Revisiting the Feddes reduction function for modeling root water uptake and crop transpiration

•The empirical Feddes reduction function is widely used without onsite calibration.•Adapted and modified versions improve model predictions within limits.•Predictions of compensated water uptake need process-based models.•Inverse modeling to calibrate the Feddes reduction function is feasible.•Some of the Feddes parameters are insensitive to SWAP model outputs. The widely used Feddes reduction function (FRF) was reviewed in the context of its merits and shortcomings for modeling root water uptake (RWU), crop transpiration, and soil water balance components. The empirical nature of the FRF was contrasted to few calibrations reported in the literature, and the conditions where the correct operation of the FRF was of limited importance were discussed. Some alerts regarding the feasibility of this function in hydrological studies were provided and reinforced by a case study with sugarcane under subtropical conditions. Available modifications of the FRF can improve RWU predictions under aeration stress but some conceptual imperfections make it inaccurate under drought stress and high compensated uptake. In such cases, the use of mechanistic RWU models in replacement of the FRF is recommended to accurately predict the actual transpiration dynamics. The calibration of the FRF by inverse modeling is feasible, although some of the FRF parameters are insensitive to model outputs.