Biases in CMIP6 Historical U.S. Severe Convective Storm Environments Driven by Biases in Mean‐State Near‐Surface Moist Static Energy. Issue 23 (9th December 2022)

• Record Type: Journal Article
• Authors: Chavas, Daniel R.; Li, Funing
• Is Part Of: Geophysical research letters. Volume 49:Issue 23(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-12-09
• Subjects: severe convective storms -- CMIP6 -- climate model -- CAPE -- moist static energy -- severe weather; Geophysics -- Periodicals; Planets -- Periodicals; Lunar geology -- Periodicals; 550
• DOI: 10.1029/2022GL098527
• Languages: English
• ISSNs: 0094-8276

Abstract

Abstract: This work evaluates how well Coupled Model Intercomparison Project 6 models reproduce the climatology of North American severe convective storm (SCS) environments in ERA5 reanalysis and examines what drives biases across models. Biases in spring SCS environments vary widely in magnitude and spatial pattern, though most models do well in reproducing the climatological pattern and a few (MPI and CNRM) also reproduce the overall magnitude. SCS biases are driven by biases in extreme convective available potential energy. These biases are ultimately found to be driven by biases in mean‐state near‐surface moist static energy, indicating that the SCS environments depend strongly on the near‐surface mean state. Results are similar for fall, but not summer or winter when free‐tropospheric biases are also important. Biases differ strongly across parent models but weakly across child models of the same parent. These outcomes help identify models well‐suited for studying climate effects on SCS environments. Plain Language Summary: Climate models are useful tools for studying how severe thunderstorms may change with climate change. Models cannot simulate the storms themselves but can simulate environments that support severe thunderstorms. Using the most recent set of models used to simulate future projections of climate change, we find that some models can simulate the historical climatology of these environments very well, while others do not. We also show that model errors in severe thunderstorm environments arise primarily due to errors in the average temperature and moisture of near‐surface air. Hence, simulating these extreme environments well depends strongly on simulating average conditions well. Key Points: Most models reproduce the pattern of severe convective storm (SCS) environments over North America, and some also reproduce the magnitude Spring SCS environment biases are driven by mean‐state near‐surface moist static energy biases Hence, simulating SCS environments well requires simulating the near‐surface mean state air properties well