Rob Larter checks the satellite feed the way a doctor checks a pulse: not expecting improvement, but needing to know the number. The marine geophysicist at the British Antarctic Survey has been watching fractures accumulate in the Thwaites Eastern Ice Shelf throughout 2025 and into early 2026 — fractures that thread outward from the pinning point, the submarine ridge on the ocean floor that has held the shelf's floating face in place for decades. He has watched similar shelves fail before. Which is why, in the months before anyone else filed a single headline, he sat down and wrote the death notice.
Breaking Point
The Satellite Image, January 2026
The numbers from January 2026 are stark by any measure. The Thwaites Eastern Ice Shelf is now moving at just over 2,000 metres per year. That is three times its speed in January 2020. And the acceleration has continued. In the five months since that measurement was taken, the shelf has picked up pace again.
"Essentially in free fall now," is how one researcher involved in the monitoring described it. That phrase does not mean the ice is dropping off a cliff. It means the mechanisms that normally slow glacial movement — friction with bedrock, the buttressing weight of the floating shelf, the grip of the pinning point — are one by one giving way. The shelf is no longer behaving like ice. It's behaving like a dam that's losing its footing.
The fractures visible in the latest satellite imagery are concentrated where the shelf presses against that submarine ridge. Huge gashes, in the description of researchers who've been watching the data for years. They weren't there in 2020. Some appear to have opened over the course of a single Antarctic winter.
What makes this moment different from every previous Thwaites alarm is the pre-written obituary. Larter and his colleagues at the British Antarctic Survey have drafted a press release, ready to issue the moment the shelf detaches. It's titled for the death of a geological feature. "Its final demise could happen suddenly," Larter said, "and to avoid being caught on the hop, we have already prepared an 'obituary' press release."
Scientists do not do this casually. A pre-written press release is the kind of preparation you make when you believe something could happen between one morning's satellite pass and the next. Not in decades. Not in years. Any day.
To understand what that means for the rest of us, you need to understand what an ice shelf actually does — and what happens to the glacier behind it when it's gone.
The Buttress
The Thing That Was Holding Everything Back
An ice shelf is the floating extension of a glacier. Where the main body of the ice sheet sits on land, the ice shelf extends beyond the coastline, floating on the ocean but still attached at the grounding line — the submerged boundary where ice leaves bedrock and begins to float. The shelf is thinner and lower than the grounded ice behind it, and on its own it contributes nothing directly to sea level rise. It's already floating; its displacement is already accounted for in ocean volume.
What it does, instead, is act as a brace. A glacier flowing toward the ocean is slowed by the resistance of its own floating tongue. The ice shelf pushes back against the flow of the grounded ice inland. Remove the shelf, and the grounded ice accelerates. More ice per year reaches the sea. Sea level rises.
Marine Ice Sheet Instability — Why Thwaites Is Uniquely Dangerous
Most glaciers sit on a bed that rises as you go inland. That means if they start retreating, the grounding line moves to shallower water, where the ice is thinner and less vulnerable to ocean melt. The retreat naturally slows. It's self-limiting.
Thwaites is the opposite. Its bedrock deepens as you move inland — what glaciologists call a retrograde bed. If Thwaites' grounding line begins retreating, it encounters deeper water. Deeper water means more exposure to the warm ocean. More melt. Faster retreat. Which exposes deeper water again. The process is self-reinforcing, not self-limiting.
This feedback loop is called Marine Ice Sheet Instability, and it's why Thwaites is not just a large glacier at risk — it's a large glacier where retreat, once it passes a critical threshold, may be physically impossible to stop by any natural mechanism.
If You Were Ten
Imagine you're trying to push a heavy box off a ramp. Most ramps go up as you push, which makes it harder and harder to keep moving — so the box eventually stops. Thwaites sits on a ramp that goes down in the direction it's moving. The further it goes, the easier it is for it to keep going. There's no natural stopping point built into the slope.
