We are increasingly using hydrological models to study the potential impacts of future climate change on catchment runoff. However, the impact studies often present the results predicted by any particular hydrological model. Such model predictions represent only the results of that model and the reliability will depend on how well the model structure is defined and how well the model is parameterized. This might raise concern that hydrological impact studies using a single hydrological model may lead to unreliable conclusions. There are varieties of hydrological models with different levels of complexities but only little information is known about the impact of these model structural differences on the predictions. Therefore, this study analyses two different model conceptualizations to have an idea about the uncertainties related to hydrological model structure in the climate change impact predictions. For the intercomparison, two distributed models are considered, namely MIKE SHE and WetSpa. The models were built and calibrated using identical datasets for Grote Nete catchment, a mediumsized catchment in Belgium. Both models were able to capture the flow dynamics of the catchment runoff with high efficiencies. High flows were well modeled by WetSpa, but slightly overestimated by MIKE SHE, whereas the low flows were better approximated by MIKE SHE, and slightly overestimated by WetSpa. Regarding their predicting capacities, both models showed to overpredict the high flow changes, by which they might underestimate the impact of climate change. However, this is mainly seen for at the high flow changes and is more pronounced for the MIKE SHE model. The MIKE SHE model was built with a 3-D groundwater flow and river model, which allowed to simulate discharges at the - three - internal stations within the catchment. The model performance results downstream the catchment could be passed to these internal stations. The observed and simulated runoff series match reasonably well, athough the efficiencies were lower than downstream the catchment. At these internal gauge stations the low flow minimas are very well estimated, whereas the peak flow performance was less good but acceptable at Vorst and Meerhout. The groundwater heads in MIKE SHE and their seasonal variation had a high model performance.
Both calibrated models were forced with the climate scenarios for the 2050 and 2100 future periods and their impacts on the runoff regime were assessed by means of changes in the high and low flow extremes. The estimated flow changes by the MIKE SHE and WetSpa model showed similar tendencies. The high flow extremes could decrease as well as increase depending on the applied climate change scenario. For the high scenario, climate change will act positively with higher expected extremes, whereas decreases are expected in case of the low climate scenario.