HR: 16:15h
AN: B22E-02    [PDF]
TI: Multi-year southern Africa MODIS burned area product generation and validation
AU: * Roy, D P
EM: droy@kratmos.gsfc.nasa.gov
AF: Department of Geography, 1113 LeFrak Hall, University of Maryland, College Park, MD 20742 United States
AU: Jin, Y
EM: yjin@hermes.geog.umd.edu
AF: Department of Geography, 1113 LeFrak Hall, University of Maryland, College Park, MD 20742 United States
AU: Justice, C O
EM: justice@hermes.geog.umd.edu
AF: Department of Geography, 1113 LeFrak Hall, University of Maryland, College Park, MD 20742 United States
AU: Giglio, L
EM: giglio@hades.gsfc.nasa.gov
AF: Science Systems and Applications, Inc., NASA Goddard Space Flight Centre, Code 923, Greenbelt, MD 20771 United States
AU: Lewis, P
EM: plewis@geog.ucl.ac.uk
AF: Remote Sensing Unit, Department of Geography, University College London, 26 Bedford Way, London, WC1H0AP United Kingdom
AB: A MODIS burned area product has been developed that maps the 500m location and approximate day of burning using a change detection algorithm based on a bi-directional reflectance model-based expectation approach. The algorithm has been applied to recently reprocessed 500m daily MODIS land surface reflectance data to produce burned area data sets for all of southern Africa, including Madagascar, for 2000 onwards. This paper presents the MODIS southern Africa burned area product, describes the protocol developed to validate it, and the validation results. The validation protocol is based upon interpretations by members of the Southern Africa Fire Network (SAFNet) of multitemporal Landsat Enhanced Thematic Mapper plus (ETM+) data to derive maps of the location and approximate date of burning. The protocol was implemented in 2000, 2001 and 2002 at eleven Landsat ETM+ scenes distributed across southern Africa, from Namibia to Mozambique, to encompass representative regional variation in the conditions for which the MODIS burned area product was generated and to capture the more important factors that influence product performance. Statistical comparisons between the Landsat ETM+ derived independent burned area data and the 500m MODIS burned area product for 2000, 2001 and 2002 are described. Comparisons between the MODIS burned area products and MODIS 1km active fire detections are also presented. Constrains on our ability to precisely define the limits of MODIS burned area detection are discussed. The implications of this work for improved regional and country level burned area estimates and the need for development of validation reporting metrics specific to the information needs of different users in southern Africa are also discussed.
DE: 0400 Biogeosciences
DE: 1630 Impact phenomena
DE: 1640 Remote sensing
DE: 1699 General or miscellaneous
DE: 9305 Africa
SC: Biogeosciences [B]
MN: 2003 Fall Meeting