery important because of its potential effect on the plumbing of the region. Faults can be anideal path for fluid to follow because they are more open and connected than the pores of therocks.

4.7.2 GeologyDuring the preliminary stages of field camp, the team worked together to create a preliminary

cross section of the student site, which can be found in the geology section of the paper. Theinitial cross section, figure 3.1, combined well data, seismic data, and surface measurements tomake a rough interpretation of the subsurface. Well data and surface measurements showed adiscrepancy in the depths and dips of the formations seen in the seismic data. This discrepancywas hypothesized by the fault that was believed to exist based on previous studies of the region,described in subsection 4.7.1. Because of regional geologic trends, a normal fault was assumed.After gathering the data from this year’s field session and more closely reviewing the seismicinterpretations of past field sessions, it became clear that the fault is most likely normal butwith the opposite orientation. This new interpretation, figure 4.6, required us to add anotherfault to reconcile the depths of the formations by keeping the dip constant so that the formationswere shallower on the West side of the line compared to the East. This lined up nicely with theinterpretations of previous field camps that believed a second fault could be seen southwest ofPagosa Springs along the San Juan river. With these two faults, the measured surface dips couldbe correlated well with the depths measured by the P-1 well and seismic data.

4.7.3 DC ResistivityDC resistivity held considerable significance on Reservoir Hill this year, with DC resistivity

crews taking surveys at three different sites along the student site: lines S0, S1, and S2. Theinverted data sets indicate a large resistive body characterized by a shallow, eastward dip witha prominent, relatively conductive discontinuity running through it. The geologic interpretationsof the region delineate the resistive body as Dakota Sandstone. A more conductive layer lies atopthe Dakota Sandstone. This layer is indicative of the Mancos Shale and is in concurrence withregional geology. The discontinuity within the Dakota Sandstone layer linearly correlates to allthree lines. The conductive nature of the anomaly likely indicates the infiltration of water.

At first glance, the anomaly observed replicates PAGO-02, a DC profile published in Dr. Re-vil’s paper, almost perfectly. PAGO-02 is shown for reference in Figure 4.42. Comparing theseresults further confirms past claims of a fault. These results have also been verified by theother methods–self potential, EM, hammer seismic, and gravity. The findings from each of theseresults are compared to DC findings below.

4.7.4 Self PotentialSelf potential surveys occurred across lines S0 and S1. As shown in Figure 4.43 and Fig-

ure 4.44, the results from these surveys support the findings proposed by DC resistivity. Bothlines, in particular line S0, show a transition from an increasing change in potential to a de-


creasing change in potential, a classic sign of upwelling. Upwelling occurs when upward fluidflow through the subsurface is intercepted by a physical boundary, which changes the directionof fluid flow. Evidence of upwelling coincides with the conductive anomaly found in the DC in-versions, further confirming the presence of a subsurface physical boundary on reservoir hill.Evidence of upwelling may also indicate that the anomaly affects fluid flow in the subsurface.

4.7.5 EMTwo different types of EM surveys, frequency domain EM (FDEM) and time domain EM

(TDEM), were taken on and flanking Reservoir Hill. Frequency domain surveys–utilizing bothEM 31 and EM 34 instrumentation–were taken along student site lines S0 and S1. Both sur-vey sites show a decrease in conductivity moving westward along the survey line, with S0 veryclearly illustrating this trend. Such a trend indicates that the Mancos Shale–a relatively conduc-tive body–is changing with respect to depth across the line. Additionally, on the eastern side ofboth survey lines S0 and S1, there is a highly conductive spike, anomalous to the rest of thedata points. Lithologic depth changes like this one, in conjunction with a conductive anomaly,also characterize the findings in the DC data. The correlation between these results for lines S0and S1 can be seen in Figure 4.43 and Figure 4.44.

The TDEM surveys, which were taken at three different locations on Reservoir Hill, furthersupport the claims proposed in the FDEM surveys. The vertical sounding profiles producedby TDEM show a significant offset, especially between sites two and three. Such results areindicative of geologic uplift, specifically a fault with a westward hanging wall. Additionally, thevertical sounding profile at site two aligns almost perfectly with the S1 DC data, yielding furthersupport for the results produced by the DC surveys. Figure 4.45 overlays the TDEM survey takenat site two onto the DC data to show this comparison.

