Crops such as soybean ( Glycine max L.) are grown predominantly under rainfed conditions where water is a major limiting factor and the interannual variability in rainfall pattern is high. Crop modeling has proven a valuable tool to evaluate the long-term consequences of weather patterns, but the candidate crop models must be tested and calibrated for new regions prior to their use as extrapolation tools to predict optimum cultivar choice and sowing dates. The objectives of this paper were to calibrate the CROPGRO-soybean model for growth and yield under rainfed conditions in Galicia, northwest Spain, and then to use the calibrated model to establish the best sowing dates for three cultivars at three locations in this region. The starting point of the calibration process was the CROPGRO-soybean version previously calibrated for non-limiting water conditions. The original model, when simulated versus rainfed soybean field data sets, tended to simulate more severe water stress than actually occurred. In order to calibrate growth and yield for the actual soil we tried several ways for the modelled crop to have access to more water. Modifications were made on soil depth, water holding capacity, and root elongation rate. In addition, other changes were made to predict accurately the observed water-stress induced acceleration of maturity. Long-term simulations with recorded weather data showed that soybean is more sensitive to planting date under irrigated than rainfed management, in the three studied Galician locations.
Read full abstract