Fifteen Miles in Fifteen Months

I. What happened to Hektoria

Hektoria Glacier sits on the eastern edge of the Antarctic Peninsula, on the side facing the Weddell Sea. By Antarctic standards it is small — about 115 square miles, roughly the area of Philadelphia. Between late 2022 and early 2023, the glacier retreated approximately fifteen miles in fifteen months, with eight kilometres of that retreat occurring in just sixty days during November and December 2022. NASA Landsat imagery captured the sequence. A study published in Nature Geoscience in late 2025 reconstructed the mechanism. On 19 May 2026, the NASA Earth Observatory issued an update characterising the event as the fastest modern collapse of a grounded Antarctic glacier ever recorded.

The mechanism is what makes Hektoria significant beyond its own size. Glaciers in Antarctica typically retreat in a slow, predictable manner that climate models can simulate with reasonable accuracy. Hektoria did not. The glacier had been resting on a flat plain of bedrock that lies below sea level. As the ice above thinned over years of warming, the bedrock's shape allowed seawater to push beneath the ice. Once seawater intruded under a large area at roughly the same time, an enormous section of grounded ice — ice resting on the seafloor and thus contributing to sea level when it melts — lifted off the bedrock simultaneously and began floating. Floating ice is structurally vulnerable in ways grounded ice is not. The lifted section began calving away in massive blocks, and the calving accelerated as more ice was exposed to the ocean.

The researchers led by the University of Colorado Boulder and Swansea University describe this as a “buoyancy event” — a transition from grounded to floating ice that, once initiated, can accelerate retreat by an order of magnitude. The conditions that triggered it at Hektoria are not unique to that glacier. They exist beneath several much larger glaciers in West Antarctica, including the Thwaites Glacier — the so-called Doomsday Glacier — whose collapse would commit roughly two feet of global sea-level rise on its own.

In late May 2026, the International Thwaites Glacier Collaboration — the joint U.S.–U.K. research consortium that has spent the past seven years studying the glacier — went public with a warning that the eastern ice shelf in front of Thwaites is now expected to break up within 2026. Robert Larter, the British Antarctic Survey marine geophysicist who co-leads the collaboration's U.K. arm, told Live Science that the shelf's final disintegration is “very likely to happen sometime this year.” BAS has prepared an obituary press release in advance, given that the breakup is expected to happen suddenly. Two specific measurements anchor the warning: the shelf's flow speed increased more than threefold between January 2020 and January 2026, now exceeding 2,000 metres per year, and the glacier ice formerly held back by the shelf has sped up by roughly 33 percent over the same period. The breakup of the floating ice shelf itself will not directly raise sea level — floating ice already displaces its water volume. What it will remove is the buttress that has slowed the flow of Thwaites' grounded ice into the ocean. The grounded ice is what raises sea level when it reaches the sea.

II. What the models said, and what they did not

The standard climate models used by the Intergovernmental Panel on Climate Change to project Antarctic ice loss treat glacier retreat as a relatively gradual process. The models account for thinning, for surface and ocean warming, for shifts in snowfall, and for some calving dynamics. They do not, in general, include buoyancy-driven sudden retreat events at the scale Hektoria has now demonstrated. Most current sea-level-rise projections, including the IPCC's sixth assessment report, assume that significant West Antarctic ice loss unfolds over centuries.

A February 2026 study by the Potsdam Institute for Climate Impact Research and the Max Planck Institute of Geoanthropology, published in Nature Climate Change and led by Ricarda Winkelmann, made a related but distinct point. The Antarctic Ice Sheet, the researchers argued, does not behave as a single tipping element. It behaves as a set of interacting basins, each with its own critical temperature threshold. At present warming — approximately 1.3°C above the pre-industrial average — the researchers found that approximately 40 percent of the ice stored in West Antarctica may already be committed to long-term loss. At 2 to 3°C of warming, additional basins in East Antarctica would cross thresholds beyond which loss becomes irreversible over centuries to millennia. “It's not one single threshold we need to watch in Antarctica,” Winkelmann said. “It's a sequence.”

The two findings, read together, suggest that the rate at which Antarctic ice may be lost is potentially faster than standard projections indicate (Hektoria's mechanism), and that the total amount of ice already committed to loss under current warming is potentially larger than commonly stated (the Winkelmann analysis). Neither finding is final. Both are part of an evolving scientific picture. Both push in the same direction.

III. The connection to where people live

Antarctica contains approximately 90 percent of the world's ice and roughly 70 percent of its fresh water. Complete loss of the West Antarctic Ice Sheet alone would raise global sea level by approximately 3 to 5 metres. Complete loss of all Antarctic ice would raise it by approximately 58 metres. These figures, framed at this scale, can feel abstract. They become concrete at the scale of human exposure.

A half-metre rise in global sea level — within the range of plausible 2050 outcomes under current emissions — would place roughly 800 million people in 570 coastal cities at acute exposure to flooding. At that level, approximately 11 percent of Bangladesh's land area would be lost, displacing 15 million people. Seventy percent of Europe's largest cities would face new coastal risk. Seventy-eight million Chinese in low-elevation coastal zones would be exposed, a population growing by approximately three percent each year. A one-metre rise would expand all of these figures and bring the loss of land currently inhabited by an additional 230 million people. A two-metre rise, which the IPCC's sixth assessment described as “not ruled out” by 2100 under high-emissions and rapid-ice-loss scenarios, would redraw the world's coastlines. The Hektoria mechanism and the Thwaites ice-shelf warning sit inside these projections; they suggest, without proving, that the higher end of the range may need to be taken more seriously than the lower.

