lccs land cover classification system general background
TRANSCRIPT
Land cover classification system
LCCS is a comprehensive methodology for description, characterization, classification and comparison of most land cover identified anywhere in theworld, at any scale and at different levels of detail.
WATH IS:
Land cover classification system
LCCS is a new world wide “language” to allow a rational use and an easy exchange of land cover informationbetween different countries, institutions, end users.
In the last 10 years the GIS capabilities (spatial analysis)became widely diffuse at any level.
GIS software's cheaper, more powerful and user friendly.
Spatial analysis has became an essential tool in Natural Resources assessment, modeling and planning.
WHY A NEW CLASSIFICATION SYSTEM?
Land cover classification system
• strong need of new data• problem of data integrability and flexibility•problem of miss use of the data
Derived consequences:•gigantic increase of digital data users•increased necessity/possibility of data exchange betweendifferent users and disciplines
Today the value of a data-set not only depends by the quality and the amounts of details present but also bythe level of flexibility, integrability with other type of dataand standardization of the information.
LCCS ALLOWS THE PRODUCTION OF DATA FULLY COMPATABLE TO THE GIS ENVIRONMENT
Land cover classification system
Land cover classification system
Major advantages of the method
•LCCS gives to the user the possibility to define a large number of land cover features (app.100.000)
•despite the large number all the classes have a verydistinctive and unambiguous meaning
•LCCS create an absolute level of standardizationbetween the different users
Land cover classification system
Major possible applications:
•all the land cover/natural resources mapping activities at any scale in any part of the world
•bridging system between land cover mapsthat use local or not standardized legends
•possible link between small scale global mapping activities and detailed large scalenational mapping programs
•bridging system to link natural resources statistical census with mapping programsat national or regional level
Land cover classification system
Background:
• LCCS is the result of many years of efforts of FAO (Food and
Agriculture Organization) and UNEP(United Nation Environmental Program) to harmonize data collection and management and to take a step towards an internationally agreed reference base for land cover and and land use.
•The initial concept of LCCS was discussed and endorsed by an international working group on Classification and Legend in Senegal (1996). More than 60 experts from different countrieswere present.
Land cover classification system
•While is fully development (1997-1999) the system links withother international ongoing initiatives:
FDGC (U.S. Federal Geographic Data Committee)
ECWG(U.S. Vegetation Subcommittee and Earth Cover Working Group)
USDA Forest Service
IGBP (International Geosphere-Biosphere Programme)
LUCC (Land Use and Land Cover Change)
South African National Land Cover Database Project
Land cover classification system
Results:
•Since is publication (November 2000) more than 1300 copies of the manual/software has been distributed to international and national institutions all over the world.
•LCCS is the Land Cover Classification endorsed by FAO andUNEP.•It has been adopted by the IGBP network in the LUCC program.
•It has been adopted by JRC for the GLC2000 initiative
•More than 8 millions of Km are currently mapped at 1:200000scale with this method under the Africover program.
• Many translation exercises of old legends in LCCS are ongoing.
2
Land cover classification system
•The following workshops/seminars on the use of LCCS havebeen or will be undertaken during this year:
IGBP-GLUCC seminar (Ispra, October 2000)GLC2000 seminar on classification and legend (Ispra, December 2000)Workshop/seminar in USGS (Sioux Falls, May/June 2001) E.C.A. Regional Conference (Addis Ababa, Sept 2001)Africa - GIS Conference (Nairobi, Oct. 2001)
Workshops in New Zealand and Japan (date to be established)
From next month a web page fully dedicated to LCCS users will beestablished
Land cover classification system
Future actions:
The FAO will assure a regular update of the concept/software.
The release of the new software version is expectedat the beginning of the year 2002.
The new version will be done in a new programming language, and will include some modifications asindicated by the user community.
The compatibility between the different LCCS versionswill always be maintained
Planning
Wild life
UNIQUE SOURCE OF BASIC ENVIRONMENTAL INFORMATION
Standardization process - Comparability- Increasing efficiencyStandardization process - Comparability- Increasing efficiencyCost reduction - Easy to update - Conform international standardsCost reduction - Easy to update - Conform international standards
MULTIPURPOSEAFRICOVER
DATABASE FORENVIRONMENTAL
RESOURCES
Rangeland Agriculture
Forestry
Land cover classification system
STUDY AREAINTERPRETATION RESULT
SPATIAL DATABASE GEN.
