Precipitation Extremes in western U.S. urban areas: How reliable are regional/global climate model projections?

Understanding the nature of precipitation extremes under projected future climate is essential to stormwater infrastructure planning. Several studies based on observational data suggest that the hydrologic cycle is intensifying as the climate of the western U.S. warms, and extreme precipitation events are increasing. Future climate projections rely primarily on simulations of global climate models (GCMs) downscaled using either statistical models or regional climate models (RCMs). However, relatively little work has been done evaluating the ability of dynamical downscaling using RCMs to simulate future precipitation extremes. The key question we seek to answer is: how effective are RCM-downscaled climate simulations in reconstructing observed sub-daily to daily scale frequency distributions of precipitation extremes over the major urban areas of the western United States? We use hourly precipitation data from National Climate Data Center (NCDC), sub-daily precipitation output from the selected RCMs that participated in the North American Regional Climate Change Assessment Program (NARCCAP), and GCM output from the models that provided the NARCCAP boundary conditions. Both annual and seasonal observed precipitation extremes at sub-daily and daily scales are compared for 14 urban areas. Empirical fitted distributions and summary statistics of annual maxima and peaks over threshold are from both observations and the climate models.