The economic figures scale similarly. The UK National Oceanographic Centre has estimated the global annual cost of unmitigated sea-level rise at approximately $14 trillion per year by 2100. A 2025 study by Dutch and Italian researchers, modelling 271 coastal regions, estimated combined United Kingdom and European Union losses at €872 billion by century's end. World Bank economists modelling 136 major coastal cities found that without comprehensive adaptation, annual losses from coastal flooding alone could approach $1 trillion by 2050; with adaptation — dikes, seawalls, drainage upgrades, managed retreat — those losses fall to approximately $60 billion. The gap between the two figures, roughly $940 billion per year, is the approximate annual cost of failing to act. United States shoreline armouring costs alone are forecast at $300 billion by 2100. Approximately 300 American energy facilities currently sit within four feet of high tide.

Different coastlines will rise at different rates. The US East Coast is projected to see 14 to 21 inches (0.36 to 0.54 metres) of rise by 2050; the western Gulf of Mexico, where rise is fastest in the United States, somewhat more than half a metre. The Pacific Northwest, where tectonic plates are pushing land upward, will see slower rise. In Asia, Kolkata, Mumbai and Dhaka are the cities with the most people projected to be at risk from coastal inundation — between eleven and fourteen million each by 2070. Miami, Guangzhou and New York hold the highest exposed asset values: between two and three-and-a-half trillion dollars per city. Pacific Island nations face the most acute proportional exposure: at one metre of rise, Tuvalu, Kiribati, the Marshall Islands and the Maldives confront questions of state survival. Australia and New Zealand carry relatively low population exposure but very high concentrations of high-value coastal property — Brisbane, Sydney's eastern suburbs, Auckland's reclaimed land, the entire low-lying Bay of Plenty. Specific national projects already underway include the 8 June 2026 launch of Tuvalu's National Ocean Policy — the same week this edition goes to press, in a country averaging less than 4.5 metres above sea level — the Half Moon Bay seawall project in California, the Thames Barrier upgrade in London, the Maeslantkering surge barrier in the Netherlands, the relocation of Indonesia's capital from Jakarta, and the Miami-Dade flood-management programme. The list extends.

Hektoria Glacier itself, in terms of its direct contribution to sea level, is modest. A 115-square-mile glacier on the Antarctic Peninsula has nowhere near the volume to substantially shift global sea level on its own. What Hektoria has demonstrated is a mechanism. That mechanism, if it applies to Thwaites, to Pine Island, to the larger glaciers along the Amundsen Sea coast, would change the timeline by which projections about coastal cities, low-lying nations, and global infrastructure need to be revised. The science to determine whether and where the mechanism applies is now underway. Both NASA and ESA satellite programmes have begun targeted mapping of bedrock topography under vulnerable glaciers. Results will be published over the coming years.

Two observations close the picture. First, the standard scientific practice in projection work is conservative; this means that on questions where the data is still emerging, formal projections will lag behind the leading edge of what the field actually expects. The IPCC's next assessment cycle, AR7, is currently in early stages with publication targeted for 2028–2029. By then, the Hektoria mechanism will likely have been studied at multiple sites, and the projections will have updated. Second, the timeline within which decisions about coastal infrastructure are made — whether to build a seawall in this decade or the next, whether to relocate a port or upgrade it, whether a city plans for one metre of rise or two — runs in parallel to that scientific work but does not wait for it. The science is updating. The infrastructure decisions are happening now. They are happening in different cities at different scales, on different timelines, with different costs falling on different populations. The gap between the updating science and the decisions already being made — in dollars, in metres, in displaced lives — is its own form of risk.

Sources

Hektoria collapse: ScienceDaily / NASA Earth Observatory on the 19 May 2026 update characterising Hektoria's 15-mile retreat in 15 months; ScienceDaily on the original Nature Geoscience study describing the buoyancy mechanism; Earth.com on the broader implications for vulnerable glaciers.

Thwaites ice-shelf warning (late May 2026): Live Science interview with Robert Larter (British Antarctic Survey) on the predicted 2026 breakup; The Cool Down on the “obituary” press release prepared by the British Antarctic Survey; GB News on the threefold increase in ice-shelf flow velocity since January 2020.

Tipping-points framework: Potsdam Institute for Climate Impact Research on the Winkelmann et al. study; The study itself in Nature Climate Change; Down to Earth (India) on the findings and their implications.

Regional context: AntarcticGlaciers.org on the broader Antarctic Peninsula context (13 billion tonnes/year ice loss).

Population exposure and human impact: C40 Cities — The Future We Don't Want analysis on 800 million people in 570 cities at risk by 2050; Quanta Magazine on WAIS collapse and the 230 million figure; World Economic Forum Global Risks Report on Bangladesh and Chinese exposure; NRDC on US East Coast and Gulf Coast projections to 2050.

Economic projections: UK National Oceanographic Centre's $14 trillion annual estimate by 2100 via Bloomberg / Zurich Insurance; 2025 Dutch/Italian 271-region study on $3 trillion+ in coastal city losses and the €872 billion UK-EU figure; Climate Central on the World Bank/Hallegatte study and the $1 trillion vs $60 billion adaptation calculus; PNAS on US shoreline armoring costs forecast at $300+ billion by 2100.

Fifteen Miles in Fifteen Months

I. What happened to Hektoria

II. What the models said, and what they did not

III. The connection to where people live

Sources

Further Reading

One honest edition. Every Saturday.