A Look at the Remote Sensing Applications Program of the National Agricultural Statistics Service

Abstract

This paper presents a summary of the procedures used by the National Agricultural Statistics Service (NASS) in its crop area estimation program during the years 1980-1987. It includes briefly some of the results with more detailed information provided by Allen and Hanuschak (1988). In its application program, NASS formed a regression estimator for crop acreage by using satellite data in conjunction with ground data which were collected during the annual June Enumerative Survey. The track record shows that the Landsat based crop area estimates for major producing regions of the U.S. were closer than the June Enumerative Survey (JES) direct expansion estimates to the Agricultural Statistics Board final estimates most of the time. The basic methodology, data processing techniques, and concepts used in the Landsat estimating program were developed through various research projects during the 1972-1979 period and are introduced briefly here. The timing of this description is appropriate with 1987 being the final year of operational crop estimation using data from the Landsat Multispectral Scanner Sensor (MSS). In the future there will most likely be a return to the use of satellite data in the estimating program, but for now NASS will no longer be using this data to produce timely crop estimates. The primary reason for this discontinuation is the uncertain status of the current Landsat satellites which have already outlived their expected design lives. In addition new satellite technology in the United States, France, Japan, India, and the USSR has produced data far superior to that yielded by Landsat's MSS. However, in order for NASS to take advantage of the new data, more research is required to assess the feasibility of its use so that when a new program is implemented, the anticipated improvement in the accuracy of the results will be cost effective.