Changes in the frequency or intensity of extreme rainfall events could have serious socio-economic and environmental impacts. Within the Northeastern United States increases in the frequency of extreme rainfall events (more than 2 inches of rain falling in less than 48 hours) have already been documented. Given projected increases in extreme rainfall events and the associated consequences, it is of great importance to evaluate the ability of coupled atmosphere-ocean general circulation models (AOGCMs) to simulate observed extreme rainfall.

This study compares historical observations from the Northeastern United States with simulated and downscaled extreme rainfall from three of the latest Intergovernmental Panel on Climate Change (IPCC) AR4 coupled AOGCMs, the United States National Oceanographic and Atmospheric Administration/Geophysical Fluid Dynamics Laboratory CM2.1, the United Kingdom Meteorological Office Hadley Centre Climate Model version 3 and the United States Department of Energy/National Center for Atmospheric Research Parallel Climate Model.

The most extreme precipitation events over a thirty-year time period at model grid boxes were compared with the stations records and the North American Regional Reanalysis (NARR). The station records were corrected with an Areal Reduction Factor (ARF) to allow more accurate comparison with the model grid boxes. Overall, the models tended to underestimate the most extreme events. However, there was a great deal of variation among the locations and models, and the seasonal pattern of extreme rainfall was well captured.