Thwaites' grounding line currently sits between 800 and 1,200 metres below sea level. The retrograde bed behind it descends further still. In 2024, a paper published in PNAS confirmed something researchers had long feared: seawater is intruding beneath the grounded ice at tidal frequencies, extending six kilometres or more inland of the grounding line. The ocean is not just touching the front of the glacier. It's reaching underneath it.
"Projections of sea level rise might be significant underestimates," said Alex Bradley, an ice dynamics researcher at the British Antarctic Survey and lead author of the study. The mechanisms that drive melt at the grounding zone are more aggressive than most models have assumed. The ice-ocean interface is more permeable than anyone expected.
"Its final demise could happen suddenly, and to avoid being caught on the hop, we have already prepared an ‘obituary’ press release."Rob Larter — Marine Geophysicist, British Antarctic Survey, 2026
The warm water doing this work is Circumpolar Deep Water: a layer of relatively warm Atlantic-derived water that upwells onto the West Antarctic continental shelf and finds its way toward the base of the glacier. It doesn't need to be dramatically hot to cause dramatic damage. Water just a degree or two above freezing point melts ice at rates of 200 metres per year at the most vulnerable sections of the grounding zone — rates that have been directly measured, not modelled.
The Eastern Ice Shelf's pinning point is a submarine ridge on the ocean floor. The shelf's floating face rests on this ridge the way a door rests in its frame. Without it, the floating shelf has nothing to brace against. The fractures opening around that point in 2025 and 2026 are the physical record of that bracing force being overwhelmed.
Eighty Years
What the Glacier Has Already Given Up
The Thwaites retreat didn't begin last year. It began in the 1940s. That knowledge comes from sediment cores and seafloor mapping — physical records of where the grounding line once sat — and from the work of researchers who pieced together the pre-satellite history of West Antarctica's ice. Significant thinning and retreat was already underway before a single Earth-observing satellite had been launched.
What the satellite era revealed, beginning in the early 1990s, was how fast that retreat had become. Between 1992 and 1996 alone, the grounding line retreated 1.4 kilometres. The main trunk of the glacier accelerated by 33 per cent between 1973 and 1996. Then it accelerated by another 33 per cent between 2006 and 2013. Ice discharge was increasing not just in absolute terms but in its rate of increase — the acceleration was itself accelerating.
FIGURE 1 — THWAITES GLACIER RETREAT TIMELINE, 1940s–2026
By 2017, satellite data showed sectors of Thwaites retreating at 0.8 kilometres per year, with basal melt rates of 200 metres per year at the fastest points. Other sectors were melting at a tenth that speed — the glacier is not uniform, which is part of what makes modelling its future so difficult. Some parts are racing. Others are walking. The racing parts are the ones that now have fractures running through them.
The International Thwaites Glacier Collaboration, a joint US-UK research programme launched in 2018, has been the most sustained scientific effort to understand what the glacier is actually doing at the grounding line. Researchers have drilled through more than a kilometre of ice to instrument the ocean cavity beneath the shelf. They've sent autonomous submarines into spaces no human will ever reach. In early 2026, a PBS film crew documented researchers drilling 1,000 metres down to reach the base of the glacier — a project years in planning, driven by the recognition that what happens in those few hundred metres of dark water beneath the ice determines the future of every coastal city on Earth.
| Ice Body | Sea Level Potential | Current SLR Contribution | Status |
|---|---|---|---|
| Thwaites Glacier | ~65 cm | 4% of global SLR | ACCELERATING |
| Pine Island Glacier | ~50 cm | ~3% of global SLR | ACCELERATING |
| Amundsen Sea sector (full) | ~1.2 m | 8% of global SLR | AT RISK |
| West Antarctic Ice Sheet (cascade) | 3.3–5 m | Indirect | TRIGGERED BY THWAITES |
| Greenland Ice Sheet | ~7 m | ~1 mm/yr | ACCELERATING |
| Global sea level (current rate) | — | 4.5 mm/yr total | ONGOING |
TABLE 1 — ICE BODIES, SEA LEVEL POTENTIAL, AND CURRENT STATUS. Sources: NSIDC, BAS, Copernicus Climate Change Service.
