Evaluating three cotton simulation models under different irrigation regimes

Abstract

Crop growth/irrigation scheduling simulation models are becoming an integral part of crop management schemes designed to maximize input use. The selection of a particular model over others must be based on the model's performance under local conditions. Three crop simulation models, COTTAM, GOSSYM and IRRSCH, were evaluated for their accuracy in approximating field-observed crop water stress index (CWSI) of cotton grown under different irrigation regimes in a semiarid environment. The purpose is to use CWSI measured in the field to evaluate and to select an appropriate simulation model to schedule irrigation for cotton production. Generally, CWSI predicted by the models was lower than the measured values. Comparatively, IRRSCH was the best of all the models in predicting CWSI regardless of the irrigation treatments. Statistical analysis showed no significant difference (at 95% significance level) between the measured and the predicted CWSI by IRRSCH when cotton was irrigated at −0.40 MPa in 1983. Analysis of variance of paired data showed that COTTAM and IRRSCH predicted similar CWSI ( P = 0.05) in cotton irrigated at −1.50 MPa in 1983. This study showed IRRSCH, a simple water balance model, is superior to COTTAM and GOSSYM in approximating CWSI at water stress levels most encountered in irrigated fields (−0.15 and −0.40 MPa) and best approximates field water balance.