4.7.6 Hammer SeismicHammer seismic surveys were conducted along the eastern parts of student site lines S0 and

S1. Hammer seismic’s relatively shallow depth of investigation focused mostly on calculatingand modeling the depth to the Mancos Shale. The results showed relatively consistent overbur-den thickness, with the exception of one anomaly. On line S1, at approximately flag 59, theoverburden thickness decreases significantly. This anomaly, shown in figure 4.9 aligns with theconductive anomaly found in the DC resistivity data, implying that the conductive anomaly isdirectly associated with lithologic change. The scale of the hammer seismic survey was too smallto overlay with DC data, however hammer seismic results can be viewed in section 4.3.8.

4.7.7 GravityA gravity survey was completed along line S0. The results were inconclusive due to severe

vertical precision errors, so little can be interpolated from the S0 gravity profile. It is worth noting,however, that two geologically referenced gravity models were made, one featuring a fault wherethe DC conductive anomaly is shown, and one without the fault. The model with fault fits thedata set much better, thus supporting the idea of a geologic boundary as the source of the DCanomaly.


4.8 ConclusionAll four methods surveyed on Reservoir Hill have indicated an anomalous presence on lines

S0 and S1, with nearby vertical soundings providing additional support. The anomalous bodycan be traced from flag 40 on line S0 to flag 60 on S1. This infers a linear physical feature, suchas a fault, buried stream channel, or anthropogenic source. When comparing the DC resultson line S1 to the finalized geologic cross section, the results match well. This discerns thatnormal faulting should be seriously considered as the source of the conductive discontinuity.The comparison is shown in Figure 4.47. There also seems to be a very clear disruption in thesubsurface saturation, meaning that if this fault exists, it is directly influencing the fluid flowingin the area. The exact nature of its influence once the fluid passes the fault is still unknown andthere still isn’t evidence that this particular fault is a source of the hot water, but further studyof the region can hopefully provide a more complete explanation.

The ultimate goal for the student site was to better characterize the anomalies seen by pre-vious studies in the area. These anomalies are geologically important because faulting in theregion could potentially be influencing, if not driving, the fluid movement in the area of PagosaSprings. This correlation between the results of the varying methods strengthens the credibilityof the data, allowing for a more confident interpretation of what may be happening in this area.

4.9 RecommendationsSeveral recommendations are suggested to continue this study and improve the accuracy.

The first recommendation is to conduct a survey perpendicular to the the lines conducted withinthis report. Previous years field camps in this area have conducted surveys that are essentiallyparallel with this years survey lines. It is difficult to characterize a fault and fluid flow with asurvey only extending in one direction, as faults can look very different by moving in differentdirections. A survey which shows a perpendicular direction can help to confirm the subsurfaceanomaly in addition to providing additional information. Along with this recommendation, weadvise a grid survey that centers around flag 60 of line S1. This will allow for a 3D inversion ofthe data that will ultimately provide a better quality image of the subsurface.

Borehole geophysical methods are, unfortunately, outside of the scope of the Colorado Schoolof Mines Field Camp capabilities, but we believe that a series of exploratory wells could providekey information in this complicated system. We’d specifically like to see a well drilled north westof Reservoir Hill, south east of Reservoir Hill, and through the anomaly itself. These positionswere chosen based off of the speculated fluid flow through the region.


(a) A map of the 2017 field camp survey lines on Reservoir Hill

(b) A map highlighting the survey lines discussed in the paper by Andre Revil et al [37].

(c) A map of the 2012 field camp Hot Springs Area survey lines [45].

Figure 4.41: Figure 4.41c and Figure 4.41b show the survey parameters of the historic DC resistiv-ity surveys collected on Reservoir Hill. Figure 4.41a shows the survey parameters for each of themethods conducted the 2017 field camp.


Figure 4.42: A conventional inversion of DC Resistivity Line PAGO-02. Data for this line wascollected at the 2012 field camp, and was later processed and published in Dr. Revil’s paper. Notethe relatively conductive anomaly at distance x=1000m. This anomaly appears in almost the exactsame section as the anomaly found in the S1 DC profile. DC line S1 can be found as a part ofFigure 4.44


Figure 4.43: Plots comparing the SP, EM31, EM34, and DC resistivity results on line S0. The reddashed lines show correlations between the SP and EM methods. These correlations line up nicelywith the conductive and resistive bodies in the DC Resistivity inversion. Specifically, the spikein the data seen in SP, EM-34, and EM-31 line up with the conductive body on the west side ofReservoir Hill. This is followed by a drop in the data which lines up with the beginning of theresistive body a little bit farther to the east.


