Modelling the hydrology of a catchment using a distributed and a semi-distributed model

Various hydrological models exist that describe the phases in the hydrologic cycle either in an empirical, semi-mechanistic or fully mechanistic way. The way and level of detail for the different processes of the hydrologic cycle that needs to be described depends on the objective, the application and the availability of data. In this study the performance of two different models, the fully distributed MIKE SHE model and the semi-distributed SWAT model, was assessed. The aim of the comparative study was to examine if both models are equally able to describe the different phases in the hydrologic cycle of a catchment, given the availability of hydrologic data in the catchment. For the comparison, historic data of the Jeker river basin, situated in the loamy belt region of Belgium, was used. The size of the catchment is 465 km(2). The landscape is rolling, the dominant land use is farmland, and the soils vary from sandy-loam to clay-loam. The daily data of a continuous period of 6 years were used for the calibration and validation of both models. The results were obtained by comparing the performance of the two models using a qualitative (graphical) and quantitative (statistical) assessment, such as graphical representation of the observed and simulated river discharge, performance indices, the hydrograph maxima, the baseflow minima, the total accumulated volumes and the extreme value distribution of river flow data. The analysis revealed that both models are able to simulate the hydrology of the catchment in an acceptable way. The calibration results of the two tested models, although they differ in concept and spatial distribution, are quite similar. However, the MIKE SHE model predicts slightly better the overall variation of the river flow. Copyright (C) 2005 John Wiley Sons, Ltd.