Date of this Version
As the debate on potential climate change continues, it is becoming increasingly clear that the main concerns to the general public are the potential impacts of a change in the climate on societal and biophysical systems. In order to address these concerns researchers need realistic, plausible scenarios of climate change suitable for use in impacts analysis. It is the purpose of this paper to present a downscaling method useful for developing these types of scenarios that are grounded in both General Circulation Model simulations of climate change, and in situ station data. Free atmosphere variables for four gridpoints over the Missouri, Iowa, Nebraska, Kansas (MINK) region from both control and transient simulations from the GFDL General Circulation Model were used with thirty years of nearby station data to generate surface maximum and minimum air temperatures and precipitation. The free atmosphere variables were first subject to a principal components analysis with the principal component (PC) scores used in a multiple regression to relate the upper-air variables to surface temperature and precipitation. Coefficients from the regression on station data were then used with PC scores from the model simulations to generate maximum and minimum temperature and precipitation. The statistical distributions of the downscaled temperatures and precipitation for the control run are compared with those from the observed station data. Results for the transient run are then examined. Lastly, annual time series of temperature for the downscaling results show less warming over the period of the transient simulation than the time series produced directly from the model.