Figure 4.44: Plots comparing the SP, EM31, EM34, and DC resistivity results on line S1. The reddashed lines show correlations between the SP and EM methods. These correlations line up nicelywith the conductive and resistive bodies in the DC Resistivity inversion. Specifically, there is aclear drop in data in the SP, EM-34, and EM-31 where the resistive body begins, a spike wherethis body ends and transitions into a more conductive zone, and another drop where the secondresistive body begins.


Figure 4.45: A plot showing the DC inversion from line S1 overlayed with the vertical profile takenat TDEM site two. The sounding was taken in the middle of the conductive anomaly and fits theresulting DC inversion perfectly, further confirming the credibility of the DC results.

Figure 4.46: A comparison between the gravity model and line S0. The model aligns a fault tothe anomaly of the DC survey. Positioning the fault here decreases the associated error with thegravity data.


Figure 4.47: A comparison of the DC data to the finalized geologic cross section. The area of theDC survey is outlined with a blue rectangle.

5. Appendix

Contents


5.1 Report on Investigation of Local Wells in Pagosa SpringsIt is important to not that this summary was executed based purely on second-hand reports

and secondary interpretation of published papers. This paper was completed during data pro-cesing for the remainder of the report, and does not necessarily reflect the findings of this reportand those cited. It is an exercise purely for establishing a general foundation on which to builda knowledge base regarding the Pagosa Springs geothermal system.

5.1.1 Verification of Local ReportsOne of the missions upon returning to CSM what to gather as much additional information

to corroborate the local anecdotes from Pagosa Springs as possible. The first place to start wasthe Colorado Oil and Gas Conservation Commission (COGCC), which has the most completeinformation on drilled wells for Archuleta County. According to JR Ford, he had six wells thathad been previously drilled on his property by an oil and gas exploration company in 1951. Ifany records from those locations from that time existed, the well information could be found onthe COGCC interactive map. There was only one record in the region that corresponded with oneof the locations provided by JR. Well Fee #1 is a wildcat well drilled in January 1941, operatedby someone named K.L. Kendrick [50]. Fee #1 roughly corresponds with the location of JRs thirdnorthernmost well, seen in Figure ??. Comparing the wildcat status of this well to the shape ofthe locations of the other wells according to JR, it was then a reasonable assumption to makethat those wells were also wildcat wells and corresponded with the same time frame as Fee #1.The wildcat wells existence provides support to JRs report, particularly with how close the wellstrue location is to where JR drew it very roughly on a map. The only real discrepancy between theonline information and JRs report is the difference in time, 1941 [50] versus 1951 as reportedby JR. This is a perfectly reasonable error, and doesnt severely undermine the information.

The second stage of information gathering was to conduct a well permit search through theColorado Division of Water Resources. The first search was to discover water well permitting

212 Chapter 5. Appendix

information in a half mile radius around Fee #1, with the hope of discovering if any of the otherwells had be instead permitted as a water well after failing to find oil in 1941. This search hadno luck, so a second attempt was made to find information by instead searching for well permitsby JR. This search did pan out, and returned a permit for a well he drilled in July 2001 [51]. Thedate of this permit corresponded generally with when JR reported his wells started drying out,beginning roughly at the turn of the century and finally completely drying out in 2002. JR hadstated that he had 6 excellent and flowing water wells on his property, so if all wells had stayedhot and flowing there would have been no need to drill a seventh well for personal use. This iscorroborating evidence backing up JRs claim that the 6 exploration wells were drying out andthat water flow in the region dropped.

Unfortunately, that was the extent of the information available through well permits becausethere was very little information available online. The search was expanded to include waterrights, because while water rights might be on a larger scale they can still say something aboutthe behavior of geothermal water flow in the region. There was only one result for water rightsin Pagosa Springs for the Pagosa Hot Water Well. The P-1 well is located just across the SanJuan river from the Mother Spring, and a few of its properties were described in various papers.The Colorado Water Court documents provided information on the original well application anddecree from the 1960s, 2 well inspection reposts from 1978, and a map detailing the wells exactlocation. This information was extremely useful, and led to the discovery that usage in January1978 was 1.5 times greater than in March of the same year [52].

