Storylines of Southern Hemisphere climate change from CMIP6: Antarctica and the Southern Ocean

The manifestation of regional changes in climate at high latitudes is notoriously uncertain according to climate models, aside from the expected polar amplification of the global warming trend. This poses a particular issue for policymakers in designing targeted adaptation and mitigation strategies, in limiting the impact of human-induced climate change. The leading mode of uncertainty arises from the unknown response of the atmospheric circulation to global warming, resulting from inconsistencies in the model representation of key physical processes, which are typically parameterised as they operate on sub-grid spatial scales. 

Using the latest suite of state-of-the-art climate models as part of the sixth phase of the Coupled Model Intercomparison Project (CMIP6), a storyline approach is adopted to derive physically plausible scenarios of climate change over Antarctica and adjacent regions of the Southern Ocean for the end-of-the-century (2070-2099), according to both emission scenarios SSP3-7.0 and SSP5-8.5. We identify two robust features of climate change simulated across all models known to lead to a strengthening and poleward displacement of the eddy-driven mid-latitude jet: (1) the decline in Southern Hemisphere sea ice extent and (2) the strengthening of the wintertime stratospheric polar vortex and delayed summertime breakdown. Whilst the response of these two aspects is consistent in sign, a large intermodel spread exists in terms of magnitude. Using a multi-linear regression framework, we generate storylines of climate change conditional upon the magnitude of change in each driver centred around the multi-model mean response. We examine the response of both the near-surface climate (Southern Annular Mode, air temperature, precipitation) and Southern Ocean (SSTs, salinity and mixed layer depth) for both an extended austral summer (DJFM) and winter season (JJA). Through ancillary evaluations of model historical performance with respect to ERA-5 reanalysis over the 1985-2014 period, including investigation of both prognostic and diagnostic variables, the presented storylines are refined to help minimise the impact of model deficiencies in affecting the realism of the generated storylines. Our results highlight the merit of using the storyline approach to identify future potential scenarios of regional climate response and constrain uncertainty in Antarctic climate predictions. Storyline impact assessments for the Southern Ocean marine ecosystem, particularly the ecologically sensitive South Scotia Sea region, will later be investigated within the EU Horizon 2020 ‘PolarRES’ project through storyline-guided downscaling experiments using regional climate models.