IWRA Proceedings

AbstractThe province of Manitoba, Canada, is blessed with abundant surface water resources but lacks in weather stations where the water resources are. As a result, hydrological modeling and climate change impact assessments for water resources management face difficulty mainly due to limited input data. Recent studies found that the North American Regional Reanalysis (NARR) has high potential for use as input data for hydrological modeling and statistical downscaling of GCM data. The objective of this study is (1) to utilize the NARR data for hydrologic modeling and statistical downscaling of GCM data and (2) to assess the climate change impact on Manitoba water resources. Two catchments (the Taylor and Burntwood River basins) in northern Manitoba and two (the Sturgeon and Troutlake River basins) in north-western Ontario were selected for this study. The SLURP model was set up with meteorological input data from NARR and calibrated for each catchment against the observed streamflow data. K -Nearest Neighbour (k-NN) resampling, a statistical downscaling technique, was used to downscale the output from the recent Canadian GCM (CGCM3) simulation under the IPCC SRES A2 emission scenario (2081-2100). The downscaled CGCM3 data were used as input to the calibrated SLURP model to assess the future climate change impact on water resources. The results indicate that (1) the SLURP hydrological model can be reasonably calibrated with the meteorological input data from NARR, (2) the results from the statistical downscaling with NARR are comparable to those with weather station data, and (3) the warmer and wetter climate in the future is likely to increase the runoff. NARR is found to be a good alternative to weather station data for climate change impact studies in data-scarce central Canada, where higher risk of flooding and lower risk of extended droughts are projected due to climate change.