Regional-scale climate prediction from the GISS GCM

The complexity of the global climate system suggests that the only way in which we can develop a reasonable assessment of future climate change is through the use of general circulation models (GCMs) of the atmosphere and oceans. Current climate change experiments, however, are limited by the inability of the models to produce accurate simulations of the regional climate. The Goddard Institute for Space Studies (GSISS) 4° × 5° GCM is shown to simulate the present-day sea level pressure over the United States very accurately, but this paper also demonstrates large regional biases in the GISS simulations of present-day temperatures. An analysis of circulation-temperature relationships with observational data shows that accurate transfer functions can be derived between sea level pressure and temperature. When applied to the GCM circulation, these transfer functions result in a temperature distribution that matches the long term mean more accurately than does the model's actual temperature field. Applying these transfer functions to the 2 × CO₂ circulation presents the possibility for an alternative approach to regional-scale climate change predictions that combines GCM simulations with empirically observed circulation-climate relationships.