NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR) FAO Land Cover Mapping methodology,

tools and standards&

GLC–SHARE database

Renato Cumani & John LathamLand and Water Division (NRL)

International Symposium on Land Cover Mapping for the African

ContinentJune 25-27, 2013

UNEP HQ & RCMRD, Nairobi, Kenya

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Content FAO Global Land Cover

Network Standards for Land Cover

mapping FAO Land Cover Mapping

Toolbox African land cover databases Global Land Cover SHARE

database Conclusions

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Main activities of the organization Putting information within reach Sharing policy expertise Providing a meeting place for nations Bringing knowledge to the field

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

FAO Global Land Cover Network (GLCN)Main Objectives:

To improve linkages between global, regional and national studies on land cover and the environment

To improve standardization, homogenization, compatibility and efficiency of information provided by different applications

To provide information that improves design and efficiency of sampling for validation of land cover products at global, regional and national levels.

To increase use and sharing of remote sensing data and its derived datasets

To provide comparable products at global, regional, and national and lower levels

To undertake capacity development and institution strengthening to maximize benefits for developing countries

To support operational development and use by national stakeholders of products emanating from the programme

multi-date landsat imagery

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

FAO GLCN Core activities Establish global network Develop Land cover mapping

methodology Standards development (LCCS/LCML) Land Cover Mapping Toolbox

(LCCS/MadCat/ADG) Technical assistance to national

experts for land cover mapping activities

Preparation of guidelines, manuals, templates, workshops, technical papers, metadata

Capacity building Awareness raising workshops, training

resources and sessions Dissemination and outreach (FAO

GeoNetwork and FAO GLCN website) Enable use of the land cover

information multi-date landsat imagery

Sudan

• Nepal

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Standards and Classification SystemLCCS / LCML / ISO 19144-2:2012LCCS: Comprehensive methodology for description, characterization, classification and comparison of most land cover features identified anywhere in the world, at any scale or level of detail: basis for comparative classification. (6 UN official languages)

Created in response to a need for a harmonized and standardized collection and reporting on the status and trends of land cover

Standards and Tools

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

LCCS databasesGlobal Land Cover

(GLC) 2000

1 km resolution

The dataset was sponsored by

members of the VEGETATION programme,

including JRC. Each partner used the

Land Cover Classification System (LCCS)

produced by FAO and UNEP, which

ensured that a standard legend was used across

the globe

Standards and Tools

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

LCCS databasesGlobCover ~2006

300 m resolution

The GlobCover Land Cover product is based on

ENVISAT MERIS data at full

resolution from January 2005 to June 2006.

The GlobCover Land Cover product is labelled

according to the UN Land

Cover Classification

System

Standards and Tools

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

FAO Land Cover Mapping Toolbox

ACCURACYANALISYS

MULTI USER DATA BASEBROWSER

THEMATIC & CART. ASPECT

INTERPRETATIONEFFICIENCY

Land

cover

class.

Syst.

Map

Acc.

Prog.

Advanc.

Database

GATEW.DAT

A B

AS E

DATA PRODUCERS DATA USERS

Standards and Tools

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

LCCS 2: 2001 (use LCCS)

LCCS 3: 2013 (use LCML/UML)

ADG 2: 2003 ADG 3: 2013

ADG 3 for ArcGIS 10.x: 2013

Standards and Tools

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Application designed by FAO Uses object-base classification Current version – Release June 2013 Wizard driven installation Implemented using .Net Framework Coding with LCCS2 and LCCS3 Requires Windows XP / Vista / 7 /8 (32 and 64

bit) Free to use for FAO programmes One time activation needed:

Institution, User Name, Address, PC CODE send request by email

Mapping Device – Change Analysis Tools (MADCAT)

Standards and Tools

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Country scale (30m or better resolution) • on going ECONET Ethiopia • 2012 Fouta Djallon

Highlands Malawi• 2011 Sudan • 2010 South Sudan Tunisia

Kenya Update• 2007 Somalia• 2006 Kenya LCC • 2005 Senegal • 2004 Libya• 1998- 2002 AFRICOVER

FAO’s Land Cover Mapping in Africa

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLCN/AFRICOVER: East Africa Module Project facts: Mapped area: 8.5

million Km2 Countries covered: 10 Landsat Scenes used:

more than 400 Period of activity:

1998-2004 Result: Multipurpose

Africover Database for the Environmental Resources produced at a 1:200,000 scale (1:100,000 for small countries and specific areas)

Development of a regional database and regional aggregation

Burundi, DR Congo, Egypt, Eritrea, Kenya, Rwanda, Somalia, Sudan,

Tanzania and Uganda.

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Fouta Djallon AOI: ca. 400,000 Km2

5 Countries within the AOI: Guinea, Guinea-Bissau, Mali, Senegal, Sierra Leone

LANDSAT coverage (30m res) 1990-2005

ASTER coverage (17 m res) 2008-2011

RapidEye coverage (5 m res) ~2005

Fouta Djallon Highlands land cover change

AG NVH NVS NVT URB WAT0

2000400060008000

100001200014000 1990

2008

Fouta Djallon Highlands land cover changes

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Malawi Land Cover change database (1990’s-2010’s)

From other to cropland (square blue) and vice versa (circle orange)

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Primary datasets: Landsat 30m (~2000, ~2005-2007 )

Spot4 imagery

2009-2010, 2.5-5m, 10-20 m res IRS 2007, 15-22m res Aster 2005-2010,15 m res.Ancillary: Googe Earth high resolution imagery Africover dataset (dated 1999-2000)

Sudan Land Cover

ATLAS of Sudan Digital ATLAS (DVD) Posters Skin Implemented in Sudan Capacity to undertake land cover

assessments

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Primary: Landsat ETM imagery

(GLS), 30m res., false color. • 2000 circa• 2005-2007

Spot4 imagery• 2006-2008, 10-20 m

res., true color.

