Comparisons of North American Regional Climate Change Assessment Program (NARCCAP) GFDL & CRCM Simulations with Observations

The North American Regional Climate Change Assessment Program (NARCCAP) is an international program to produce high resolution climate change simulations in order to investigate uncertainties in regional scale projections of future climate and generate climate change scenarios for use in impacts and decision aid research. A set of regional climate models (RCMs) driven by a set of atmosphere-ocean general circulation models (AOGCMs) over a domain covering the United States and adjacent land areas have been performed by various international modeling groups. The AOGCMs have been forced with the IPCC's Special Report on Emissions Scenarios (SRES) A2 emissions scenario for the 21st century. The A2 scenario is at the higher end of the SRES emissions scenarios. Simulations with these models were also produced for a historical period. The RCMs are nested within the AOGCMs for the historical period 1968-2000 and for the future period 2038-2070.

We evaluate the runs produced by the Geophysical Fluid Dynamics Laboratory (GFDL) time slice experiments as well as the runs produced by the Canadian Regional Climate Model (CRCM). We are interested in comparing temperature, rainfall and clouds from these models to surface reporting stations obtained from the Global Historical Climatology Network (GHCN). The cloud simulations from the RCM's are compared to a satellite derived cloud analysis from the NOAA GOES series of satellites over the United States.

Results indicate that large differences between the RCM derived clouds and those from the satellite analysis. The GFDL model has a difficult time simulating the correct occurrence of the marine stratus clouds off the Southern California coastline. On the other hand the model greatly over predicts the occurrence of clouds over the peninsula of Florida. The temporal correlations of clouds at various locations in CONUS indicate an unusual pattern, which appears unrealistic when compared to observations. Temperature and precipitation patterns from both the GFDL and CRCM runs show systematic biases, which vary in magnitude depending on location. Additional more detailed results will be reported on at the conference.