Yield prediction models based on remotely sensed data

Estimation crop yield before the harvest is one of the greatest concerns in agriculture, since variations in crop yield from year to year impact international trade, food supply, and market prices. Early prediction of crop yield on the global and regional scales offers useful information to policy planners. Appropriate recognition of crop productivity is essential for sound land use planning and economic policy (Hayes and Decker 1996). At the field-scale, crop yield information helps farmers to make quick decisions for upcoming situations, such as the choice of alternative crops and whether to abandon a crop at an early stage of growth. More recently, assessment of crop productivity at the within-field level has become an important issue in precision farming (Stafford 2000; Yang et al., 2001a). The relationships between various environmental factors -typically meteorological information and soil parameters -and crop yield, has been the most common approach to predicting crop yields in past years (Varcoe, et al. 1990; Drummond et al., 2003). More recently, relationships between chlorophyll content in crop leaves and grain yield have been explored using SPAD chlorophyll meters. Prediction of crop yield based on remotely sensed data has not yet become the norm.