SPATIAL ANALISYS
LOST OF INFORMATION
CONSTANT AMOUNT OF INFORMATION
Land cover classification system
CLASSIFICATION/LEGEND
CARTOGRAPHIC STANDARDS
INTERPRETATION PROCESS
GENERALIZATIONSPATIAL TEMATIC
INH
ER
EN
T C
HA
RA
CT
ER
IST
IC O
F T
HE
ST
UD
Y A
RE
A -
TH
E T
RU
TH
RE
SU
LT
OF
TH
E P
RO
CE
SS
INFORMATION EXTRACTION PROCESS
Land cover classification system
LEGEND
INHERENTCHAR. AREA
SCALE
CLASSIFICATIONSYSTEM
INTERPRETATIONPROCESS
ACCURACY
DATA TYPE- ANC.DATA- FIELD WORK- INT. SKILL
INFORMATIONCONTENT
CARTOGRAPHIC STANDARDS
TIMEBUDGET
Land cover classification system
Land cover classification system
NEW CONCEPT OF MULTI USER ORIENTED NATURALNEW CONCEPT OF MULTI USER ORIENTED NATURAL RESOURCES DATA BASERESOURCES DATA BASE
CLASSIFICATIONSYSTEM
INTERPRETATIONSOFTWARES
ACCURACYANALISYS
MULTI USER DATA BASEBROWSER
INTERP.STANDARDIZ.
Land
cover
class.
Syst.GEOVIS
AFRICOVER INTERPR. ANDMAPPING SYSTEM
AIMS
IMAGE PROCESSINGBASED
GIS BASED
DA
TA
BA
SE
Map
Acc.
Prog.
Africov.
Database
GATEW.
Interp.
Africov.
Database
Why a new classification?
Despite the importance of land cover, an internationally accepted reference base does not exist.
• Many classifications are heterogeneous with respect to quality, scale and nomenclature;
• Class definitions are imprecise, ambiguous and/or absent;
• Some classifications have a limited ability to accommodate the whole range of potential classes.
There is a strong urge to have a common classification system.
• Land cover types cross national boundaries;
• To facilitate data exchange, a common “language” is required.
Objectives
Why land cover?
Land cover is one of the most important elements for description and study of the environment.
• The main resource controlling primary productivity for terrestrial ecosystems can be defined in terms of land (and its cover);
• Land cover is the easiest detectable indicator of human interventions on the land.
Land cover is a critical parameter for environmental databases.
• Land cover changes quickly over time;
• Land cover is the basic geographic feature that other environmental applications can use as reference base.
Objectives
Definition
Land cover is the observed (bio)physical coverLand cover is the observed (bio)physical cover
on the earth’s surfaceon the earth’s surface.
It includes vegetation and man-made features as well as bare rock, bare soil and inland water surfaces.
Classification Concepts
Difference between Classification and Legend -1-
Classification is defined as the “ordering or arrangement of ordering or arrangement of objects into groups or sets on the basis of their objects into groups or sets on the basis of their relationshipsrelationships” (Sokal, 1974). It is an abstract representation of the situation in the field using well-defined diagnostic criteria.
Classification Concepts
Abstract representation of a classification consisting of a continuum with two gradients: circles and triangles in red and white (Küchler and Zonneveld, 1988).
Concrete situation in the field in a particular area (Küchler and Zonneveld, 1988). One should note that not all possible combinations of circles and triangles in black and white here represented occur in the left side (abstract representation).
Land Cover Classification SystemLand Cover Classification System
L C C SClassification is defined as the “ordering or arrangement of objects into groups or sets on the basis of their relationships” (Sokal, 1974). It is an abstract representation of the situation in the field using well-defined diagnostic criteria.
cover
height
100%
50%
10%
>50m 30m >5m1 2 3 4 ...
