Assessment of Various Statistical Downscaling Methods for Downscaling Precipitation in Florida

Several statistical downscaling models have been developed in the past couple of decades to assess the hydrologic impacts of climate change by projecting the station-scale hydrological variables from large-scale atmospheric variables simulated by general circulation models (GCMs). GCMs have a good chance to capture the large-scale circulation patterns and correctly model smoothly varying fields such as surface pressure, but it is extremely unlikely that these models properly reproduce nonsmooth fields such as precipitation. This study compares five statistically downscaling models: multiple linear regression (MLR), positive coefficient regression (PCR), stepwise regression (SWR), support vector machine (SVM), and bias-correction spatial disaggregation (BCSD) for estimation of rainfall in the state of Florida, USA. The performance of the models is evaluated using various performance measures, and it was found that the SVM model performed better than all the other models in reproducing most monthly rainfall statistics at 18 locations. Output from the third-generation Canadian Global Climate Model (CGCM3) GCM for A1B scenario was used for future precipitation projections.