A few feet of thinning could cause glaciers to lift off and break apart
In late 2022, a glacier on the eastern Antarctic Peninsula retreated five miles in two months — a pace never before recorded in the modern scientific era. Hektoria Glacier did not fail because of unusually warm seas or dramatic surface melt, but because the flat seafloor beneath it quietly offered the ice a lift, allowing buoyancy and gravity to do what no storm or heat wave needed to. The event is less a singular catastrophe than a clarifying lesson: the architecture of the earth beneath the ice may matter as much as the warmth of the water around it, and that architecture exists elsewhere.
- Hektoria Glacier collapsed at half a mile per day in late 2022, a rate so extreme it forced scientists to reconsider the basic mechanics of how glaciers fail.
- The trigger was not a heat spike or a melt surge — it was the disappearance of seasonal sea ice that had been quietly holding the glacier's front in place.
- Once unmoored, the glacier lifted off its flat seafloor, toppled forward in massive slabs, and generated six glacial earthquakes as icebergs capsized in rapid succession.
- Satellites recorded a sixfold jump in flow speed and thinning of roughly 80 meters per year — changes invisible to ground observers but unmistakable from orbit.
- The same flat-bed geometry lurks beneath other major Antarctic outlets, and researchers warn that similar sudden collapses could compress sea-level rise timelines by decades.
In late 2022, Hektoria Glacier on the eastern Antarctic Peninsula retreated five miles in two months, peaking at roughly half a mile per day — the fastest known modern retreat of any grounded Antarctic glacier on record. Naomi Ochwat of the University of Colorado Boulder led the investigation, and when she flew over the site nearly two years later, she was struck by the sheer scale of what had disappeared. Without continuous satellite coverage, she noted, the loss of two and a half kilometers in two days might never have been detected at all.
The mechanism was deceptively simple. Beneath Hektoria lay an ice plain — a stretch of flat seafloor where the glacier's weight barely pressed against the rock below. Once the ice thinned enough, buoyancy took over. The glacier lifted off the bed and began to float, sending massive ice slabs toppling forward in a process called buoyancy-driven calving. Six glacial earthquakes marked the largest breakups. Crucially, none of this required warm ocean water or surface melt — the only trigger was the loss of seasonal fast ice that had been dampening waves and steadying the glacier's front.
The deeper concern is what Hektoria reveals about the continent's broader vulnerability. Its flat-bed geometry is not rare; ice plains exist beneath several major Antarctic outlet glaciers. Paleoclimate records suggest retreat rates on such beds can reach hundreds of meters per day under the right conditions. Senior researcher Ted Scambos warned that if similar conditions develop elsewhere, sea-level rise from Antarctica could accelerate far beyond current projections — potentially shifting timelines by decades.
Most climate models treat glacier retreat as gradual and predictable, and do not account for sudden flotation events that can trigger tenfold surges in ice loss. Scientists are now mapping the continent for other glaciers resting on flat beds, treating these zones as early-warning areas. The study, published in Nature Geoscience, calls for glacier models to be updated — not as a dramatic revision, but as a straightforward correction to account for what happens when the ground beneath the ice decides to let go.
In late 2022, a glacier on the eastern Antarctic Peninsula did something scientists had never quite seen before. Hektoria Glacier retreated five miles in two months—not gradually, not over a season, but in a sustained burst of collapse that peaked at roughly half a mile per day. Researchers using satellite data would later confirm it was the fastest known modern retreat of any grounded Antarctic glacier on record.
The speed was so extreme that it forced a reckoning with how glaciers actually fail. Naomi Ochwat, a postdoctoral researcher at the University of Colorado Boulder's Cooperative Institute for Research in Environmental Sciences, led the investigation. When she flew over the glacier in early 2024, nearly two years after the collapse, she was struck by the sheer scale of what had vanished. "If we only had one image every three months, we might not be able to tell you that the glacier lost two and a half kilometers in two days," she said. The satellites had captured something that ground-based observation alone could never have revealed.
