Arctic Sea Ice in the Community Earth System Model version 2 (CESM2) over the 20th and 21st Centuries

Over the past decades, Arctic sea ice has declined in thickness and extent and is shifting toward a seasonal ice regime. These rapid changes have widespread implications for ecological and human activities as well as the global climate, and accurate predictions could benefit a wide range of stakeholders, from local residents to governmental policy makers. However, many aspects of the polar transient climate response remain poorly understood, particularly in regard to the response of Arctic sea ice to increasing atmospheric CO₂ concentration and warming temperatures. The Coupled Model Intercomparison Project Phase 6 (CMIP6) provides a useful framework for understanding this response, and the participating climate model simulations are a powerful tool for advancing our understanding of present and future changes in the Arctic climate system.

Here we explore the current and future states of Arctic sea ice in the Community Earth System Model version 2 (CESM2), the latest generation of the CESM and NCAR’s contribution to CMIP6. We analyze changes in Arctic sea ice cover in two CESM2 configurations with differing atmospheric components: the “low-top” configuration with limited chemistry (CESM2-CAM) and the “high-top” configuration with interactive chemistry (CESM2-WACCM). We find that the two experiments show large differences in their simulation of Arctic sea ice over the historical period. The CESM2-CAM winter ice thickness distribution is skewed thin, with an insufficient amount of ice thicker than 3 m. This leads to a lower summer ice extent compared to the CESM2-WACCM and observations. In both experiments, the timing of first ice-free conditions is insensitive to the choice of future emissions scenario (known as the shared socioeconomic pathways, or SSPs, in CMIP6), an alarming result that points to the current vulnerable state of Arctic sea ice. However, if global warming stays below 1.5°C, the probability of an ice-free summer remains low, consistent with other recent studies. By the end of the 21^st century, both experiments exhibit an accelerated decline in winter ice extent under the high emissions scenario (SSP5-8.5), leading to ice-free conditions for up to 8 months and an open-water period of 220 days or more depending on the region. Initial results show that the CESM2 simulates less ocean heat loss during the fall months compared to its previous version, delaying the formation of sea ice and leading to lower winter ice extent. Given that the CESM2 reaches a higher atmospheric CO₂ concentration and thus warmer global and Arctic temperatures by 2100, these results suggest the presence of emerging processes associated with a state of the Arctic climate that has never been sampled before.