USING SATELLITE AND REAL-TIME WEATHER DATA TO PREDICT MAIZE PRODUCTION

Large-scale assessments of crop conditions prior to harvest are critic al for providing early estimates of production. Satellite and weather information provide the opportunity for near real-time crop monitoring . The objective of this research was to develop an operational assessm ent system for crop production utilizing data from these sources. Maiz e (Zen mays) production was assessed in 42 Crop Reporting Districts (C RDs) across the United States Corn Belt, which produce 60% of all maiz e grown in the United States. Satellite, climatolocal, and agricultura l data were collected for 8 years, 1985-1992, and aggregated into CRDs . A model predicting the normalized maize yields for each CRD was deve loped that included as independent variables a satellite data variable . the Vegetation Condition Index, and a climatological variable, the C rop Moisture Index. This model explained approximately three-quarters (R-2=0.73) of the variation observed in the normalized yields, and was examined both for its accuracy and its timeliness in providing produc tion estimates. Predicted seasonal yields were summed to provide a mai ze production estimate for the entire Corn Belt study region. Producti on estimates deviated from the final USDA statistics, which become ava ilable several months after harvest, by less than 10% for all eight gr owing seasons. In addition, the production estimates were available ap proximately 2 months prior to the completion of the maize harvest. Thi s system has the potential for providing timely information to organiz ations monitoring regional or global agricultural production for human itarian or economic benefits.