A spatial analysis of corn and soybean yields and weather relations

This thesis studied the relationship between corn and soybean yields and weather conditions in U.S. Corn Belt area. Three statistical models were tested, including two fixed effect models and one geographically weighted regression. Due to different underlying assumptions and specifications of each model, results indicated different implications. Major findings included severe weather conditions during plant.s reproductive stage had much greater impact than vegetative stage; the favorable effect of moderate heat could not compensate the yield loss caused by extreme heat at the same magnitude; and heterogeneous crop-weather relations across crop reporting district prevented further aggregation of cropping region. As another objective of this thesis, predictive powers of each model were also studied