CMIP6 Representation of Declining Sea Ice and Arctic Cyclones in the Current Climate

Abstract The Arctic climate system is changing rapidly with important implications in the Arctic and beyond. The interaction between the sea ice and Arctic cyclones makes it an important topic to be understood in the warming climate. We analyzed an ensemble of Coupled Multimodel Intercomparison Project (CMIP6) model simulations from 1985 to 2014 to determine model skill in depicting Arctic cyclones and their relationship with sea ice. A comprehensive climatology of Arctic cyclones and sea ice concentrations (SIC) was produced and compared to the ERA5 reanalysis product. The models reproduced the observed sea ice spatial patterns and trend well. However, the models struggled to capture the concurrent patterns and trends in Arctic cyclone characteristics that were evident in the reanalysis data. The models underestimated local cyclogenesis in the Arctic, which led to an overall underestimation of Arctic cyclone counts. Lead/lag analysis of ERA5 data suggests that reduced sea ice in the warm season can drive increased cyclone counts in the following cold season, which then reduces SIC in the next warm season in a feedback cycle that appears to be missing from the CMIP6 models. The results also revealed deviations between CMIP6 and ERA5 cyclone intensities. The magnitude and sign of the intensity differences varied based on model resolution, surface roughness parameterization, and skill in the representation of cyclogenesis location. This work highlights the need to improve sea ice‐atmosphere interactions and the representation of synoptic systems in the next generation of global models.