Sensitivity Analysis and Evaluation of Physical and Experimental Models of Root Water Uptake in Tomato

Root water uptake is considered as a major factor for  predicting plant transpiration and the product yield. In general, the water root uptake models are divided into two macro- (experimental) and micro- (physical) groups. The physical models require more hydraulic parameters, while empirical models are simpler and require less input data. The aim of this study was to compare the performance of two empirical models and a physical model to predict the root water uptake of tomato under greenhouse conditions. Hence, the generalized likelihood uncertainty estimation (GLUE) was used to calibrate the models (hydraulic parameters of soil and root water uptake). The results of the sensitivity analysis of different models showed that the root water uptake is more sensitive to soil hydraulic conditions than the root characteristics. The results indicated that among the soil hydraulic parameters, the root water uptake shows more susceptibility to the coefficients of shape (λ) and the saturation moisture content (Өs). Among the root characteristics, the most sensitive parameter is the longitudinal density of the root. The results also revealed that the Van Genuchten model combined with the GLUE method has well simulated the root water uptake, as the parameters of R2, NS, NRMSE, MAE, ME, and d were 0.79, 0.69, 14.12, 0.75, 0.69, 0.86, respectively.  The integration of root water uptake models as a sub-model in the agro-hydrological models appears to be a useful tool for predicting the plant transpiration, the product yield as well as water management in the field.