Generation of regional climate change scenarios using general circulation models and empirical downscaling
Coupled ocean-atmosphere general circulation models (GCMs) are the best tools available for examination of climate change due to increases in atmospheric greenhouse gas concentrations. Due to large computational requirements, these numerical models run at horizontal resolutions that are inadequate for climate impact studies and, hence, require parameterization of many small-scale processes important for characterization of regional climate. The aim of this research was to develop and evaluate a methodology for generating regional climate change scenarios for the Midwest region of the USA using GCM simulations and empirical downscaling methods. The research focuses on (1) identification of relationships between large-scale predictors and three surface parameters (local maximum and minimum daily surface air temperature and total daily precipitation) at 84 stations in the study region, (2) evaluation of variables simulated by two GCMs, and (3) development and evaluation of empirical downscaling tools to generate projections of the surface parameters for the 21 st century.
The results of the analysis indicate that the large-scale atmospheric predictors explain a large proportion of the variability in the surface parameters, but that GCM simulations of the large-scale predictors do not exhibit an acceptable level of agreement with observations at the grid point level. Therefore, the downscaling models applied in this study are based on (1) relationships between GCM simulated variables and the surface parameters and (2) spatially aggregated predictor information.
The downscaled climate change scenarios indicate strong warming at most stations consistent with projected increases in greenhouse gases. Averaged over all stations, the downscaled results indicate year-round warming, but the magnitude of the 21st century temperature change is inconsistent between results downscaled from the two GCMs used. These results show that, under the emissions scenarios used by the GCMs, important climate change impacts such as increases in heat wave frequency may be realized, although there is a high degree of uncertainty associated with these findings. The downscaled precipitation scenarios are less consistent than those for temperature (in terms of both the direction and magnitude of precipitation change and its spatial coherence), resulting in lower confidence for the precipitation scenarios relative to those for temperature.