Evaluation of the CMIP6 models for simulating the trend of the Barents-Kara Sea compound heatwaves in boreal autumn

Abstract

Compound heatwaves (CHWs) pose more severe environmental and socio-economic risks than individual heatwave events, with their frequency increasing rapidly worldwide. In the Arctic, particularly the Barents–Kara Sea (BKS), CHWs are intensifying, accelerating ice melt, permafrost thaw, and ecosystem disruptions. Accurate climate projections require a robust assessment of climate models' ability to simulate these extreme events. This study evaluates the performance of 24 CMIP6 models in reproducing autumn CHW trends in the BKS, revealing substantial intermodel discrepancies in trend magnitudes and spatial patterns. The observed BKS regional CHW trend is 0.55 events per decade. Group A models (e.g., ACCESS-CM2, CanESM5, GFDL-CM4) exhibit smaller errors, with a mean absolute bias of −0.03 events per decade, while Group B models (e.g., FGOALS-g3, GFDL-ESM4, MIROC6) show larger discrepancies, with a mean absolute bias of −0.40 events per decade. Further analysis reveals that Group A models better simulate the enhanced moisture transport from the Atlantic to the BKS, a key driver of CHWs, with smaller deviations in latent heat flux and net longwave radiation. The results highlight the significant role of water vapor in the occurrence of CHWs and suggest that improving the representation of these processes could enhance the accuracy of future Arctic CHW projections. This study underscores the importance of model evaluation for improving climate predictions and understanding the dynamics of extreme heat events in high-latitude regions.