Search

Summary: The impact of climate change on the hydrology of the Chute-du-Diable watershed (Quebec, Canada) is studied by comparing statistics on current and projected future discharge resulting from a wide range of climate change scenarios. The use of 10 equally weighted climate projections from a combination of 5 general circulation models (GCMs) and 2 greenhouse gas emission scenarios (GHGES) allows for the definition of an uncertainty envelope of future hydrologic variables. GCM data is downscaled using the change factor approach for 30-year time slices centered around years 2020, 2050 and 2080. To estimate natural variability, synthetic time series are then computed for each horizon and for each climate change scenario, using a stochastic weather generator (30 series of 30 years), and are entered into a hydrology model. Future hydrological regimes are then compared to the control period (1961–1990) using the annual and seasonal mean discharge, peak discharge and timing of peak discharge criteria. Results indicate a 1–14°C increase in seasonal temperature and a −9 to +55% change in seasonal precipitation. The largest increases in both temperature and precipitation are observed in the winter and spring seasons. The main hydrologic impact observed is a spring flood appearing 1–5 weeks earlier than usual and a modification of its amplitude from −40 to +25%. Most scenarios suggest increases in the winter, spring and fall discharge, whereas summer is expected to see a decrease in discharge. While there is still a large scatter in projected values, the uncertainty analysis projects a better view of the most probable future hydrologic behaviour of the watershed. Of all sources of uncertainty considered in this study, the largest comes from the choice of a GCM. Accordingly, all impact studies based on results from only one GCM should be interpreted with caution. [Copyright &y& Elsevier]

This study aims at evaluating the hydrologic impacts of climate change on the Châteauguay 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 B2 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. [ABSTRACT FROM AUTHOR]