When Peter Convey first stepped onto Signy Island in the South Orkney archipelago in 1989, a small rock was visible through the McCloud Glacier’s surface. It was named Manhaul Rock, though it hardly warranted one. You could easily approach it, ski by it, or drive a skidoo across the ice next to it. By 2024, Manhaul Rock had transformed into a full nunatak, a jagged mass rising above the scant remnants of the glacier surrounding it. The ice that once concealed it had simply disappeared.
Convey, a biologist at the British Antarctic Survey, has visited the Antarctic Peninsula for 35 years. “For a casual visitor, the initial impression remains that the area is predominantly ice-covered,” he states. “However, for those of us fortunate enough to return multiple times, there are evident changes over time.”
These changes are now the focus of an extensive new analysis, released today in Frontiers in Environmental Science, that seeks to accomplish something both simple and alarming: modeling the potential futures of the Antarctic Peninsula under three different scenarios. A group of 24 researchers from institutions across the UK, the US, Germany, and other locations analyzed a variety of factors, including atmospheric temperatures and sea ice, as well as penguin colonies and moss beds, evaluating data for low, medium-high, and very high emission scenarios. The disparity between the best and worst outcomes is, in fact, vast. Moreover, the timeframe for deciding which trajectory we take is rapidly shrinking.
“The Antarctic Peninsula is a unique location,” says Bethan Davies of Newcastle University, who spearheaded the study. “Its future is contingent upon the decisions we make today.”
This may sound like typical climate discourse, but it’s not. The Peninsula is warming significantly faster than the global average, with Vernadsky Station on the west coast recording about 0.45°C of warming per decade since 1951; that equates to more than 3°C over 73 years. The all-time highest temperature recorded on the Antarctic mainland, 18.6°C, occurred near the Peninsula’s northern tip in February 2020. That same summer, George VI Ice Shelf faced a 32-year record melt event. Two years later, another record surface melt occurred in the Peninsula during an extreme heat episode caused by an atmospheric river, a long, narrow channel of moisture that transports heat from the subtropics to the poles.
These events are no longer isolated incidents. The years from 2022 to 2024 recorded the three lowest extents of Antarctic sea ice in the satellite era. Atmospheric rivers impacting the Peninsula have been increasing at approximately 0.89 per decade since 1979. Marine heatwaves are becoming more frequent and intense in the Southern Ocean. In essence, the system is already under significant strain.
What the new study does is forecast that strain into the future. The three scenarios align with global temperature increases of 1.8°C, 3.6°C, and 4.4°C above pre-industrial levels by the century’s end. Under the lowest scenario, the Peninsula warms by about 2.3°C compared to pre-industrial temperatures. Under the highest scenario, that figure escalates to around 6.1°C. The ramifications ripple through every aspect of the system in ways that become nearly irreversible under high emissions.
Consider sea ice. Under low emissions, seasonal reductions remain mild, between 1 and 2 percent. In the very high scenario, winter sea ice around the Peninsula decreases by nearly 20%, with the Weddell Sea losing over a fifth of its autumn coverage. This is significant because krill, the tiny crustaceans that form the basis of much of the Antarctic food web, rely on winter sea ice for their young to thrive. Lose the ice, lose the krill. Lose the krill, and you begin to lose the whales, seals, and penguins that feed on them. The 2022 record low sea ice in the Bellingshausen Sea has already led to breeding failures among local emperor penguin colonies. With further warming, such catastrophic seasons could become standard.
Then there are the ice shelves, those immense floating platforms that support the glaciers behind them. When Larsen B broke apart in 2002, the Hektoria Glacier behind it receded more than 16 kilometers in less than two decades. Under very high emissions, the study suggests that both Larsen C and the Wilkins ice shelves are likely to collapse by 2100. Surface melt would saturate the porous firn layer, forming meltwater ponds that penetrate the ice through hydrofracture. This represents a type of gradual structural failure. George VI Ice Shelf, although already experiencing extraordinarily high surface melt rates, might endure longer due to a compressive flow regime that enhances its resistance to fracturing, but even that resilience has its limits. Should it collapse