Abstract:
Climate change will have important impacts on hydrological systems, especially in semi-arid regions, where water resource is limited. To predict these changes, various hydrological models have been used. I investigated here the pros and cons of using integrated hydrological models for this purpose. Integrated hydrological model are distributed, surface/subsurface coupled hydrological models, which aim to reproduce the water processes using physically-based relationships. Because of their long calibration time and because of their high number of parameter, they are also difficult to calibrate at catchment scale. To accelerate this process, I developed a calibration method based on a hierarchy of grids of increasing precision. This method was applied in a research catchment in north-east Spain. This catchment underwent a rapid transition from non-irrigated to irrigated agriculture. I could reproduce this transition with my model, suggesting that it could be effective in predicting climate change effects. Then, I modeled hydrological effect of climate change in this catchment under different irrigation conditions. Irrigation onset has a large impact on the catchment responses to climate change. For example, actual evapotranspiration is predicted to decrease in the future in the scenario without irrigation while it is predicted to increase in scenarios with irrigation. Finally, I studied the future impacts of meteorological droughts. I compare the outputs from my model with a simpler method based on drought indices (summary metrics representing the dryness level). Drought indices were shown to have important weaknesses compared with the model outputs, notably concerning the predictions of the future drought intensities. The conclusion of this thesis is that integrated hydrological models are useful to predict hydrological climate change impacts, notably in catchments where land-use change or surface/subsurface interactions are important.
Print-on-Demand