5.1.2 Analysis of Literature on Pagosa SpringsThe information provided above does little to truly characterize the geothermal system in

Pagosa Springs. More information was sought, particularly with regard to any and all well datadescribing water and water temperature as close to the Pagosa Springs area as possible. FieldCamp 2017 was graciously granted access to a paper written by Paul Morgan that attemptsto characterize the geothermal behavior of Pagosa Springs. However, since the paper remainsunpublished, secondary corroboration for its information was sought out in the forms of the1980 technical report of Pagosa Springs geothermal system written by Michael Galloway, andthe 1994 geothermal assessment of Archuleta County and Pagosa Springs written by Fraser Goffand Janice Tully.

For his report, Galloway conducted borehole measurements of temperature, pH, conductivity,and discharge at the O-2 and P-1 wells, located in downtown Pagosa Springs. These particularwells are close in proximity to the Mother Spring, with O-2 located 450 feet away and P-1 located405 feet away. According to the temperature logs, O-2 reaches a peak temperature of 57 degreesCelsius at 30 meters depth [53]. The P-1 well also has a temperature peak at around 30 metersdepth, but reaches its true peak temperature of 61.5 degrees Celsius at approximately 90 metersdepth [53]. Both wells increase in temperature until they reach their maximum, and then thetemperatures overturn. Galloway hypothesized that the two wells demonstrated evidence of twothermal zones, one in the Mancos Shale and one in the Dakota Sandstone. The single thermalzone in the O-2 is located solidly within the Mancos Shale, while the P-1 well has a thermalzone in both the Mancos Shale and the Dakota Sandstone[53]. Galloway also concluded that thesource for the water most likely comes from meteoric water [53] that has been heated at depth,and that the water travels up fissures in the Precambrian basement to the surface to fuel thehydrothermal system.

The report written by Goff corroborates the story painted by Galloway. He conducted sig-nificant geochemical studies of the water in the region, and determined that chemistry of thewater is consistent with that of meteoric water that has been heated at depth [54]. The paperhypothesizes a potential mechanism for the source of the hot springs based on this conclusion.The meteoric water sinks to the Precambrian basement where it is heated, and then rises backto the surface. There is potentially a main fissure that feeds the Mother Spring that funnelsthe water directly from depth and produces the highest temperatures seen on the surface. Thewater also spreads out to secondary fissures and along formation interfaces and mixes with the


groundwater, producing the varying chemistry and temperatures radiating out from the centersource [54]. The most relevant portion of the report to current field camp investigations are theanalyses of the Stinking Springs, which was close to the 2014 Field Camp main line, and thatof the Pagosa Springs aquifer. The results show that each source is comparatively different fromone another [54], implying that the various hot springs systems of Pagosa Springs and Chromoare most likely not connected.

The information from the reports was analyzed, particularly with regard to temperature infor-mation, and then combined tp paint a picture of the geothermal system in the Pagosa Springsarea. That information was then expanded upon thanks to the efforts of Paul Morgan. He cor-roborated with Pagosa Verde, Amax, and the Colorado Geologic Survey to expand the amountof information on the thermal gradients of the Pagosa Springs area. While the data is still com-paratively sparse, with little to no information available for the region located to the northwestof Pagosa Springs [55], the recontoured map shows relatively the same anomaly as presented byGalloway and agreed upon by Goff [55]. Morgan also agrees with Goffs mechanism of fluid trans-port [55], with a large central pipe feeding the Mother Spring and other water seeping to forma geothermal reservoir in the Dakota Sandstone. He then takes the information a step farther,and asserts that the Mother Spring is unique and exhibits few signs of being a fault controlledfeature [55].

