HR: 16:00h
AN: B22E-01    [PDF]
TI: Application of MODIS Land Products to Estimate Regional Cropland Area and Production
AU: * Lobell, D
EM: dlobell@stanford.edu
AF: Carnegie Institution Dept of Global Ecology, 260 Panama St, Stanford, CA 94305 United States
AU: * Lobell, D
EM: dlobell@stanford.edu
AF: Stanford University, Dept. of Geological and Environmental Science, Stanford, CA 94305 United States
AU: Asner, G
EM: gasner@globalecology.stanford.edu
AF: Carnegie Institution Dept of Global Ecology, 260 Panama St, Stanford, CA 94305 United States
AU: Asner, G
EM: gasner@globalecology.stanford.edu
AF: Stanford University, Dept. of Geological and Environmental Science, Stanford, CA 94305 United States
AU: Ortiz-Monasterio, I
EM: i.ortiz-monasterio@cgiar.org
AF: International Maize and Wheat Improvement Center (CIMMYT), Wheat Program, Apdo. Postal 6-641, 06600 Mexico D.F., Mexico, Mexico D.F., 06600 Mexico
AB: The spatial and temporal coverage of MODIS offers unique opportunities for agricultural applications. Here we investigate the application of 250m MODIS vegetation index composites to map crop areas and yields in various agricultural regions in Mexico and the United States. Traditional 'hard' classification of MODIS data can lead to significant errors when estimating crop areas because a MODIS pixel is often large relative to typical field sizes. We present an approach that uses temporal reflectance signatures to determine the sub-pixel extent of various crops using linear unmixing. Endmember sets are constructed using Landsat data to identify pure pixels, and uncertainty resulting from endmember variability is quantified using Monte Carlo simulation. This approach allows endmembers to be used over broad regions and in different years, facilitating operational estimates of crop area with well-defined uncertainties. We then apply a light-use efficiency model to relatively pure pixels to determine the productivity of different crops. The resulting estimates of total crop area and production are compared with reported harvest statistics, and various sources of uncertainty are discussed.
DE: 1640 Remote sensing
DE: 1694 Instruments and techniques
SC: Biogeosciences [B]
MN: 2003 Fall Meeting