The Arctic surface climate in CMIP6: status and developments since CMIP5

TL;DR

This paper assesses the Arctic climate simulation accuracy in CMIP6 models, highlighting improvements over CMIP5 in sea ice representation but noting persistent biases, and projects future Arctic climate stabilization under low emission scenarios.

Contribution

It provides a comprehensive evaluation of CMIP6 Arctic climate simulations, comparing them to CMIP5 and reanalysis data, and offers future projections under low emission scenarios.

Findings

Improved sea ice volume and extent representation in CMIP6.

Persistent biases include winter cold biases and variability issues.

Arctic climate expected to stabilize by 2060 under SSP126 scenario.

Abstract

Here we evaluate the sea ice, surface air temperature, and sea-level-pressure from 31 of the models used in the Coupled Model Intercomparison Project Phase 6 (CMIP6) for their biases, trends, and variability, and compare them to the CMIP5 ensemble and the ERA5 reanalysis for the period 1979 to 2004. The principal purpose of this assessment is to provide an overview of the ability of the CMIP6 ensemble to represent the Arctic climate, and to see how this has changed since the last Phase of CMIP. Overall, we find a distinct improvement in the representation of the sea ice volume, but also in the sea ice extent, mostly linked to improvements in the seasonal cycle in the Barents Sea. However, numerous model biases have persisted into CMIP6 including too-cold conditions in the winter (4 K cold bias) and a negative trend in the day-to-day variability over ice in winter. We find that under the low emission scenario, SSP126, the Arctic climate is projected to stabilize by 2060 with a sea ice extent of around 2.5 million km2 and a temperature 4.7 K warmer than the early 20th century average, compared to 1.7 K of warming globally.