National Land-Cover Data and Census of Agriculture Estimates of Agricultural Land-Use Area Differ In the Northeastern United States

摘要

The landscape of the northeastern United States is a complex mixture of forest, agriculture, and development. Many urgent social and environmental issues require spatially-referenced information on land-use, a need potentially met by the National Land-Cover Data (nlcd). nlcd non-site-specific accuracy from 1992 and 2001 was assessed by comparing county proportions of cropland and agricultural grassland (pasture/hay) area to those recorded in the usda Census of Agriculture (coa) for 1992 and 2002. Estimates of total agricultural area corresponded closely (r~2 = 0.91 for 1992; 0.85 for 2001/2), but nlcd performed poorly at distinguishing between cropland and grassland (r~2= 0.64 and 0.80 for cropland in 1992 and 2001; 0.50 and 0.44 for grassland in those years). High forest cover reduced accuracy. Landscape complexity and patch size may impede classification accuracy. Accuracy varied greatly by state, even within a single mapping region. The nlcd is the best available source of spatial land-cover, but may not be suitable for applications requiring agricultural subclass data.