Analysis of Extreme Heat in Historical and Projected Climate Simulations for Regional Climate Planning Purposes in the U.S.

Abstract:

The U.S. National Climate Assessment (NCA) states that global climate models predict more extreme temperatures and more frequent, intense, and longer heat waves on a regional basis as global temperatures rise throughout the 21^st century, but a thorough test of whether these models can simulate observed heat metrics and trends over the historical period was not included in the assessment. Understanding the capabilities of climate models over the historical period is crucial to assessing our confidence in their predictive ability at regional scales.

Our work fills this research gap by evaluating the performance of Coupled Model Intercomparison Phase 5 (CMIP5) models as compared to observational data using multiple heat metrics. Our metrics are targeted for the southwest United States, but our regional analysis covers the entire continental U.S. and Alaska using 7 of the regions delineated by the NCA. The heat metrics include heat wave and cold wave frequency, intensity, and duration, overnight low temperatures, onset and length of the hot season, and human heat stress. For the best performing models, we compute the same heat metrics for the RCP scenarios. In addition to presenting the results of our CMIP5 historical and RCP analyses, we also describe how our results may be applied to the benefit of our community in Southern Arizona as a case study. Our research will be used by NOAA's Climate Assessment for the Southwest (CLIMAS) and by an interdisciplinary collaborative team of researchers from the University of Arizona working with an electric utility to integrate climate information into their strategic planning.