Investigation of uncertainties in assessing climate change impacts on the hydrology of a Canadian river watershed

It is known that climate change will affect water resources. The impact of climate change on river regimes attracted the attention of hydropower companies over the recent years. Although general trends in future river flows can be reasonably well assessed using climate and hydrological models, their overall uncertainty is much more difficult to evaluate. An improved evaluation of uncertainty in the hydrological response of watersheds to climate change would help designing hydropower projects as well as adapting existing systems to better cope with the anticipated changes in flow regimes.;The uncertainty in assessing the potential hydrological impacts of a changing climate is of high interest for the scientific community. A framework for evaluating the uncertainty of climate change impacts on watershed hydrology is developed in this thesis. The sources of uncertainty studied comprise the global climate model (GCM) structure, climate sensitivity, natural variability and, to a lesser extent, hydrological model structure. Climate sensitivity is the global mean climatological temperature change due to a doubling of atmospheric CO2 concentration. Natural variability refers to the uncertainties resulting from the inherent randomness or unpredictability in the natural world. Climate projections under IPCC A2 scenario for the 2080 horizon were downscaled to regional scale using the change factor method and developed into long time series with WeaGETS, a stochastic weather generator developed at the Ecole de technologie superieure. The predicted future climate scenarios were forced into four hydrological models to simulate future flows in the Manicouagan River Basin, located in the province of Quebec, Canada. The Monte Carlo sampling method was implemented as a probabilistic approach to trace out the magnitude of uncertainty in accordance with the weighting schemes attributed to the different sources of uncertainty. In particular, equal and unequal weights were attributed to the GCMs to see whether this would have a significant impact on the resulting flow return periods. This experiment was motivated by the fact that GCM structure is usually the most significant contribution to the overall uncertainty in projected flows. Experiments using equal and unequal weights on climate sensitivities were also performed.;Modelling results indicate that the future hydrological cycle will intensify as precipitation and temperature will increase in the future for all GCMs projections. The spring peak flow will occur earlier by a few weeks for all GCMs investigated and will increase for a majority of them.;The uncertainty related to GCM structure, climate sensitivity, natural variability and to hydrological model structure were assessed separately. Uncertainty due to the GCM structure was found to be significant and to vary seasonally and monthly especially during the month of April. Variations in climate sensitivity and natural variability introduced moderate changes in the hfydrological regime when compared with the uncertainty due to GCM structure. The choice of hydrological model will also result in a non-negligible uncertainty in climate change studies.;At last, the magnitude of uncertainty under various weighting schemes attributed to GCMs and climate sensitivities was evaluated. Weight scheme experiments indicate that assigning equal or unequal weights to the GCM structure had a marginal to small effect on the return periods calculated for the hydrological variables studied, i.e. monthly discharge and spring runoff volume. However, for climate sensitivity, the weights assignment notably influenced the probability of occurrence of large hydrological events. The choice of the hydrological model also had a significant impact on return periods of large hydrological events. It should be given due attention in selecting hydrological models and assigning weights on climate sensitivities in uncertainty assessment of climate change impacts in designing water resources systems.