Data Mining for Forecasting Mississippi Cropland Data Layers

In 1999, Mississippi became an early adopter of the National Agricultural Statistics Service (NASS) Cropland Data Layer (CDL) program. With the support of the NASS Spatial Analysis Research Section (SARS), we have progressed from an annual crop picture to a pixel by pixel history of Mississippi farming. Much of our early work for Mississippi agriculture is now easily provided from the web based application CropScape, released by SARS in 2011. In this study, pixel history data from CDLs has been mined to give forecasts of Mississippi crop acres. Traditionally, such agricultural data mining emphasizes the trends of early adopters driven by factors such as global warming, technology, practices, or the marketplace. These studies provide forecasted CDL products produced using See5° and Imagine°, the same software used in Mississippi CDL production since 2006. Mississippi CDL forecasts were made using historical information available as soon as the CDL for the previous year was completed. For example, the CDL forecast for winter wheat, produced at a date when winter wheat was planted but not most crops, gave results of 104.6 +/- 5.4% of the official NASS estimates for winter wheat for the years 2009-2011. In 2012, all of the states of the contiguous US will have the historical CDL data to do this type of study. A CDL forecast is proposed as a useful addition to CropScape.