After a gradual increase from the start of regular satellite observations, in 2016 Antarctic sea ice extent experienced a dramatic reduction of around 1 million km^2 that persists to this day. It is not clear whether this change is part of normal internal variability, or a response to external forcing which could have pushed the system into a potentially irreversible critical transition. Based on dynamical systems theory, previous research interpreted observed increases in variability and autocorrelation as signs of critical slowing down preceding a critical transition. Building on this framework, we show that autocorrelation and standard deviation have been relatively stable since the 2000s when computed with respect to the changing mean. Using the longer CMIP6 historical time series, we show that models do not capture the observed changes in mean-state, and we detect no consistent change in autocorrelation and standard deviation. Some CMIP6 models simulate significant changes in the seasonal cycle of Antarctic sea ice extent, which can lead to spurious trends in standard deviation and autocorrelation if not taken into account. Our results suggest that the new low sea ice state represents a dramatic reduction in the mean state without a fundamental change in variability. Further work examining indicators of critical transitions is needed to better contextualise the low sea ice state and the possibility of future sea ice recovery.

Antarctic sea-ice has declined over the past decade with growing evidence linking these losses to warming of Circumpolar Deep Water (CDW). Yet the sensitivity of Antarctic sea-ice to CDW warming is poorly constrained, and it is unclear whether continued ocean warming could trigger an irreversible tipping point in the sea-ice system. Here we investigate the response of Antarctic sea-ice to CDW warming using ensemble climate model experiments and a theoretical framework for winter ocean stability. Using ACCESS-CM2, we conduct three experiments, with 10 ensemble members each, in which CDW is warmed by 0.25°C, 0.5°C, and 1.0°C. Antarctic sea-ice declines in all experiments, but the response is nonlinear with similar losses in the 0.25°C and 0.5°C experiments, and substantially larger declines in the 1.0°C experiment. Despite these losses, all experiments stabilise at reduced sea-ice states with no runaway declines. This stabilisation is driven by enhanced winter ocean heat loss, which acts as a negative feedback on further sea-ice loss. Our results suggest that CDW warming and the sea-ice albedo feedback are insufficient to trigger an irreversible sea-ice tipping point. To investigate further, we apply a theory for winter ocean stability to estimate the warming required to prevent winter sea-ice formation in the marginal ice zone, providing new constraints on the sensitivity of Antarctic sea-ice to ocean warming.

While it is well-established that Zonal Wave Three (ZW3) can drive regional expansion or contraction of Antarctic sea ice, it is unclear whether it has an effect on circumpolar total sea ice area (SIA). Here, we provide evidence to challenge previous studies’ attribution of large overall Antarctic SIA decreases to ZW3. We show that the circumpolar-mean relationship between ZW3 and SIA tendencies is weak due to mostly offsetting meridional wind anomalies and associated opposing sea ice responses. This is despite moderately-strong, positive and significant local meridional wind-sea ice relationships prevailing along the ice-edge. Furthermore, relationships between circumpolar-mean meridional winds and total SIA become weak and can even reverse in some months. An idealised stochastic model reproduces this behaviour under specific conditions where local meridional wind anomalies do not completely cancel out and local SIA responses to wind-forcing vary in strength.

Landfast ice serves as a seasonal extension of the Antarctic coastline and is critical to operations at all Australian Antarctic research stations. The processes preceding and inducing rapid landfast ice retreat during the mid-season, breakout, and end of season, break-up, of the landfast ice remain poorly understood and pose a sizable safety concern to occupants of Antarctic research stations. A new wave hindcast dataset from ECMWF accurately represents wave propagation through the surrounding Antarctic pack ice, allowing estimation of the wave energy reaching the landfast ice edge near the time of these rapid retreat events. Combined with the circumpolar landfast ice dataset of Fraser et al. (2020), this enables the identification of consistent patterns linking increased wave energy at the landfast ice edge to subsequent landfast ice break-up near Australian Antarctic research stations. These results support the hypothesis that wave energy is a primary driver of landfast ice break-up. The findings provide a foundation for potentially developing an early warning system for future landfast ice break-up events around Casey, Davis, and Mawson stations.

The last decade has been marked by a consistently low Antarctic sea ice cover. The most remarkable individual feature of the generally low Antarctic state was the low winter sea ice cover in 2023. This was the first ever winter extreme sea ice event record in Antarctica and was 5-7 standard deviations below the mean (dependent on baseline climatology). Whilst multiple studies have explored the drivers of the last decade’s low state, few studies have explored the drivers of this specific extreme event. By applying a newly developed analysis technique called “Stability at Freezing Temperature (SaFT)” to an ocean reanalysis, I am able to quantify the relative importance of atmosphere and ocean drivers. This analysis demonstrates that the dominant driver of the 2023 winter extreme was penetration of a subsurface warm anomaly into the winter mixed layer. It remains unresolved whether this event was a natural extreme anomaly, or a response to anthropogenic ocean change,

Turbulent ocean mixing plays a key role in regulating heat, salt and freshwater distributions in polar oceans, influencing sea-ice formation and melt, ice-shelf-ocean interactions, ocean circulation, and Antarctic ecosystems. Yet direct observations of mixing south of 60°S remain scarce, resulting in large uncertainties in our understanding of Southern Ocean mixing processes and their representation in ocean and climate models.

Here, we present microstructure turbulence observations collected over three consecutive austral summers (2023-2025) in the King Haakon VII Sea. Our observations show enhanced subsurface mixing over the Antarctic continental slope associated with tidal forcing, with dissipation rates an order of magnitude higher than typical open-ocean values below the mixed layer.

We investigate the potential influence of the enhanced mixing on sea-ice growth and melt by combining turbulence and temperature observations to estimate vertical heat fluxes, which quantify the upward transport of subsurface heat toward the surface ocean.
We also pair the mixing observations with δ¹⁸O measurements to quantify freshwater fluxes and assess how freshwater from sea-ice melt and meteoric sources is redistributed by ocean mixing.

These findings demonstrate the value of direct microstructure observations for constraining Southern Ocean mixing processes and highlight the need for improved parameterizations of turbulent mixing in models.

Microscale processes within sea ice govern the ocean–atmosphere heat flux in regions of critical importance for Earth’s climate. These processes are represented in numerical sea-ice models via an effective thermal conductivity, which is highly uncertain due to the practical challenges associated with the field observations on which it is based. Using a novel laboratory apparatus, here we grow sea ice in controlled conditions and measure its internal temperature with unprecedented resolution and precision to develop a corrected expression for sea ice thermal conductivity. The thermal conductivity is shown to increase with sea ice temperature, which is in direct contrast to existing models which predict a decrease. We show that the corrected thermal conductivity will lead to increased sea ice thickness at warm temperatures and thereby help to resolve persistent biases in Antarctic sea ice that have plagued successive iterations of climate models.