97 98 99 100
leaf type
leaf phenologyevergreen deciduousmixed
broadleaved(very large)
broadleaved(very small)
aphyllous
needleleaved
1 2 3 4 ...
25
natural vegetationnatural vegetation
Geographic
area
Land Cover Classification SystemLand Cover Classification System
L C C SL C C S
Legend is defined as the “application of a classification in a particular geographic area ”
100%
50%
10%
>50m 30m >5m1 2 3 4 ...
97 9899100
3347
7883
Reference Classification System
DEFINED
GEOGRAPHIC
AREA
mapping
units
derivedlegend
Inherent characteristicof the area determining the legend classes at a given scale and data type
Interpretation activity
A priori versus a posteriori
classification
Classification Concepts
CONCRETE FIELD SITUATIONA PRIORI CLASSIFICATION
A POSTERIORI CLASSIFICATION
ADVANTAGES:
- HIERARCHICAL ORDER- STANDARDIZATION
DISADVANTAGES:
- IMPLICIT RIGIDITY OF THE SYSTEM
ADVANTAGE:
- HIGH DEGREE OF FLEXIBILITY
DISADVANTAGES:- DEPENDING ON AREA- NO STANDARDIZATION
OR HIERARCHICAL ORDER
Current a-priori classifications in relation to flexibility
Conceptual Basis
high level of standardization
lowflexibility
incr
easi
ng
nu
mb
er o
f cl
asse
s
understandable list of class names
enormous list ofclass names with
unclear or too narrowclass boundary
limited number ofgeneric classes
limited number ofgeneric classes
adequate numberof detailed classes
high flexibility
low level of standardization
By increasing the number of classes in an a-priori system,this will increase the flexibility.
Class boundary definitions are based on very slight differences, this will decrease the level of standardisation.
Class boundary definitions will be clear and classes will contain aggregated land cover types. Therefore low flexibility.
Flexibility that will allow the accommodation of any occurring land cover.
New approach to classification -1-
Increasing flexibility while maintaining mapabilitymapability
The classification should be flexible in the sense that it should address the potential for the classification system to describe enough classes to cope with the real world.
At the same time flexibility should adhere to strict class boundary definitions that are unambiguous and clear.
Basic principle
A land cover class is defined by the combination of a set of independent diagnostic attributes, the so-called classifiersclassifiers that are arranged to assure a high degree of mapabilitymapability, i.e. geographical accuracy.
Conceptual Basis
New approach to classification -2-
Issues impeding application of the new approach
Land cover deals with a heterogeneous set of classes, therefore one single set of classifiers cannot be used to describe all features.
The set of classifiers is tailored to each major land cover feature in order to create meaningful combinations of classifiers.
Furthermore, with the aim of achieving a logical and functional hierarchical arrangement of classifiers, certain design criteriadesign criteria have been applied.
Conceptual Basis
Land cover classification system
New approach for land cover classification
Basic principle
A land cover class is defined by the combination of a set of independent diagnostic attributes, the so-called classifiersclassifiers . The combination of various types of classifier (trees, shrubs, herbaceous, closed, open, sparse, broadleaf,
needleleaf etc.) give to the user the possibility to define a wide variety of different land cover features
Land cover classification system
It is possible to state LCCS is a new language to describein a standardised way the different land cover features.
In a language words and a syntax allow to create a semantic concept.
In LCCS it is the same, the classifiers are the words the classification rules are the syntax and the land cover features the concepts to be described.
Conceptual Basis
How to create Land Cover classes in LCCS ?:
The same is in LCCS, the user has to find out the right combination of classifiers to describe a certain land cover feature.
Like using a language, the all concepts exist.The problemis then to find out the good combination of words to describe them.
No pre-defined list exist. The user must create each single class he/she needs one by one. This involves a process of passing from the idea of the class he/she has to the creation of the representation of the idea using a meaningful sequence of classifiers that are able to exactly represent the idea he/she has of the specific land cover feature.
Conceptual Basis
How to create Land Cover classes in LCCS :
trees
closed
open
shrubsherbaceous
sparse
evergreenbroadleaved?
??
