How Credibly Do CMIP6 Simulations Capture Historical Mean and Extreme Precipitation Changes?

• Published in: Geophysical research letters Vol. 50; no. 14
• Main Authors: Donat, Markus G., Delgado‐Torres, Carlos, Luca, Paolo, Mahmood, Rashed, Ortega, Pablo, Doblas‐Reyes, Francisco J.
• Format: Journal Article
• Language: English
• Published: Washington John Wiley & Sons, Inc 28.07.2023; Wiley
• Subjects: Availability; Climate change; Climate models; Climate prediction; CMIP6; Evaluation; Extreme weather; extremes; Future precipitation; Global warming; historical climate simulations; Intercomparison; Modelling; Precipitation; Precipitation trends; Regions; Simulation; Trends; Trustworthiness; Asia; North America; Europe
• ISSN: 0094-8276; 1944-8007
• DOI: 10.1029/2022GL102466

Future precipitation changes are typically estimated from climate model simulations, while the credibility of such projections needs to be assessed by their ability to capture observed precipitation changes. Here we evaluate how skillfully historical climate simulations contributing to the Coupled Model Intercomparison Project Phase 6 (CMIP6) capture observed changes in mean and extreme precipitation. We find that CMIP6 historical simulations skillfully represent observed precipitation changes over large parts of Europe, Asia, northeastern North America, parts of South America and western Australia, whereas a lack of skill is apparent in western North America and parts of Africa. In particular in regions with moderate skill the availability of very large ensembles can be beneficial to improve the simulation accuracy. CMIP6 simulations are regionally skillful where they capture observed (positive or negative) trends, whereas a lack of skill is found in regions characterized by negative observed precipitation trends where CMIP6 simulates increases. Plain Language Summary Climate models are the primary tools to predict future changes in precipitation related to global warming. These predictions can however only usefully inform adaptation measures if they can be trusted. Here we evaluate the trustworthiness of climate model‐simulated precipitation changes based on their capability to correctly capture observed precipitation changes. We apply skill measures commonly used for the evaluation of seasonal to decadal climate predictions to historical climate simulations. We perform this analysis for total precipitation accumulations and indicators of precipitation extremes. The level of skill differs between regions and can be sensitive to the number of available simulations, with some regions benefitting from very large simulation ensembles. Mean and extreme precipitation are skillfully predicted in similar regions, including large parts of Europe and Asia. Lack of skill typically occurs in regions where observed precipitation is characterized by downward trends but Coupled Model Intercomparison Project Phase 6 models simulate increases. This study helps understand the trustworthiness of climate simulations to realistically capture precipitation changes, identifying regions where current models are more or less capable. Key Points Coupled Model Intercomparison Project Phase 6 (CMIP6) realistically simulates observed changes in mean and extreme precipitation in large parts of Europe and Asia and other land regions In regions with moderate skill and observed precipitation subject to multi‐decadal variations the availability of very large ensembles is beneficial Lack of skill occurs primarily in regions where negative precipitation trends are observed but CMIP6 simulates increases