HR: 11:35h
AN: B42A-05 [Abstracts]
TI: Multi-resolution methods for mapping tropical forest cover change
AU: * Hansen, M C
EM: Matthew.Hansen@sdstate.edu
AF: Geographic Information Science Center of Excellence, Wecota Hall, Box 506B
South Dakota State University, Brookings, SD 57007
United States
AU: Townshend, J R
EM: jtownshe@geog.umd.edu
AF: Department of Geography, 2181 LeFrak Hall
University of Maryland, College Park, MD 20742
United States
AU: Loveland, T R
EM: loveland@usgs.gov
AF: United States Geological Survey, National Center for Earth Resources Observation and Science, Sioux
Falls, SD 57198
United States
AU: Pittman, K
EM: Kyle.Pittman@sdstate.edu
AF: Geographic Information Science Center of Excellence, Wecota Hall, Box 506B
South Dakota State University, Brookings, SD 57007
United States
AU: Carroll, M
EM: marc@geog.umd.edu
AF: Department of Geography, 2181 LeFrak Hall
University of Maryland, College Park, MD 20742
United States
AB:
Automated methods for mapping forest cover change are presented using MODIS and Landsat data sets. Vegetation Continuous
Fields of percent tree cover, a standard MODIS Land Science Team product, allow for the creation of forest change indicator
maps. Inputs to the algorithm are time-series composites that are used to generate time-integrated annual metrics. Metrics
capture the salient features of vegetation phenology and enable continental to global scale mapping of vegetation attributes
such as tree cover. Combined spectral/thematic change difference images are then used to derive maps undergoing likely
change. Change indicator maps reveal relative increases or decreases in annual conversion rates for areas exhibiting
extensive change, such as South America and Insular Southeast Asia. Change indicator maps also enable a targeted sampling
scheme of Landsat data to produce more accurate areal estimates of change. For an area such as Central Africa, MODIS change
maps do not reveal change on an annual basis. For such an area, MODIS 250 meter maps of tree cover are used to calibrate
mapping of Landsat data. This inverse scaling approach enables the synoptic MODIS signal to be used in training Landsat
images over a wide region. Such a methodology points the way to operationalize wall-to-wall fine-scale mapping of forest
change for the entire tropical zone.
DE: 1632 Land cover change
DE: 1640 Remote sensing (1855)
SC: Biogeosciences [B]
MN: Fall Meeting 2005