Evaluation of Several Steady State Models for Predicting Soil Salinity in Temperate Region of Fars Province

In the arid and semi arid regions, accumulation of excess soluble salts in plants root zone is a limiting factor for crop production. The level and distribution of salinity in cropped lands vary in different times during the growth season as a result of interaction of rainfall, irrigation, evapotranspiration, leaching, and drainage. These relationships and interactions are complicated and not fully understood. Therefore, computer models are useful for predicting the response of the system to such complicated conditions and final evaluation of irrigation water quality. In the present study, three steady state models including WATSUIT, Hoffman and Van Genuchten models (exponential, trapezoid and 40-30-20-10 water uptake pattern) and crop water production function (CWPF) were evaluated. The required input data for running the models and the measured soils salinity were collected from experimental plots as well as farmer's cropped fields in different years in Fars Province. The results showed that soil salinities predicted by WATSUIT model and CWPF model were, respectively, lower and higher than that by the other models.  The models were evaluated by 2 statistical methods including root mean square error (RMSE) and coefficient of determination (D). Statistical index showed that the least error belonged to Hoffman and Van Genuchten models (40-30-2-10 water uptake pattern). Therefore, the capability of this model to predict soil salinity is more than the other models. Both statistical indexes showed that CWPF model predicted soil salinity with considerable error. Thus, this model was not suitable for predicting soil salinity in the present experimental conditions.