Comparative evaluation of three-crop growth models for the simulation of soil water balance in oilseed Brassica

Performance of three-crop growth models for dynamic simulation of soil water balance in the sandyloam soils cropped to oilseed Brassica was evaluated. The model parameters were subjected to sensitivity analysis, modified and calibrated for local environment and verified with experimental field values. Simulated root zone (0–120 cm) moisture from Campbell–Diaz model was more sensitive to stepwise changes in input parameters as compared to the O'Leary and SWASIM models. While calibrating these models during post-rainy season of 1992–1993, simulated profile moisture from the Campbell–Diaz and SWASIM models, on an average, did not deviate by more than ± 5% from measurements except on one or two occasions whereas the O'Leary model gave slight overestimates up to seed filling stage and underestimates of the order of −6% in the post-seed filling stage. Campbell–Diaz model requiring least number of inputs yielded the lowest error estimates (6.01–11.06 mm) followed by the SWASIM (6.46–18.88 mm) and O'Leary (8.3–20.6 mm) models. The Campbell–Diaz and SWASIM models can also be successfully used with common coefficients and also to simulate layerwise soil moisture contents, respectively.