Basic concept of a land cover class (the idea)the idea)
Use of LCCS method (the language)(the language)
Elaboration of the conceptin the codified LCCS language (the concept expression)(the concept expression)
Trees A3
Closed A10
Height 14-7m B6
Needeleaved D2
Evergreen E1
=A3+A10+B6+D2+E1
Conceptual Basis
LCCS is a concise language, it use the minimum setof classifiers to characterise a land cover feature.
WHY? For three main reasons:
- a difference exist between the process of classify or describe a feature.
- the use of the minimum set of classifiers necessary reduce the interpretative errors and facilitate the homogenisation of the interpretation.
- the characterisation of a land cover feature with the a minimum set of classifiers will reduce and simplify the field check.
AFRICOVER - Eastern Africa ModuleAFRICOVER - Eastern Africa Module
If a description of an elephant must be done all the possible elements can be used
Big hears
Long tooth's
Long noose
Big foots
Grey skin
Higher than 2.5 m
If a classification must be done only few elements will beused, only the ones that allow a distinct identification.
Long tooth's
Long noose
THE DESCRIPTION OF A RABBIT FOR INSTANCE DOES’NOT AUTOMATICALLY IMPLIES THE USE OF THE SAME ELEMENTS
THE CLASSIFICATION OF A RABBIT WILL USE THE SAME ELEMENTS SHOWING THE DIFFERENCE IN RANGE OF THEIR VALUES
Land Cover Classification SystemL C C S
There is a strong urge to have a common classification system.
• Land cover types cross national boundaries;
• To facilitate data exchange, a common “language” is required.
Land Cover Classification SystemLand Cover Classification System
L C C Sin
crea
sin
g n
um
ber
of
clas
ses
high level of standardization
lowflexibility
understandable list of class names
enormous list ofclass names with
unclear or too narrowclass boundary
limited number ofgeneric classes
limited number ofgeneric classes
adequate numberof detailed classes
high flexibility
low level of standardization
Land cover classification system
Overview of the software application
build up legend
edit classes
create user-defined land cover classes
display legend
save and retrieve
export
FIELD DATA MODULEFIELD DATA MODULE
standardized general field data collection
specific field data collection
automatic extraction of land cover class
from field data
saving of field data in synthetic form
print and export
TRANSLATOR MODULETRANSLATOR MODULE
translation of external classifications
into LCCS
similarity of external single classes through
LCCS
comparison of twoexternal classifications
through LCCS
comparison of twoLCCS classes
all classifiers and attributes
glossary
conditions to create land cover classes
database of all possible classes including name, code and
description
images and interpretation database
CLASSIFICATION MODULE
LEGEND MODULELEGEND MODULE
Africover - Eastern Africa ModuleAfricover - Eastern Africa Module
Africover Database Gateway (ADG) it will be a user-friendly software that will allow an easy and fast re-grouping of the Land Cover polygons not only by class names but also, in a much more flexible and detailed way, by classifiers used.
It will give to the Multi-Purpose database the tool to be really functionalfor multi-user applications.
Africover Interactive Database for interpretation (AID) it will be an interactive software to guide and homogenize the land cover interpretation.It will be the basic tool interpreters of different disciplines and countries will use to learn, improve and homogenize Land Cover interpretations.
AFRICOVER - Eastern Africa ModuleAFRICOVER - Eastern Africa Module
LEGEND GENERATION PROCESS
A FIVE PHASES PROCESS IS APPLIED:
•PRODUCTION OF A SO CALLED “TEORETICAL LEGEND”
•PRELIMINARY INTERPRETATION TO ARRIVE TO PRODUCEA”PRELIMINARY LEGEND”
•FIELD SAMPLES TO PRODUCE AN “INTERMEDIATE ACCURACYANALISYS” OF THE CLASSES INTERPRETED
•PREPARATION OF THE “FINAL LEGEND”
•FINAL INTERPRETATION TO PRODUCE THE “FINAL LEGEND”
THE LEGEND IS ALWAYS THE FINAL PRODUCT OF THE “INFORMATION EXTRACTION PROCESS”
From the concept to the practical application:- a set of new dedicated tools is developed that are explained below-
Africover Interpretation and Mapping System (AIMS)Geographic Vector Interpretation System (GeoVIS)is a new concept of on-screen interpretation software that combines the advantages of the digital automated classification and traditional visual interpretation in a synergetic way. It improves the interactive use of ancillary data and any other type data.
