Extreme events over the continuous United States portrayed in a CESM1–WRF dynamical downscaling framework

Dynamical downscaling in the form of regional climate models (RCMs) has been commonly employed to produce high-resolution simulations over limited regions given computational constraints. Relative to coarse global climate model (GCM) fields, RCMs yield a better representation of complex land-atmosphere interactions such as topographically-forced circulations and mesoscale organization associated with deep convection. These improvements make RCMs a suitable tool for modeling extreme weather events like heavy rainfall and heat waves. In this study, we use output from the Community Earth System Model Version 1.0 (CESM1) to drive the Weather Research and Forecast model (WRF) as a RCM in order to explore the distribution and trend of extreme events across the contiguous United States. RCM simulations are conducted over the majority of North America for the last five decades of the 20th and 21st centuries (1950-1999 and 2050-2099). The trend of 90%, 95% and 99% percentile values from probability density functions (PDFs) of daily precipitation (PR), daily maximum temperature (TMAX), and daily minimum temperature (TMIN) are used to quantify extreme events. The behavior of the RCM is broadly consistent with the large-scale pattern present in the GCM. Over much of the western United States, the RCM produced an increasing trend of extreme events for both precipitation (both frequency and intensity) and temperature. Over the central Great Plains area, the frequency and intensity of extreme precipitation events changes little with time, indicating that the trend in extremes is strongly regionally dependent. Because an RCM forced by GCM output will include large-scale teleconnections, the analysis explores the dependence of extreme events on the phase of the El Nino/Southern Oscillation (ENSO). Thresholds corresponding to extreme values of precipitation and temperature exhibit a strong dependence on ENSO phase and location.