Examining the forced response of past regional climate to guide selection of general circulation models for regional analyses

Both the IPCC and the North American Regional Climate Assessment Project (NARCAP) are focusing on the need to understand regional climate change with improved short- (decadal) and long-term predictive confidence.  This emphasis puts increasing demands on the ability of GCMs to accurately represent regional climate variability and its responses to forcings, whether the GCMs are used directly or to provide boundary conditions for regional climate and impacts models (cf. figure).  Meeting this need requires high temporal and spatial resolution paleoclimate reconstructions because the inherently greater climate variability at the regional scale makes it more difficult to detect and attribute the nature of forced and unforced climate change using only limited instrumental data.  Work is presented that targets this need in the western North American region, using new reconstructions of primary climate variables to drive analyses of regional responses to forcings, especially the post-volcanic temperature response.  The agreement between the reconstructed response patterns and the corresponding responses in GCM simulations of the existing IPCC-archive (CMIP3) will then be examined, with the goal of identifying a regionally best-performing set of GCMs to force regional impacts analyses.

Caption:  Uncertainty in modeled 21st century fire probabilities, driven solely by the GCM data used to provide climate inputs into the fire probability model (Westerling and Bryant, 2006, California Energy Commision. CEC-500-2005-190-SF).  For example, note the change of sign between the scenarios for the highly-populated coastal mountain region of southern California.