Role of sea surface temperature and sea-ice changes in high-latitude climate change: a multi-model experiment within the CRiceS H2020 project.

Dramatic sea ice loss has recently occurred at both poles. Multiple studies have suggested that changes to sea ice can impact weather in both the polar regions and mid-latitudes. However, the current generation of climate models disagrees on the rate and location of sea ice loss, and on the rate of warming in the polar regions.Thus, the atmospheric response to sea ice loss within and outside the polar regions remains highly uncertain. To reduce this uncertainty, we have performed a set of coordinated simulations with four different atmospheric general circulation models (AGCMs) within the project “Climate Relevant interactions and feedbacks: the key role of sea ice and Snow in the polar and global climate system” (CRiceS). A baseline simulation and six perturbation simulations were performed, all of which were 40-years long and had prescribed  sea surface temperatures (SSTs) and sea ice concentration. In the perturbation simulations, the SSTs and sea ice concentration were changed independently, and then both were changed together. The SST and sea ice concentrations were obtained from CMIP6 simulations with the Australian Earth system model ACCESS-ESM1.5. Monthly-mean SST and sea-ice area averaged over 20 years of simulation were taken from 1) the historical simulation (years 1950-1970, Baseline simulation), 2) the scenario SSP1-2.6 simulation (years 2080-2100), and 3) the scenario SSP5-8.5 simulation (years 2080-2100) and were then used as perpetual monthly average values of SSTs and sea ice fraction in our model simulations, thus eliminating inter-annual variability in SSTs and sea ice. This array of perturbation experiments, performed with four AGCMs, allows us to isolate atmospheric responses in polar regions and mid-latitudes that are due to SST or sea ice changes, examine the linearity of these feedbacks, and investigate the robustness of the atmospheric responses. The results of this coordinated modelling experiment show that the models agree well on the magnitude and spatial distribution of the 2-m temperature and precipitation response. Increasing SSTs has a larger and more spatially extensive impact on the overall response than decreases in sea ice,  which primarily only cause a localised response in regions where sea ice disappears (most notably, a strong warming over the Arctic ocean in winter). The models agree less well on the magnitude and spatial distribution of the mean sea level pressure response, in particular over northern Europe and Antarctica, suggesting that modelled uncertainties associated with atmospheric circulation are larger than uncertainties associated with thermodynamics. These results and others, along with information about the openly available dataset, will be presented.