Operationalization of NASS Crop Area and Yield Estimation Programs.

For the past 15 years, the United States (US) Department of Agriculture / National Agricultural Statistics Service (NASS) has produced an annual crop specific land cover classification coined the Cropland Data Layer (CDL). It is derived from Landsat or Landsat-like satellite imagery and depicts more than 100 unique crop categories across the major farming areas of the US. In 2008 the product became national in scope, and today the entire historical CDL dataset is available to the public for visualization, analysis and dissemination via the CropScape web portal. The CDL has many uses internal to NASS. During the growing season the CDL is utilized to derive statistically-based acreage estimates of the major commodity crops. The CDL gets analyzed during times of natural disasters to estimate acreage reductions for regions likely not to be harvested. And finally, the CDL acts as a cropland “mask” to sort out which crops are growing where for use in the NASS remote sensing yield estimating program. The yield research, reliant on tracking regional average vegetation indices and surface temperatures over the crop season, was initiated by Paul Doraiswamy and has run alongside the CDL program for over a decade. The convergence of his work, the CDL production, expanded phenology datasets and modeling refinements has made estimating crop yields from space an operational reality.