5.1.3 Analysis and HypothesisBefore commencing studies, Field Camp 2017 was presented with two potential options for

the source feeding the Mother Spring. The first has been discussed, above, with a fractured base-ment feeding heated water up to the surface and pooling in reservoirs around Pagosa Springs.The unfavored option also mentioned is that the water originates near 8-Mile Mesa and travelsfrom depth up through various formations and faults up to form a reservoir in the Dakota Sand-stone and feed the Mother Spring. The reason this method is unlikely is that there is little to nosurface geology indications of spring deposits to support the fractured flow theory. A third char-acterization for the Mother Spring, based on a thought presented by Paul Morgan, also exists.Morgan mentioned that the flow seen in the P-1 well and by extension the Mother Spring couldbe the result of a convection flow in the subsurface. This could mean that the source feeding theMother Spring and its direct surroundings is a completely separate system than that of PagosaSprings. The hypotheses are many, but actual evidentiary support is small. Field Camp 2017 en-countered the presence of a fault in the basement along its main line, and evidence for a secondfault perpendicular to the lines on the student site. While more information is clearly needed toactually characterize the nature of the geothermal system in Pagosa Springs, it can however beargued that based on well informations the Mother Spring and direct surroundings behave in amanner unique even within the Pagosa Springs geothermal system.

Research is ongoing for this project, particularly expanding the knowledge base for the prin-cipal wells in and around Pagosa Springs. Well information was gathered on the P-1, O-2, TG-1,TG-3, TG-5, and Brown Federal Wells, with the thought that these wells are all tied to the geother-mal system in Pagosa Springs. This geophysical information, including temperature gradients,sonic velocities, and lithology lists, from around the Mother Spring can create an extrapolatedimage of what the geology directly beneath Pagosa Springs is like. Understanding the geology indetail allows for modeling of secondary systems, such as hydrothermal fluid flow. Much is stillnot known about what lies directly under Pagosa Springs, and with limited access for student-run surface geophysics the best information available to create that geologic model comes fromwell data.

5.1.4 Accomplishments and RecommendationsField Camp 2017 is responsible for creating the first masterlist for local well information in

Pagosa Springs. This was done in the hopes that having all the information in one place willassist future field camps in finally successfully characterizing the nature of the Pagosa Springs

214 Chapter 5. Appendix

geothermal system, and by extension the nature of the Mother Spring. A master list will alsoenable future field camps to begin expanding local well knowledge, instead of having to rediscoverall the same information every summer. The people who live in Pagosa Springs have a far greaterunderstanding of the local system than students who spend two brief weeks in the region, andusing that knowledge will improve the accuracy of data interpretations. A recommendation forfuture field camps is to have a dedicated team of students actua lly go out into the communityand talk to locals to learn whatever information they can on whats happening beyond whatcan be seen on the survey. These people can be assigned, but also can be volunteers suchas those involved in community outreach. Even with the limited knowledge gained this year,Field Camp 2017 has already benefited dramatically from learning what exactly is happening inPagosa Springs.

Bibliography

[1] M. Galloway, “Hydrogeologic and geothermal investigation of pagosa springs, col-orado”, Denver, CO, Tech. Rep., 1980, Accessed:2017-05, pp. 49–54 (cit. on pp. 22–30, 80, 157).

[2] V. Kelly, “Pre-cambrian rocks of the san juan basin”, in Guidebook of the sanjuan basin- new mexico and colorado- first field conference, Accessed: 2017-05, NewMexico Geologic Society & New Mexico Bureau of Mines and Mineral Resources,1950 (cit. on p. 22).

[3] R. Anderson and D. Kirkland, “Origins, varves, and cycles of jurassic todilto forma-tion, new mexico”, Bulletin of the american association of petroleum geologists, vol.44, pp. 37–52, 1 1960, Accessed:2017-05 (cit. on p. 23).

[4] R. OSullivan, “The jurassic wanakah and morrison formations in the telluride-ouraywestern black canyon area of southern colorado”, U.s. geological survey bulletin,1924, Accessed:2017-05 (cit. on p. 23).

[5] USGS. (1979). Dakota sandstone and burro canyon formation. Accessed: 2017-05(cit. on p. 23).

[6] K. C. M. William A. Cobban, “Structural map of pagosa springs”, 2011, Accessed:2017-06 (cit. on p. 24).

[7] E. Beaumont and C. Read, “Geologic history of the san juan basin area”, in Guide-book of the san juan basin- new mexico and colorado- first field conference, Accessed:2017-05, New Mexico Geologic Society & New Mexico Bureau of Mines and MineralResources, 1950, pp. 49–54 (cit. on pp. 24, 26, 29).

