Retrieval of irrigated and rainfed crop data using a general maximum entropy approach

This paper presents a method to separate harvested area and yield for irrigated crops from rainfed crops in a region, given gross harvested area and yield, and climatic, agronomic and economic data for crops. The method is based on the principle of general maximum entropy, which combines incomplete data, empirical knowledge and a priori information to derive desired information. The model is applied to three large basins with aggregated climatic and agricultural conditions, and to five counties in Texas and California. The modeled results and assessed values in these study areas are compared. While the dependability of model outputs relies on empirical knowledge and judicious parameter estimation, the model remains reliable even for the significant level of uncertainty produced by subjectively predetermined major parameters. The model can be applied to retriving historical data for irrigated and rainfed crops; it can also be used for irrigated and rainfed agriculture planning based on climatic and technological projections. Moreover, the model provides other useful information, including water allocation by crop, water use efficiency and the impact of other agricultural inputs.