Application of a distributed physically based hydrological model to a medium sized catchment

Current distributed physically based models are often complicated to the extent that often because of lack of data, they are rarely calibrated and validated thoroughly. Rarely, internal non-calibrated wells or discharge stations are included in a model evaluation. In this study the fully distributed physically based MIKE SHE code was applied to the 600 km(2) catchment of the Gete, Belgium. First, a rigorous method for the data flow of a data intensive code such as MIKE SPIE was established. The model was then calibrated both against daily discharge measurements and observed water levels. After the calibration, the model was first validated using a split-sample (SS) test. The results show that observed discharges are simulated well in both the calibration and the validation period, while results for the piezometric levels differ considerably among the wells. In addition, a multi-site (MS) validation test was performed for 2 internal discharge stations and 6 observation wells, which showed inferior results for the discharge stations and comparable results for the water table wells. As in the calibration and the SS validation test, water table fluctuations are predicted well in some wells, but with little agreement established in others. This is partly due to scale effects and the poor quality of the data in certain areas of the catchment. Mainly, the lack of data makes it as yet difficult to simulate time series of internal catchment variables with acceptable accuracy. Because of this, the calibrated and validated model is unable to provide reliable predictions of the water table over the entire catchment.