Land Cover Classification System (LCCS) is a new concept of land cover classification system that is flexible, able to describe any type of land cover feature in the world, at any scale or level of detail, in a standardized manner. It provides a common language to bridge users and existing systems.
Africover - Eastern Africa ModuleAfricover - Eastern Africa Module
Land cover classification system
It is possible to state LCCS is a new language to describein a standardised way the different land cover features.
In a language words and a syntax allow to create a semantic concept.
In LCCS it is the same, the classifiers are the words the classification rules are the syntax and the land cover features the concepts to be described.
AFRICOVER - Eastern Africa ModuleAFRICOVER - Eastern Africa Module
If a description of an elephant must be done all the possible elements can be used
Big hears
Long tooth's
Long noose
Big foots
Grey skin
Higher than 2.5 m
If a classification must be done only few elements will beused, only the ones that allow a distinct identification.
Long tooth's
Long noose
THE DESCRIPTION OF A RABBIT FOR INSTANCE DOES’NOT AUTOMATICALLY IMPLIES THE USE OF THE SAME ELEMENTS
THE CLASSIFICATION OF A RABBIT WILL USE THE SAME ELEMENTS SHOWING THE DIFFERENCE IN RANGE OF THEIR VALUES
Land cover classification system
NEW CONCEPT OF MULTI USER ORIENTED NATURALNEW CONCEPT OF MULTI USER ORIENTED NATURAL RESOURCES DATA BASERESOURCES DATA BASE
CLASSIFICATIONSYSTEM
INTERPRETATIONSOFTWARES
ACCURACYANALISYS
MULTI USER DATA BASEBROWSER
INTERP.STANDARDIZ.
Land
cover
class.
Syst.
AFRICOVER INTERPR. ANDMAPPING SYSTEM
AIMS
IMAGE PROCESSINGBASED
GIS BASED
DA
TA
BA
SE
Map
Acc.
Prog.
Africov.
Database
GATEW.
Interp.
Africov.
Database
AFRICOVER - Eastern Africa ModuleAFRICOVER - Eastern Africa Module
LEGEND GENERATION PROCESS
A CALIBRATION PROCESS MUST BE SET UP TO SELECT THEAMOUNT OF CLASSES THAT CAN BE “REALISTICALLY” MAPPEDWITH THE AVAILABLE “INTERPRETATIVE TOOLS” (ANCILLARYDATA, FIELD SAMPLES, REMOTE SENSING DATA, LOCAL KNOLEDGEOF THE AREA ETC)
LCCS IS AN EXTREMELY FLEXIBLE SYSTEM (APP. 100000 land cover variations can be codified)
THE FLEXIBILITY OF A CLASSIFICATION SYSTEM IS DIRECTLY LINKED WITH THE “THEMATIC GENERALIZATION” IN THE SENSE THAT AN INCREASE POSSIBILITY TO CODIFY THE “REALITY”CAN MINIMAZE THIS TYPE OF GENERALIZATION
AFRICOVER - Eastern Africa ModuleAFRICOVER - Eastern Africa Module
LEGEND GENERATION
IN AFRICOVER THE LEGEND IS THE MAXIMUM POINT OF EQUILIBRIUM BETWEEN THE INFORMATION GIVEN BY THE INHERENT CHARACTERISTICS OF THE STUDY AREA AND THE POSSIBILITY TO REPRODUCE THEM IN A BIDIMENSIONAL FORM (MAPPING ACTIVITYADOPTED = CARTOGRAPHIC STANDARDS - INTERPRETATIVE PROCESS)
NUMBER OFCLASSES
INTERPRETATIVE PROCESS(TIME-EFFORT-COST)
The requirements that led to the creation of LCCS are:
• harmonized and standardized collection of land cover data;
• availability of land cover data for a wide range of applications and users;
• comparison and correlation of land cover classes.
Objectives
LCCS is a comprehensive methodology for description, characterization, classification and comparison of any land cover identified
anywhere in the world, at any scale.