The ITGC data has refined, not softened, the picture. The grounding line is retreating. The basal melt rates are at the high end of model projections. The seawater intrusion beneath the grounded ice is more extensive than previous surveys had detected. Everywhere researchers have looked more carefully, things have been worse than they expected.
After the Shelf
The Proposals at the Edge of the World
When the Eastern Ice Shelf goes, the grounded glacier behind it will accelerate. How much, and how fast, depends on variables that even the best models can only bracket. But the direction is not in dispute. The question that matters at the policy level is not whether sea levels will rise, but how many metres, and whether it happens over decades or centuries. The difference between those timescales is the difference between managed coastal retreat and catastrophic displacement.
Globally, 250 million people live within one metre of high tide lines. A three-metre rise — the threshold that the full West Antarctic cascade could deliver — submerges substantial portions of Miami, London, Shanghai, New York, Tokyo, Mumbai, and dozens of smaller coastal cities. Bangladesh alone has 17 million people living within a metre of sea level. These aren't numbers drawn from worst-case scenarios. They come from elevation surveys of where people currently live.
The timing is the honest uncertainty. Thwaites itself holds roughly 65 centimetres of sea level equivalent. That rise doesn't happen overnight when the ice shelf fails — the ice is still on land, and it has to flow to the coast and calve into the ocean. The process takes decades to centuries. But the ice shelf detachment is the trigger. It's what changes Thwaites from a slow problem to an irreversible one.
The Proposals: Engineering the Glacier
The Curtain / 80km Wall: Scientists have proposed constructing a physical underwater curtain or berm stretching up to 80 kilometres across the Amundsen Sea, designed to block the warm Circumpolar Deep Water from reaching the glacier's base. The engineering challenge is without precedent: building in 500+ metres of water, in one of the most remote and hostile oceans on Earth, in conditions that prevent year-round access.
The Bubble Tube: A separate Cambridge proposal involves installing a tube along the seafloor to release billions of fine bubbles, creating an upwelling of cold surface water to displace the warm deep layer flowing toward the glacier. Less structurally demanding than a wall, but the scale of intervention required to affect a continental ice sheet remains untested.
Current status: In January 2026, the South Korean research vessel RV Araon surveyed a submarine ridge in the Amundsen Sea to assess whether it could serve as a foundation for proposed infrastructure. No intervention is funded. None is close to implementation. The survey represents the first serious engineering reconnaissance of the area.
The honest assessment: Every proposal faces the same core problem. The scale of energy required to meaningfully alter ocean circulation patterns around a glacier the size of Britain, using technology that doesn't yet exist, at a location accessible for only a few weeks each year, is staggering. These are not dismissed ideas — they are serious scientific proposals in early development. But none will be ready before the shelf detaches.
What comes after the ice shelf's detachment is a period of accelerated glacier flow and an unstable grounding line — the kind of retrograde-bed feedback that Marine Ice Sheet Instability theory predicts will be self-sustaining. Whether the grounding line stabilises at a new position or continues retreating into deeper water is the critical unknown. The models disagree, not because the physics is uncertain but because the initial conditions are imprecisely known. The shape of the bedrock kilometers inland of the current grounding line has been mapped, but not with the resolution needed to determine exactly where the bed shallows enough to arrest retreat.
What we can say is this: the window to intervene at the glacier itself was decades ago. The window to prepare for the consequences is narrowing. Coastal infrastructure planning, managed retreat from low-lying zones, updated flood modelling, and revised building codes are not hypothetical responses to a distant risk. They're responses to a process that is already underway and that a satellite image from January 2026 confirms has reached a new phase.
Larter's pre-written obituary sits in a folder at the British Antarctic Survey. It will say that the Thwaites Eastern Ice Shelf, which had held its position for decades, has detached from its pinning point. It will note what that means for the glacier behind it. It will be filed when the satellites confirm what the fractures already suggest.
He checks the feed before breakfast. The number this morning is the same as yesterday's. Still over 2,000 metres per year. Still accelerating.
Still waiting.
Buy me a coffee