Spatial yield variability is a complex interaction of many factors, including soil properties, weather, pests, fertility, and management. Crop models are excellent tools to evaluate these complex interactions and provide insight into causes of spatial yield variability. The goal of this study was to use a soybean crop growth model to determine the contribution of three factors that cause spatial yield variability and to test several calibration and validation strategies for yield prediction. A procedure was developed to calibrate the CROPGRO–Soybean model and to compare predicted and measured soybean yields, assuming that water stress, soybean cyst nematodes (SCN), and weeds were the dominant yield–limiting factors. The procedure involved calibrating drainage properties and rooting depth over three seasons for each grid. These procedures were tested on 77 grids (0.2 ha in size) in the McGarvey field in Perry, Iowa, for 1995, 1997, and 1999. Predicted soybean yields were in good agreement (r2 = 0.80) with measured yield after calibrating three model parameters. The calibrated model was used to quantify the effects of three yield–limiting factors on soybean. The maximum soybean yield potential in 1997 was estimated by running the calibrated model with no water, SCN, or weed stress. The model was then run for 1997, turning each yield–limiting factor off to assess its relative impact on yield reduction. Average estimated yield loss due to the combined effects of water stress, SCN, and weeds in each grid was 842 kg ha–1. Soybean yields were significantly reduced by an average of 626 kg ha–1 as a result of water stress. The presence of SCN in several grids accounted for an average yield reduction of 105 kg ha–1. The effects of weeds on soybean yield were not significant.
Available at: http://works.bepress.com/gregory-tylka/172/