[8] J. English and S. Johnston, “The laramide orogeny: What were the driving forces?”,International geology review, vol. 46, pp. 833–838, 2004, Accessed:2017-05 (cit. onp. 25).

[9] S. Lucas et al., “Stratigraphy and age of the upper cretaceous fruitland formation,west-central san juan basin, new mexico”, New mexico museum of natural history andscience bulletin, 2005, Accessed:2017-06 (cit. on p. 25).

216 BIBLIOGRAPHY

[10] USGS, “Geological survey professional paper”, Ohio State University, Tech. Rep.,1949, Accessed: 2017-05 (cit. on p. 29).

[11] “Geophysical investigation of the pagosa springs geothermal system”, ColoradoSchool of Mines, Tech. Rep., 2016, Accessed:2017-05 (cit. on pp. 31, 32, 158).

[12] “Geophysical investigation of the pagosa springs geothermal system”, ColoradoSchool of Mines, Tech. Rep., 2013, Accessed: 2017-05 (cit. on p. 29).

[13] I. S. Moeck, “Catalog of geothermal play types based on geologic controls”, 2014,Accessed:2017-06 (cit. on p. 29).

[14] P. J. Vrolijk, “Fault-seal analysis using a stochastic multi-fault approach: Aapgbulletin”, 2004, Accessed:2017-06 (cit. on p. 30).

[15] “Ground water atlas of the united states”, 2011, Accessed:2017-06 (cit. on p. 30).

[16] G. Mavko, Conceptual overview of rock and fluid factors that impact seismic velocityand impedance. Stanford Rock Physics Laboratory, pp. 73–112. [Online]. Available:https://pangea.stanford.edu/courses/gp262/Notes/8.SeismicVelocity.pdf (cit. on p. 36).

[17] P. Seismic. (). What is a seismic survey?, [Online]. Available: http://www.parkseismic.com/Whatisseismicsurvey.html (cit. on p. 37).

[18] U. of Nairobi. (). Section 6-3: Seismic methods, [Online]. Available: http : / /learning.uonbi.ac.ke/courses/SPH311/scormPackages/path_3/section_63_seismic_methods.html (cit. on p. 37).

[19] B. S. Z.W. Edward Wooley Ting-Li Lin, “The role of local soil-induced amplifica-tion in the 27 july 1980 northeastern kentucky earthquake”, Environmental andengineering geoscience, vol. 14, no. 4, pp. 267–280, 2008 (cit. on p. 38).

[20] (). Electromagnetics, [Online]. Available: http://www.naevageophysics.com/Electromagnetics (cit. on p. 42).

[21] J. Johansson and U. Lundgren. (). Theory. [Online]. Accessed: 2017-05-31, [Online].Available: http://www.dannex.se/theory/2.html (cit. on p. 42).

[22] (). What are electromagnetic fields. [Online]. Accessed: 2017-05-31, [Online]. Avail-able: http://www.who.int/peh-emf/about/WhatisEMF/en/ (cit. on p. 43).

[23] (). Electromagnetic induction explained. [Online]. Accessed: 2017-05-31, [Online].Available: http://www.geophysical.com/whatisem.htm (cit. on p. 43).

[24] (). Frequency-domain em induction. [Online]. Accessed: 2017-05-31, [Online]. Avail-able: http://zonge.com/geophysical-methods/electrical-em/frequency-domain-em/ (cit. on p. 43).

[25] (). Electromagnetic methods. [Online]. Accessed: 2017-05-31, [Online]. Available:http://www.geophysical.biz/electromagnetic-method.htm (cit. on p. 43).

[26] J. Klein and J. Lajoie. (1980). Electromagnetic prospecting for minerals, practicalgeophysics for the exploration geologist. [Online]. Accessed: 2017-05-31 (cit. onp. 44).

[27] (). Geonics em-31. [Online]. Accessed: 2017-05-31, [Online]. Available: http://www.expins.com/item/geonics-em-31 (cit. on p. 44).

BIBLIOGRAPHY 217


[29] W. E. Wightman et al. (2003). Application of geophysical methods to highway relatedproblems. Publication No. FHWA-IF-04-021. (cit. on p. 44).

[30] (). Lenz law of elecromagnetic induction. [Online]. Accessed: 2017-05-31, [Online].Available: https://www.electrical4u.com/lenz-law-of-electromagnetic-induction/ (cit. on p. 45).

