A detailed description of the methods utilized to produce the data, as well as research results, are provided in Ramankutty et al. The data sets were developed from a combination of agricultural inventory data and satellite-derived land cover data sets. MODIS land cover data and SPOT Vegetation’s global land cover 2000 (GLC 2000) data were combined to create 391 composite land cover classes. The composite land cover classes were calibrated against the agricultural inventory data by first computing land area fractions from inventory data and then applying linear models relating satellite data sets to agricultural inventory data to compute adjusted cropland and pasture area estimates. Cropland and pasture area estimates from land cover data sets were aggregated to the subnational administrative level and then compared with the inventory data to determine correction factors.
Figure 1. Global Distribution of Croplands in 2000
Figure 2. Global Distribution of Pasture Lands in 2000