The Performance of CMIP6 Versus CMIP5 in Simulating Temperature Extremes Over the Global Land Surface

Xuewei Fan,Chiyuan Miao,Qingyun Duan,Chenwei Shen,Yi Wu

DOI: https://doi.org/10.1029/2020JD033031

2020-01-01

Abstract:Simulations from the models participating in the sixth phase of the Coupled Model Intercomparison Project (CMIP6), which represent the most recent generation of climate models, are now available. Understanding the performance of these models in simulating historical climate extremes can provide a basis for producing reliable future climate projections. Here, we assess the simulation of 16 indices of temperature extremes defined by the Expert Team on Climate Change Detection and Indices using results from 24 CMIP6 models as compared with results from CMIP5. Comparisons with observations and reanalyses indicate that the CMIP6 models could capture the spatial patterns and temporal variations of the observed temperature extremes well for some indices, although less well for others. Based on spatial and temporal skill scores, CMIP6 ensemble means were more skillful in simulating absolute and threshold indices of extreme temperature than CMIP5 ensemble means were, but the performances of both the CMIP5 and CMIP6 ensemble means in simulating the spatial patterns for duration and percentile indices were relatively unsatisfactory (spatial skill scoresS < 0.3). Furthermore, our results suggest that there have been improvements in spatial pattern skill scores in some individual CMIP6 models relative to CMIP5 model scores for summer days, tropical nights, cold spell duration, and diurnal temperature range.