CAMBRIDGE, Massachusetts — Antarctica's tallest ice cliffs are collapsing faster than scientists predicted — but then they stop. A new satellite study of more than 15,000 ice cliffs, published in the journal Science on 15 October 2026, finds that the feared runaway collapse of the Antarctic ice sheet, in which one toppling cliff exposes an ever-taller cliff behind it, does not appear to happen in practice.
Illustration: A towering Antarctic ice cliff above dark ocean water with floating ice fragments, representing the ice cliffs studied by satellite.
What Happened
Researchers from the Massachusetts Institute of Technology (MIT) and the Woods Hole Oceanographic Institution (WHOI), working with the University of Michigan, analysed satellite observations of ice cliffs around Antarctica over four years. Their paper, "Empirical constraints on ice cliff calving", was published in Science on 15 October 2026. The team was led by Caroline Needell, a doctoral researcher in the MIT-WHOI Joint Program.
The central finding is double-edged. When ice cliffs did collapse, they retreated faster than existing models predicted — meaning more ice loss, and a higher contribution to sea-level rise, than the models assumed for those events. But the retreat was short-lived: the glaciers behind the cliffs settled into shorter, more stable shapes instead of unravelling inland. "We found that when the cliffs were retreating, they were doing so faster than what has been modeled, which would mean more ice loss and higher contributions of sea level rise," Needell said, according to MIT News. "But because we found that the retreat was short-lived, and all the glaciers eventually stabilized, that would suggest that things won't be so bad. We don't see this runaway effect."
The study also measured how tall Antarctic cliffs actually are. Most averaged around 30 metres — roughly 100 feet — well below the roughly 100-metre threshold that models have used to trigger marine ice cliff instability, the process in which a cliff taller than about 100 metres collapses under its own weight, exposing a taller cliff inland, which collapses in turn. Only about 5% of Antarctic cliffs exceeded about 52 metres, and a handful rose above 100 metres, with one reaching 127 metres. The tallest cliffs were concentrated where glaciers flow quickly toward the ocean.
Background
The fear of runaway ice-cliff collapse has shaped some of the highest sea-level projections for this century. The logic is straightforward: ice shelves — the floating extensions of glaciers — act as buttresses, holding back the ice behind them. As shelves weaken and break apart in a warming climate, they expose sheer cliffs of solid ice at the edge of the ice sheet. If those cliffs are too tall to support their own weight, they should topple, and the new cliff face behind them should be taller still, setting off a self-sustaining chain reaction that could drive rapid sea-level rise and threaten coastal cities.
Until now, that chain reaction was largely a product of physics calculations about how tall a cliff can stand, rather than observations of real cliffs over time. The new study instead tracked how each cliff and its parent glacier actually changed, using measurements from NASA's Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2), a laser altimeter that measures changes in ice elevation, with observations taken about every three months between 2019 and 2023. The researchers also drew on airborne and satellite laser data spanning 2003 to 2023, and examined Greenland as well as Antarctica.
What they found complicates the simple height rule. A cliff's stability depended on more than its height: the depth of the water in front of it, the thickness of the icy slush known as mélange in front of the glacier, and cracking near the glacier's edge all played a part. And where the land beneath the ice happened to rise, the topography acted as a literal speed bump, thinning the overlying ice and making further collapse less likely. As tall cliffs collapsed, they settled into shorter, more stable cliffs — which, the researchers argue, is precisely what prevents the domino effect from running away.
Illustration: A satellite with laser beams measuring Antarctic ice elevation from orbit, representing the ICESat-2 observations used in the study.
Why It Matters
Sea-level projections are not academic abstractions: they determine where coastal cities build seawalls, how insurers price flood risk, and which communities plan for managed retreat. If one of the most feared pathways to extreme sea-level rise this century is less likely than feared, the worst-case end of the projection range becomes less plausible — and planners can weigh that when deciding how much protection to build, and how fast.
But the study is not an all-clear, and its authors do not present it as one. Antarctica is still losing ice, and the cliffs that do collapse are doing so faster than models predicted, which means near-term ice loss from those events may be underestimated even as the long-term runaway scenario looks less likely. The finding refines the forecast; it does not reverse it. Other processes — the thinning of ice shelves from below by warm ocean water, the retreat of glaciers such as Thwaites in West Antarctica, and the steady loss of ice from Greenland — continue regardless of what happens at any single cliff face.
The study also points to what scientists most need to measure next. "Now we know what to look for," Needell said, according to MIT News. "We really need to know the bed topography, and we need to know the ice thinning rate, in order to get ice cliff instability right, because this value really matters in the context of projecting sea level rise." In other words, the next generation of sea-level models will need better maps of the ground beneath the ice, not just better physics for the ice above it.
Illustration: A coastal city skyline beside rising sea water with a protective seawall, representing what sea-level projections mean for coastal planning.
What Happens Next
The findings arrive less than a month before the COP31 climate summit opens in Antalya, Turkey, on 9 November 2026, where negotiators will again debate how fast emissions must fall to limit warming — and how much adaptation funding coastal nations, particularly low-lying island states, will receive. A refined sea-level forecast will feed into those discussions, but it is unlikely to change the central negotiating reality: every fraction of a degree of warming still accelerates ice loss, and the difference between a stabilising cliff and a collapsing one is measured in decades of emissions choices, not in a single study.
For researchers, the next steps are to extend the observational record beyond 2023, to test whether the stabilisation pattern holds as ice shelves continue to weaken, and to fold the new constraints — bed topography, thinning rates, mélange thickness — into the ice-sheet models that produce the projections used by governments. The study was supported in part by NASA, the US National Science Foundation, and the DOMINOS project, part of the International Thwaites Glacier Collaboration. If its conclusions hold up under longer observation, the worst-case sea-level story of this century may need rewriting — not to a happy ending, but to a more precise, and therefore more useful, one.
Sponsored: explore today's featured sponsored offers
FAQ
What is marine ice cliff instability?
It is a hypothesised process in which an ice cliff taller than about 100 metres collapses under its own weight, exposing a taller cliff inland, which collapses in turn — a self-sustaining chain reaction driving rapid sea-level rise.
What did the new study actually find?
Studying more than 15,000 Antarctic ice cliffs, researchers found that collapsing cliffs retreated faster than models predicted, but the retreat was short-lived and glaciers then stabilised — no runaway effect was observed.
How tall are most Antarctic ice cliffs?
Most averaged around 30 metres, well below the roughly 100-metre threshold used in models; only a handful exceeded 100 metres, with one reaching 127 metres.
How was the study done?
The team used NASA's ICESat-2 laser altimeter, with measurements about every three months from 2019 to 2023, plus airborne and satellite laser data from 2003 to 2023.
Does this mean sea-level rise is no longer a threat?
No. Antarctica is still losing ice, collapsing cliffs retreat faster than modelled, and other processes such as warm-water thinning of ice shelves continue; the study refines one worst-case pathway, it does not remove the threat.
Sources
- MIT News — "Satellite data reveal how Antarctica's ice cliffs behave over time" (15 October 2026)
- Needell et al. — "Empirical constraints on ice cliff calving", Science, DOI: 10.1126/science.aee0658 (15 October 2026)
- Nature World News — "Most Antarctic Ice Cliffs Are Too Short to Trigger the Runaway Collapse Behind Worst-Case Sea Level Forecasts, Study Finds" (20 October 2026)
Comments
Post a Comment