The Coupled Atmosphere–Ocean Response to Antarctic Sea Ice Loss

Abstract

Antarctic sea ice is projected to decrease in response to increasing greenhouse gas concentrations. Limited studies so far have examined the coupled atmosphere–ocean response to Antarctic sea ice loss. Here, we isolate the response to Antarctic sea ice loss in the atmosphere and ocean using bespoke sea ice albedo perturbation experiments with HadGEM3-GC3.1-LL, provide the first detailed examination of the global ocean response, and quantify the importance of atmosphere–ocean coupling, through comparison to uncoupled experiments with prescribed Antarctic sea ice loss. Lower-tropospheric warming and moistening over regions of sea ice loss and the nearby Southern Ocean are simulated in both coupled and uncoupled configurations but are of greater magnitude in the coupled model. A weakening and equatorward shift of the tropospheric westerly jet are simulated in both configurations, but are also larger in the coupled model. Ocean coupling allows the warming response to spread northward, and by poleward atmospheric energy transport, back to the Antarctic interior. Warmer tropical sea surface temperatures enhance atmospheric convection, driving upper-tropospheric warming and triggering atmospheric teleconnections to the extratropics, including a weakened Aleutian low. A 20% reduction in Antarctic Circumpolar Current transport and a weakening of the shallow tropical convergence cell are simulated. Surface waters warm and freshen globally, becoming more stratified and stable in the Southern Ocean, with similar changes, but of lesser magnitude, in the Arctic Ocean, where sea ice declines. Our results suggest that the climate effects of Antarctic sea ice loss stretch from pole to pole and from the heights of the tropical troposphere to the depths of the Southern Ocean.

Simple Hybrid Sea Ice Nudging Method for Improving Control Over Partitioning of Sea Ice Concentration and Thickness

“…Similarly to Ayres et al. (2022), the AGCM shows a SLP reduction over areas of sea ice loss while the coupled response (with the hybrid nudging in this case), displays a Southern Annular Mode type response.…”

“…In addition, the coupled response to Antarctic sea ice loss exhibits a tropical upper tropospheric warming (Figures 7e and 7f). This signal has been shown to arise more strongly with Antarctic sea ice loss than Arctic sea ice loss (Ayres et al., 2022; England, Polvani, Sun, & Deser, 2020). In the austral summer (Figure S5 in Supporting Information ), the AGCM surface warming remains stronger while remaining closer to the continent.…”

“…However, the most striking difference is with the hybrid nudging response (Figure 10c), where there is a strong positive SLP response at higher latitudes and weaker, but still significant, reductions in the midlatitudes, mostly over the southern part of South America. Similarly to Ayres et al (2022), the AGCM shows a SLP reduction over areas of sea ice loss while the coupled response (with the hybrid nudging in this case), displays a Southern Annular Mode type response.…”

Net Equatorward Shift of the Jet Streams When the Contribution From Sea‐Ice Loss Is Constrained by Observed Eddy Feedback

“…This is because the response to Antarctic sea-ice loss is small in the northern hemisphere and the response to Arctic sea-ice loss is small in the southern hemisphere. We note these experiments lack ocean coupling, which can lead to interhemispheric responses to sea-ice loss (Deser et al, 2015); however even with an interactive ocean, these remote responses are small compared to the more local response to sea-ice loss within the respective hemisphere (Ayres et al, 2022;England et al, 2020). For the austral jet, 9 of 10 models show an equatorward shift in response to sea-ice loss, and 8 of 9 models show a poleward shift in response to SST (Figure 4a).…”

“…The boreal (northern hemisphere) and austral (southern hemisphere) jets are simulated by climate models to shift equatorward and weaken in response to Arctic and Antarctic sea‐ice loss (e.g., Ayres & Screen, 2019; Ayres et al., 2022; England et al., 2018; Screen & Blackport, 2019; Screen et al., 2018; Smith et al., 2022; Zappa et al., 2018), respectively, albeit with uncertain magnitude. Whilst many studies have examined the responses to projected Arctic and Antarctic sea‐ice loss separately, there has been limited quantitative comparison of the two.…”