Impacts of climate change on the frequency and severity of floods in the Chateauguay River basin, Canada

This study aims at evaluating the hydrologic impacts of climate change on the Chateauguay River basin in the province of Quebec. Canada. Three global climate models (GCMs) covering a range of climate sensitivities were selected, and their output was employed to adjust the parameters of a stochastic weather generator using simple transformation rules for precipitation and temperature. Values of monthly precipitation and temperature were extracted from the GCMs for the current (1960-1990) and future (2040-2060) climate. The International Panel on Climate Change emission scenario known as 132 was selected. It represents an average scenario and corresponds approximately to a doubling of the atmospheric CO2, concentration. Resorting to stochastically generated climate scenarios allowed assessing whether the modelled effects of climate change on flows were statistically significant. Results indicate that spring and summer-fall peak flows were reduced on average by 30% and 12%, respectively, using the Echam4 model derived scenarios. The Hadcm3 model produced a weaker signal that was not statistically significant. The CGCM2 model produced a statistically significant reduction in spring peak flows of 8% on average, whereas the simulated reduction in summer flows was not statistically significant for many of the return periods considered. Many sources of uncertainties were partially considered in this study. One is the downscaling of the GCM climatology at the watershed scale. The approach employed to generate the future climate scenarios changed the precipitation variability through an adjustment of the parameters of the Gamma distribution function used to model precipitation amounts. Whether this approach is truly typical of climate change effect remains to be ascertained. Using more physically based hydrological models would help reduce uncertainties in climate change impacts studies.