85th AMS Annual Meeting

Tuesday, 11 January 2005
Influences on predictability in multi-year regional climate simulations for the continental United States
Raymond W. Arritt, Iowa State University, Ames, IA; and C. J. Anderson, W. J. Gutowski, E. S. Takle, D. Caya, C. G. Jones, J. J. Katzfey, J. W. Larson, R. Laprise, J. L. McGregor, J. Roads, and J. Taylor
The Project to Intercompare Regional Climate Simulations (PIRCS) is an international community-based project that provide a framework for systematic intercomparison of regional climate models (RCMs) and their component parameterizations. The present PIRCS experiment is 1C, which compares decadal-scale RCM simulations over North America. Presently five modeling groups are engaged in PIRCS 1C using seven RCMs.

Preliminary examination of PIRCS 1C results has focused primarily on the North American monsoon. Results indicate that RCMs can replicate the monsoon core region in northwestern Mexico but extension of the monsoon into the southwestern United States is a greater challenge for RCMs. Modeling groups also are engaged in multiple simulations in order to test uncertainties owing to influences such as internal model variability or choice of physical parameterization schemes. Internal variability has been examined by constructing an ensemble of simulations initialized at one-month intervals. Effects of internal variability are small, suggesting that these RCM simulations are strongly constrained by lateral boundary information despite the use of a large domain. The choice of convective parameterization has a much greater effect on RCM results. Different cumulus parameterizations not only produce different intensity of monsoon convection but also affect teleconnections of the North American monsoon with summer precipitation across the eastern two-thirds of the United States.

Supplementary URL: