摘要：Core Ideas Parameters of different RWU models were optimized using data from a rhizotron facility. The same soil hydraulic properties were obtained for models considering RWU compensation. Feddes–Jarvis and Couvreur models predicted similar root‐system‐scale stress functions. The obtained RWU parameters were consistent with data reported in the literature. The models with compensation predicted similar total uptake but different local RWU. The spatiotemporal distribution of root water uptake (RWU) depends on the dynamics of the root distribution and compensatory uptake from wetter regions in the root zone. This work aimed to parameterize three RWU models with different representations of compensation: the Feddes–Jarvis model that uses an empirical function, the Feddes model without compensation, and the Couvreur model that is based on a physical description of water flow in the soil–root system. These models were implemented in HYDRUS‐1D, and soil hydraulic parameters were optimized by inverse modeling using soil water content and potential measurements and observations of root distributions of winter wheat ( Triticum aestivum L.) in horizontally installed rhizotubes. Soil moisture was equally well predicted by the three models, and the soil hydraulic parameters optimized by the models with compensation were comparable. The obtained RWU parameters of the Feddes–Jarvis model were consistent with data reported in the literature, although the pressure heads h 3l and h 3h for lower and higher transpirations rates, respectively, could not be uniquely identified. Response surfaces of the objective function showed that the root‐related parameters of the Couvreur model could be identified using inverse modeling. Furthermore, these parameters were consistent with combined root architectural and hydraulic observations from the literature. The Feddes–Jarvis and Couvreur models simulated similar root‐system‐scale stress functions that link total RWU to the effective root zone water potential, suggesting that parameters may be transferable between the two models. Simulated RWU profiles differed due to different water redistribution by the root system, but the measurements were not sufficiently precise to observe this redistribution.
