Evaluation of Semiprocess-Based and Empirical Models in Simulating Root Water Uptake under Water- and Salt-Stressed Conditions

The transpiration reduction function of root water uptake (RWU) models play a crucial role in modeling crop evapotranspiration, yield, and water-use efficiency. Various empirical functions can describe how RWU responds to water and salt stresses. This study evaluated semiprocess-based (McS) and empirical RWU models that are sufficiently sophisticated to model salt-water dynamics under stress conditions. Global sensitivity analysis was performed on these models for an irrigated scenario of berseem (Trifolium alexandrinum) for a numerical crop growth experiment. The sensitivity of the model outcomes (relative transpiration, bottom flux, and average root zone electrical conductivity) to soil hydraulic and stress parameters was analyzed. The results show that the model outcomes were more sensitive to the threshold stress parameters of the McS model than to those from the empirical models. The McS model’s cumulative bottom flux (CBF), average root zone electrical conductivity, and relative transpiration (RT) information emerged as effective data for parameter calibration, whereas other models fell short in addressing uncertainties related to water stress parameters. The observed increase in uncertainty during deficit irrigation in saline soil emphasized the importance of accurately characterizing stress parameters to enhance the predictive capabilities of the model developed. The model effectively simulates water deficit and salinity impacts on RWU, showing strong sensitivity to stress parameters, particularly in dry periods and deeper root zone layers. The McS model parameters were calibrated using berseem data, and showed good alignment with observed soil moisture (R2=0.74–0.86) and electrical conductivity (R2=0.86–0.92) at various depths. The model effectively characterizes stress parameters, thereby enhancing predictive capabilities for RWU dynamics under water deficit and saline conditions.