Assessment and Ranking of Climate Models in Arctic Sea Ice Cover Simulation: From CMIP5 to CMIP6

Abstract The capability of 36 models participating in phase 6 of the Coupled Model Intercomparison Project (CMIP6) and their 24 CMIP5 counterparts in simulating the mean state and variability of Arctic sea ice cover for the period 1979–2014 is evaluated. In addition, a sea ice cover performance score for each CMIP5 and CMIP6 model is provided that can be used to reduce the spread in sea ice projections through applying weighted averages based on the ability of models to reproduce the historical sea ice state. Results show that the seasonal cycle of the Arctic sea ice extent (SIE) in the multimodel ensemble (MME) mean of the CMIP6 simulations agrees well with observations, with a MME mean error of less than 15% in any given month relative to the observations. CMIP6 has a smaller intermodel spread in climatological SIE values during summer months than its CMIP5 counterpart. In terms of the monthly SIE trends, the CMIP6 MME mean shows a substantial reduction in the positive bias relative to the observations compared with that of CMIP5. The spread of September SIE trends is very large, not only across different models but also across different ensemble members of the same model, indicating a strong influence of internal variability on SIE evolution. Based on the assumptions that the simulations of CMIP6 models are from the same distribution and that models have no bias in response to external forcing, we can infer that internal variability contributes to approximately 22% ± 5% of the September SIE trend over the period 1979–2014.