In just 15 months, Antarctica lost a stretch of glacier longer than the island of Manhattan. Between January 2022 and March 2023, Hektoria Glacier on the Antarctic Peninsula shed roughly 25 kilometers (about 15 miles) of grounded ice, according to a peer-reviewed study published in Nature Geoscience in 2025. The retreat was nearly an order of magnitude faster than any grounded-glacier collapse previously recorded in the satellite era, the study’s authors reported. As of June 2026, no comparable event has been documented anywhere else on the continent.
A glacier primed to fail
Hektoria is one of several tributary glaciers that once fed into the Larsen B ice shelf on the eastern side of the Antarctic Peninsula. When Larsen B spectacularly disintegrated in 2002, its tributaries lost the massive floating buttress that had held them in check. Research published in the years that followed showed Hektoria, along with neighboring Green and Evans glaciers, thinning steadily as relatively warm ocean water reached their undersides and melted them from below.
By the early 2020s, that slow erosion had pushed Hektoria’s “ice plain,” the gently sloping zone near where the glacier meets the seafloor, dangerously close to buoyancy. The glacier was, in effect, a building with its foundations half eaten away. It was still standing, but only because one structural element remained: a band of landfast sea ice filling the Larsen B embayment, physically bracing the glacier’s floating tongue against wave action and warm-water intrusion.
In January 2022, that band broke apart in a matter of days. The last restraint was gone.
The fastest grounded retreat on record
What followed was swift and dramatic. Successive Landsat satellite scenes analyzed by the study’s lead author, Bertie Wallis of the University of Leeds, and colleagues show the glacier’s terminus pulling back in a rapid sequence of calving events. The most intense phase came in November and December 2022, when the ice front retreated 8.2 ± 0.2 km in just two months.
A separate NASA Earth Observatory analysis independently confirmed the scale and timing, describing the episode as the highest rate of grounded glacial ice retreat observed in modern satellite records. The NASA chronology shows a relatively stable terminus before the sea-ice breakup, then a cascade of fractures as the newly exposed ice plain, no longer pressed against the seafloor, floated free and shattered.
The mechanism the researchers identified is distinct from the surface-melt-driven retreat that dominates most glacier loss worldwide. Here, years of ocean-driven basal melting thinned the ice plain until it reached flotation. Once buoyant, the ice became vulnerable to ocean forces: tides, currents, and swells that pried apart weakened ice far faster than surface warming alone could have managed. The sea-ice breakup did not cause the thinning. It simply removed the lid.
What scientists still don’t know
For all its clarity, the satellite record leaves significant gaps. No instruments were measuring ocean temperature or current speed at Hektoria’s grounding line during the collapse. Every estimate of how much thermal energy the ocean delivered to the glacier’s base relies on models or data from moorings positioned far from the calving zone. Small errors in assumed water temperature can translate into large uncertainties in melt rates, which means the precise intensity of the warm-water forcing remains an inference.
The transition from grounded ice to a floating, calving-prone state was tracked entirely from orbit. No seismic sensors, GPS stations, or ice-penetrating radar captured the mechanical sequence as the ice plain lifted off the bed. Satellites can show that the terminus retreated and estimate surface elevation changes, but they cannot directly observe the moment basal contact with bedrock was lost. That distinction matters for building models that predict where the same process might strike next, because the exact threshold at which ice begins to float depends on local thickness, density, and the shape of the seafloor beneath it.
There is also a data gap in the middle of the story. Detailed mass-loss records for Hektoria cover 2001 through 2009, when the glacier was already responding to the loss of Larsen B. Updated measurements for the specific period between 2009 and 2022 are not available in the primary sources. The causal logic connecting slow thinning in the 2000s to the abrupt 2022 collapse is strong, but the intermediate data points are sparse, leaving room for debate about whether the pre-collapse thinning was steady or came in bursts.
Why it matters beyond the Antarctic Peninsula
Hektoria is a relatively small glacier. Its 25 km retreat, while record-setting, does not by itself reshape global sea-level projections. The reason glaciologists are paying close attention is what the event reveals about process and speed.
Several far larger glacier systems draining West Antarctica, most notably Thwaites and Pine Island, also sit on ice plains that are being thinned from below by warming ocean water. Those glaciers hold enough ice to raise global sea levels by meters if they were to fully destabilize over centuries. The standard concern has been that their retreat, while potentially irreversible, would unfold gradually. Hektoria challenges that assumption. It shows that once an ice plain is thinned close to flotation and its external buttressing disappears, the transition from slow erosion to rapid collapse can happen in a single season.
That does not mean Thwaites or Pine Island will behave the same way. Each glacier sits in a different ocean environment, with different bed geometry and different sea-ice dynamics. Some rest in overdeepened basins that could accelerate retreat; others are braced by underwater ridges or colder shelf waters that may slow it. Applying Hektoria’s timeline directly to those systems would be a stretch the data cannot support.
But the event does narrow the range of what scientists consider physically possible. Before Hektoria, no observed example existed of a grounded glacier retreating this far, this fast. Now one does. And the conditions that produced it, decades of subsurface warming followed by the abrupt loss of protective sea ice, are not unique to one corner of the Antarctic Peninsula.
What happens at Hektoria next
Whether the glacier stabilizes at its new grounding line or continues to pull back remains an open question. The retreat removed a large area of grounded ice, but the bed beneath the current terminus is still relatively shallow and flat, a geometry that could leave the glacier vulnerable to further flotation if basal melting continues.
Future behavior will depend on a tug-of-war between forces: snowfall-driven thickening in the glacier’s interior pushing ice seaward, ongoing warm-water erosion at the grounding zone pulling it back, and the seasonal return (or absence) of landfast sea ice in the embayment providing or withholding a buffer. Without a longer post-2023 observational record, researchers cannot yet say whether the 25 km retreat was a discrete step change or the opening act of a more extended unraveling.
What is clear is that the event has rewritten the benchmark for how quickly grounded ice can disappear. For glaciologists monitoring vulnerable ice plains across Antarctica, Hektoria is no longer a theoretical scenario. It is measured proof that when the right conditions align, decades of slow decay can become months of dramatic collapse.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
