Evaluation of Antarctic sea ice thickness and volume during 2003–2014 in CMIP6 using Envisat and CryoSat-2 observations

Abstract

Sea ice thickness (SIT), which is a crucial and sensitive indicator of climate change in the Antarctic, has substantial impact on atmosphere–sea-ice–ocean interactions. Despite the slight thinning in SIT and reduction in sea ice volume (SIV) in the Antarctic in the recent decade, challenges remain in quantifying their changes, primarily because of limited availability of high-quality long-term observational data. Therefore, it is crucial to accurately simulate Antarctic SIT, and to assess the SIT simulation capability of state-of-the-art climate models. In this study, we evaluated historical simulations of SIT by 51 climate models of the Coupled Model Intercomparison Project Phase 6 (CMIP6) using Envisat (ES) and CryoSat-2 (CS2) observations. Results revealed that most models can capture the seasonal cycles in SIV, and that the CMIP6 multimodel mean (MMM) can reproduce the increasing and decreasing trends in the SIV anomaly based on ES and CS2 data, although the magnitudes of the trends in the SIV anomaly are underestimated. Additionally, the intermodel spread in simulations of SIT and SIV was found reduced (reduced by 43%) from CMIP5 to CMIP6. Nevertheless, based on the CMIP6 MMM, substantial underestimation in SIV of 57.52% and 59.66% was found compared to that derived from ES and CS2 observations, respectively. The most notable underestimation in SIT was located in the sea ice deformation zone surrounding the northwestern Weddell Sea, coastal areas of the Bellingshausen and Amundsen seas, and the eastern Ross Sea. The substantial bias in the simulated SIT might result from deficiencies in simulating critical physical processes such as ocean heat transport, sea ice–ocean interactions, and dynamic sea ice processes. Therefore, increasing the model resolution, and improving the representation of sea ice dynamics and physical processes controlling the sea ice–ocean interactions are essential for improved accuracy of Antarctic sea ice simulation.