[31] M. Nabighian and J. Macnae. (1991). Time domain electromagnetic prospectingmethods. Publication No. FHWA-IF-04-021. (cit. on p. 45).



[34] (). Self potential (sp) basic concept, [Online]. Available: https://archive.epa.gov/esd/archive-geophysics/web/html/self-potential_(sp)_method.html (cit.on p. 48).

[35] S. Geotechnical. (). Self potential (sp) surveys, [Online]. Available: http://www.geophysical.biz/self-potential.htm (cit. on p. 48).

[36] Revil et al., “Review: Some low-frequency electrical methods for subsurface chara-terization and monitoring in hydrogeology”, Hydrogeology journal, pp. 617–658,2012 (cit. on p. 49).

[37] Revil, “The plumbing system of the pagosa thermal springs, colorado: Applica-tion of geologically constrained geophysical inversion and data fusion”, Journal ofvolcanology and geothermal research, vol. 299, 2015 (cit. on pp. 49, 50, 205).

[38] J. Blevins. (2011). Pagosa hot spring the deepest in the world, [Online]. Available:http://blogs.denverpost.com/thebalancesheet/2011/08/17/pagosa-hot-spring-the-deepest-in-the-world/118/ (cit. on p. 49).

[39] A. Samsudin. (N.A.). Section 2: Gravity surveying (cit. on p. 53).

[40] R. E. Sheriff, “Encyclopedic dictionary of applied geophysics geophysical referencesseries”, 2002 (cit. on p. 59).

[41] (). Gtis nuseis wireless seismic node, [Online]. Available: http://geophysicaltechnology.com/wp-content/uploads/2016/02/nru-1c.png (visited on 06/07/2017) (cit. onp. 60).

[42] (2017). Trimble gps tutorial - why gps? Accessed:2017-06 (cit. on p. 61).

[43] “Archimedes principle”, 2011, Accessed:2017-06 (cit. on p. 74).

[44] D. L. Mink, “Plumbing and political will: Low temperature geothermal power ex-ploration in pagosa springs, colorado”, Tech. Rep., 2015, Accessed:2017-06-06,pp. 429–436 (cit. on p. 84).

[45] 2. G. F. Camp, “Geophysical investigation of the geothermal system in pagosasprings”, pp. 66–87, 2016 (cit. on pp. 93, 205).

[46] T. A. Drean, Velocity trends in cretaceous rocks in wyoming laramide basins. WSGS,2012, pp. 1–57. [Online]. Available: http://www.wsgs.wyo.gov/products/wsgs-2012-ri-62.pdf (cit. on p. 93).

218 BIBLIOGRAPHY

[47] M. Kokoris. (2003). Nasa em31 project diagram - north pole expedition 2003.[Online]. Accessed: 2017-06-2 (cit. on p. 190).

[48] J. Capriotti. (2016). Protem data viewer. python program written to average stackedTEM57 .Gx7 data at each site (cit. on pp. 193, 194).

[49] K. Lewis. (2017). Ix1d quick start sheet. Guide to running the IX1D inversionsoftware (cit. on p. 194).

[50] (). Well information - fee #1. Accessed: 2017-06, [Online]. Available: http://cogcc.state.co.us/cogis/FacilityDetail.asp?facid=00705107%5C&TYPE=WELL(cit. on p. 211).

[51] (2007). Summary of permit file. Accessed: 2017-06 (cit. on p. 212).

[52] (2015). Water rights for pagosa hot water well. Accessed: 2017-06 (cit. on p. 212).

[53] M. Galloway, “Hydrogeological and geothermal investigation of pagosa springs,colorado”, Colorado Geological Survey, Tech. Rep., 1980, Accessed: 2017-06 (cit. onp. 212).

[54] F. Goff and J. Tully, “Geothermal assessment of archuleta county and the pagosasprings aquifer, colorado”, in Transactions, vol. 18, Accessed: 2017-06, GeothermalResources Council, 1994 (cit. on pp. 212, 213).

[55] P. Morgan, “Origins and geothermal potential of thermal springs in archuleta county,including pagosa springs, colorado, usa (revisited)”, Accessed: 2017-06, 2017 (cit.on p. 213).

abstract - today at minesinside.mines.edu/userfiles/image/geophysics... · abstract this report...

Documents