Improved Land Cover Identification for the June Agricultural Survey Segments Using the Geospatial Cropland Data Layer

The National Agricultural Statistics Service (NASS) employs the June Agricultural Survey (JAS) to produce estimates of acreage and livestock inventories. The JAS sample is selected from an area-frame and segments of land comprise the sampling units. Every year, approximately 11,000, one-square mile sample units or segments are visited by field enumerators at the beginning of the growing season to collect crop type and acreage information across the United States. In regions of the United States with small-scale agriculture, screening to identify farm operators is often time-consuming, expensive, and subject to misclassification. Consequently, a procedure was developed to obtain land cover statistics within the JAS segments, to more accurately identify potential land use, prior to prescreening activities. The land cover statistics are based on the NASS Cropland Data Layer (CDL). The CDL is a raster-formatted, geo-referenced, crop specific, land cover product, produced annually. The most recent five years of the CDL are used to categorize the land in each JAS segment. Within each segment, percentages are calculated for several predetermined categories, such as corn, soybeans, pasture, cultivation, and impervious areas among others. This paper will describe the project objective, the methodology used to calculate the land cover statistics or percentages based on the crop specific CDL data, and the impact of incorporating geospatial data for a statistical application.