New evaluation of a regional climate modeling effort for the western United States using a superensemble from weather@home

Sihan Li, Philip W. Mote, David E. Rupp, Dean Vickers, Robert Mera, and Myles Allen (2015) Evaluation of a regional climate modeling effort for the western United States using a superensemble from weather@home Journal of Climate doi: 10.1175/JCLI-D-14-00808.1

Abstract

Simulations from a regional climate model (RCM) as part of a superensemble experiment were compared with observations of surface meteorological variables over the western United States (US). The RCM is the Hadley Centre Regional Model with improved physics parameterizations (HadRM3P) run at 25-km resolution and nested within the Hadley Centre Atmospheric General Circulation Model (HadAM3P). Overall, the means of seasonal temperature were well represented in the simulations: 95% of grid points were within 2.7°C, 2.4°C, and 3.6°C of observations in winter, spring, and summer, respectively. The model was too warm over most of the domain in summer except central California and southern Nevada. HadRM3P produced more extreme temperatures than observed. The overall magnitude and spatial pattern of precipitation were well characterized, though HadRM3P exaggerated the orographic enhancement along the coastal mountains, Cascades and Sierras. HadRM3P produced warm-dry Northwest, cool-wet Southwest US patterns associated with El Niño. However, there were notable differences including the locations of the transition from warm (dry) to cool (wet) in the anomaly fields when compared with observations, though there was disagreement among observations. HadRM3P simulated the observed spatial pattern of mean annual temperature more faithfully than any of the RCM/GCM pairings in the North American Regional Climate Change Assessment Program (NARCCAP). Errors in mean annual precipitation from HadRM3P fell within the range of errors of the NARCCAP models. Lastly, this paper provided examples of the size of an ensemble required to detect changes at local level and demonstrated the effect of parameter perturbation on regional precipitation.