Ancillary Google Earth high resolution

imagery Africover datasets

Land Cover Map of South Sudan

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

55 land cover classes

Landsat 1990’s and 2005’s

Completed in Senegal with involvement of national exerts

Senegal Land Cover Change: 2005

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Libya Land Cover database Landsat ETM+ imagery

2001’s and 2002’s 88 landsat scenes in total

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

ECO–NET Africa Earth Cover Network Africa

Sample tiles at 10 by 10 Km at half degree

Very high resolution imagery (targeted 5m or better)

Detailed land cover mapping using LCML enriched with vegetation index(es) from remote sensing

About 9,000 sampling sites

Coordinated and supervised by FAO GLCN with engagement of national experts

BASIC OBJECTS PROPRERTIES CHARACTERISTICS

Herbaceous Cover 30 – 50 % Cultivated Rainfed Field size 1 ha

Tree Cover 1 -3 % Natural Eight 5 -8 m. Disposition: irregular Leaf type: Broadleaf Leaf type % 100

Scrub Cover 2 – 4 % Natural Eight 1 – 3 m. Disposition: irregular Leaf type: Broadleaf Leaf type % 100

0.00

0.06

0.12

0.18

0.24

0.30

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35Dekad

NDVI

Ave

rage

0.00

0.02

0.04

0.06

0.08

0.10

NDVI

Sta

ndar

d De

viat

ion

AverageTotal St. Deviation

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

ECONET Outcomes Create a database with extremely

detailed information, global or regional, comparable and continually updated in support of a wide range of activities.

Serve as multi-statistical source of information at any level of detail or complexity. Using specific software on the Web, any end users worldwide will be able to define a geographical area and the categories for which area statistics will be generated.

Provide the most consistent, detailed and dynamic test site for calibration and/or accuracy assessment of any future wall-to-wall global, regional or national Land Cover mapping programmes

Africa Land Cover products

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE approach FAO System of Environmental

Economic Accounts (SEEA) London Group process

Global Consultations, interviews, comments, questions, recommendation

A significant step in improving the information accuracy of global land cover database

It integrates the best land cover data available (at sub-national, national, regional and global level) into one single harmonized database

It uses international standards: ISO TC 211 – 19144-2:2013 LMCL

GLC-SHARE database

Recommendation 19b.1: That the Land Cover Classification System (LCCS 3) developed by FAO should be adopted as the land cover classification system in the revised SEEA and that the LCML (ISO 19144-2) should be adopted at the methodology for linking to external sources of land cover data described in other land cover systems.

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE design principles Use existing available land cover databases at national,

regional and global level; Use the best available spatio-temporal land cover databases; Use the land cover legend prepared by SEEA and FAO based

on the Land Cover Meta-Language; Make use of the harmonization of the land cover elements

addressing semantic requirements; Use data fusion technology; Progressively update the database getting input from the

community of users and include additional datasets as they become available.

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE: fact-sheet Uses the FAO SEEA LCML(*) legend 30 arc-second pixel resolution 11+1 layers indicating the % share

of each class Dominant land cover layer and

quality score Overall class accuracy ~80% Designed as a platform to facilitate

crowd-sourcing Compatible with FAOSTAT

classification Designed to be used for GAEZ 2010

updateMethodology and datasets will be published in 2013

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE SEEA Legend

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE SEEA LCML Legend

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Coverage of land cover databases

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Processing Chain Multi-source and multi-resolution data fusion

through use of the land cover class elements and the LCML

Class confidence levels used in the processing chain to use the optimal spatio-temporal information at pixel level

Per product and per class rating assigned by technical experts (experts opinion)

Outputs include 11 rasters at 30 arc-second, 1 dominant land cover dataset, 1 quality indicator dataset including source date, resolution, sensor and confidence level per pixel, metadata, and technical paper

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-Share Database

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Quality AssessmentUsed ~1,000 points (ArcGIS and Google Earth validation) Overall dominant class accuracy ~80%

Distribution of the validation points

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

GLC-SHARE Summary GCL-SHARE is the first global database created

using the ISO standard for land cover classification ISO TC 211 – 19144-2 LMCL (Land Cover Meta Language) and is designed to be improved over time

GLC-SHARE will be used for GAEZ 2010 update GLC-SHARE will be used to update Land Use

Systems 2010 (FAO) Planned to be made publicly available for

comments and feedback by end of 2013 Fully documented, including metadata,

LCML/LCCS3 legend Update of the beta release with new available

datasets including the global coverage of Landsat 30 meter and new national land cover datasets

Maintained by FAO and community of practice partners such as GEO, CGIAR, JRC, IIASA

GLC-SHARE database

NAT

URA

L RE

SOU

RCES

DEP

ARTM

ENT

(NR)

Thank youContacts:

renato.cumani@fao.orgjohn.latham@fao.org

FAO

Links:www.fao.org

www.fao.org/nr/gaezwww.fao.org/geonetwork

www.glcn.org