Assessment of AOGCM biases for RCM applications over the Great Basin in the western U.S

This study addresses the impact of Atmosphere-Ocean General Circulation Model's (AOGCM) biases on the accuracy of Regional Climate Modeling (RCM) impact studies. We have analyzed AOGCM model outputs from the World Climate Research Programme's (WCRP's) Coupled Model Inter-comparison Project phase 3 (CMIP3). Most of the AOGCMs outputs show subtropical warm SST biases over the eastern oceanic boundaries. In particular, offshore of California and the Baja California Peninsula the modeled coastal SSTs are higher than observations by ~ 3-5°C with seasonal fluctuations and the warmest biases occurring between May and November. Our hypothesis is that such AOGCM SST biases will have a significant impact on simulated climate change and variability for climate parameters over the Great Basin in the western U.S. Thus, a principal question is to what extent the AOGCM SST fields need to be improved by, for example, using first order fixes, bias correction, and/or proscribing observed values to increase the accuracy of dynamically downscaled simulations. To address this question, a regional climate model (RCM) based on high-resolution WRFV3.1.1 (36 and 12 km) was integrated for a period of 1998-2007 using CCSM3 forcing with and without corrections of the SST biases. We have replaced the AOGCM SSTs with NCEP optimum interpolation (OI) SST. Even though inconsistencies in the evolution of atmospheric processes may have appeared due to the new SSTs, this approach allowed us to easily evaluate whether a priori corrections of the SST would significantly affect RCM outputs locally and remotely. Our results indicate that the precipitation fields over the Great Basin significantly change and are between 25-50% larger when the observed SSTs are used compared to the CCSM3 SSTs. An in depth analysis is presented to address the physical connections associated with such rainfall changes.