Empirical Root Water Uptake Calibration Using Output of a Physically Based Model.

Actual transpiration is estimated in hydrological simulations by root water uptake modeling. Empirical transpiration reduction models are commonly preferred because of their simplicity and lower data requirements. A drawback of these models, however, is the fact that their empirical parameters have no clear physical meaning and generally depend on soil type, atmospheric demand and plant properties. Their values are difficult to obtain and hardly documented in literature. To allow parameter calibration with a physically based model or numerical experiments, it is important to use a suitable empirical model. We evaluated some existing empirical models for root water uptake, and proposed a new empirical alternative. Models were then evaluated under different scenarios by comparing their predictions with a matric flux potential based transpiration reduction model that implicitly accounts for compensation of root water uptake. Models were embedded as sub-models into the ecohydrological model SWAP in order to be evaluated under soil water distribution scenarios. Model testing was done under a theoretical drying-out scenario with two atmospheric demands, represented by two potential transpiration rates (1 and 5 mm/d), without evaporation or precipitation. The simulations were performed for three types of soils and three root length densities exponentially distributed over depth. The parameters of the empirical models were estimated by minimizing the error in root water uptake simulated by SWAP with the physically based model and by SWAP with each of the empirical models, using the PEST software. The models, under the same root length density and soil type scenarios, were evaluated also for a whole year scenario using meteorological data from the Netherlands. We discuss simulation results and parameterizations.