The culprit was deceptively simple: a flat seabed. Beneath Hektoria lay what glaciologists call an ice plain—a stretch of seafloor at or below sea level where the glacier's weight rested only lightly on the rock below. This geometry created a precarious equilibrium. Once the ice thinned enough, gravity and buoyancy took over. The glacier lifted off the seafloor and began to float, triggering what researchers call buoyancy-driven calving. Massive blocks of ice toppled forward and shattered. Six glacial earthquakes—the seismic signature of capsizing icebergs—coincided with the largest breakups, a pattern that matched what scientists had observed at other fast-calving fronts around the world.
What made this collapse particularly striking was what it did not require. There was no unusual warming of the ocean water at the time. There was no surge of meltwater from the surface. The trigger was the removal of local fast ice—a seasonal plate of frozen ocean that had been damping waves and holding loose bergs in place. Once that plate vanished, the glacier's front destabilized. Laser altimetry from space showed the remaining ice thinning at roughly 80 meters per year. Satellites tracked a sixfold jump in flow speed as the front gave way.
The broader significance lies in what Hektoria reveals about Antarctica's vulnerability. The glacier itself is not enormous by Antarctic standards, but its underlying geometry is not unique. Ice plains exist beneath several major outlet glaciers across the continent. Paleoclimate records suggest that when grounding lines sit on very flat beds, retreat rates can reach between 55 and 610 meters per day—far faster than most modern observations. Ted Scambos, a senior research scientist at CIRES and coauthor of the study, put the risk plainly: "If the same conditions set up in some of the other areas, it could greatly speed up sea level rise from the continent."
The implications for sea-level projections are substantial. Most climate models treat glacier retreat as a steady, predictable process. They do not account for the possibility that a few feet of thinning could suddenly trigger a tenfold acceleration in ice loss. If similar flat-bed geometry exists beneath larger glaciers in West Antarctica—regions already teetering near instability—short bursts of rapid collapse could add to sea level far faster than current models predict, potentially shifting timelines by decades.
Scientists are now mapping the continent to locate other glaciers sitting on similar flat beds, using radar, seismic data, and satellite altimetry to find where ice plains might be hiding. These zones are early-warning areas. The research also demands that glacier models be updated to include sudden flotation events, the forward toppling of thick ice slabs, and the short-lived surges in motion that follow. It is not a sensational revision—it is a straightforward correction to how glaciers collapse when the bed beneath them offers them a lift. The study was published in Nature Geoscience.
Notable Quotes
When we flew over Hektoria in early 2024, I couldn't believe the vastness of the area that had collapsed— Naomi Ochwat, postdoctoral researcher at CIRES
If the same conditions set up in some of the other areas, it could greatly speed up sea level rise from the continent— Ted Scambos, CIRES senior research scientist
The Hearth Conversation Another angle on the story
Why does a flat seabed matter so much? Couldn't a glacier collapse on any kind of bottom?
A flat bed is the problem precisely because it's flat. On a sloped bed, a glacier can ground itself at different depths as it retreats. But on a flat plain, once the ice thins enough to float, there's nothing holding it back. It's like the difference between a ball rolling down a hill and a ball floating in water.
So Hektoria didn't need warm water or surface melt to collapse this fast?
No. That's what surprised the researchers. The collapse happened because the seasonal sea ice that had been bracing the glacier front melted away. Once that support vanished, gravity and buoyancy did the rest. The ocean temperature was ordinary.
How did they know the ice was actually grounded when it broke apart, not already floating?
Seismic stations recorded forceful signals during the fracturing. Those signals only happen when ice is grounded and breaking under stress. If it were already floating, the physics would be different.
What worries you most about finding similar setups elsewhere?
The speed multiplier. If a larger glacier—say, one feeding into the Ross Ice Shelf—sits on a flat bed and thins past the flotation point, it could lose grounded ice in weeks instead of years. That ice directly raises sea level. Models aren't built to handle that kind of acceleration.
Are there other flat beds under Antarctic ice right now?
Almost certainly. Paleoclimate records show they're common. The challenge is finding them. They're buried under hundreds of meters of ice. That's why scientists are now using radar and satellite data to map where they likely exist.
Could this happen to a glacier in Greenland the same way?
The geometry would have to be similar—a flat bed, thinning ice, loss of buttressing. It's possible, but Greenland's glaciers tend to sit on different kinds of bedrock. Antarctica's ice plains are a particular vulnerability of that continent.