How Marine Ice Cliffs Could Trigger Rapid Sea Level Rise

Antarctica holds enough ice to raise global sea level by around sixty metres, and the single largest uncertainty in every projection of future sea level is not how much the ocean will expand as it warms, nor how fast Greenland will melt, but how the Antarctic Ice Sheet will behave. Most of that uncertainty comes down to a question of physics that scientists have not yet resolved: whether ice, when stacked into a cliff tall enough, simply cannot hold itself up. The hypothesis known as Marine Ice Cliff Instability, or MICI, proposes that once the floating ice shelves ringing Antarctica are lost, the exposed cliffs behind them could begin collapsing in a self-sustaining chain reaction, dumping ice into the ocean far faster than any process currently observed. If correct, it would roughly double projected sea level rise this century. It is also, importantly, a hypothesis under serious and active challenge. Here is how the proposed mechanism works, and where the science actually stands.

 

1. Ice Shelves Support Glaciers

 

The starting point is well established and not controversial. Around much of Antarctica, glaciers flowing off the land do not simply calve at the coastline. They extend out over the sea as floating ice shelves, vast slabs of ice hundreds of metres thick, which run aground on seabed highs and press against the walls of the bays that contain them. This friction pushes back against the ice behind, acting as a brake on the glaciers feeding them.

Glaciologists call this buttressing, and it is the reason ice shelves matter enormously even though they are already afloat and their melting adds nothing directly to sea level. They are the restraint holding back the ice that does matter. Remove the buttress and the glacier behind it speeds up, delivering land ice into the ocean faster than before. Everything in the MICI hypothesis follows from what happens after that restraint is lost.

 

2. Warming Oceans Weaken the Ice

 

Ice shelves are attacked from two directions. Warm ocean water circulating into the cavities beneath them melts them from below, thinning them and eroding the grounding points where they make contact with the seabed. From above, summer meltwater pools on the surface and can drain into crevasses, where the weight of the water forces the cracks deeper, a process called hydrofracturing that can shatter a shelf from the top down.

Both processes are observed and understood. Basal melting driven by warm water intrusion is currently the dominant driver of Antarctic ice loss, and it is why the Amundsen Sea sector, including Thwaites and Pine Island glaciers, is the most closely watched region on the continent. Hydrofracturing is the mechanism thought to explain the most dramatic ice shelf collapse ever witnessed, and it is the trigger that MICI depends on to expose the cliffs in the first place.

 

3. Ice Shelves Break Apart

 

When a shelf is sufficiently weakened, it can disintegrate with startling speed. The definitive example came in early 2002, when the Larsen B Ice Shelf on the Antarctic Peninsula shattered over roughly five weeks, sending some 3,250 square kilometres of ice, an area comparable to a small country, into the sea. Satellite images captured a shelf that had persisted for thousands of years breaking into a slurry of icebergs in little more than a month.

What followed is the crucial observational evidence. The glaciers that had fed Larsen B, freed of their buttress, accelerated dramatically, flowing two to six times faster than before, with mass loss from the region jumping from a few gigatonnes a year to tens of gigatonnes. This confirmed the buttressing theory in the most direct way possible. Larsen B is therefore firmly established as proof that shelf collapse accelerates glaciers. What it did not clearly demonstrate is the next step in the chain.

 

4. Tall Ice Cliffs Become Unstable

 

Here is the heart of the hypothesis. When a shelf disappears, the glacier behind it terminates in an exposed cliff of ice standing above the waterline. Ice is a material with finite strength, and the taller a cliff stands the greater the stress at its base. Beyond some threshold, commonly estimated at around ninety to one hundred metres above the water, the ice should not be able to support its own weight and would fail structurally, collapsing into the sea. The glaciologist Jeremy Bassis has captured the intuition neatly: there is a reason skyscrapers can only be built so tall.

The instability comes from the geometry that follows. Much of West Antarctica sits on bedrock that deepens inland, so each time a cliff collapses, the newly exposed face behind it is taller still, which makes it even less stable. The result would be a runaway process, each collapse triggering the next, retreating inland at speeds far beyond anything observed. This was the mechanism proposed by Robert DeConto and David Pollard in an influential 2016 paper, which found that including it raised projected Antarctic contributions to twenty-first century sea level by up to an order of magnitude.

 

5. Glacier Flow Accelerates

 

In the MICI scenario, repeated cliff failure combines with the loss of buttressing to accelerate the glacier's flow toward the sea, feeding still more ice to the calving front and sustaining the retreat.

This step is also where the most serious scientific objection arises, and it is worth setting out clearly because it turns on an elegant piece of physics. When an ice shelf is removed, the glacier behind does indeed accelerate, but accelerating ice also stretches and thins. A thinner glacier presents a shorter cliff, and a shorter cliff is more stable, not less. In other words, the very acceleration that MICI relies on may be self-limiting, lowering the cliff below the failure threshold before the chain reaction can establish itself. Other stabilising factors may reinforce this, notably mélange, the dense raft of icebergs and broken ice that accumulates in front of a calving glacier and can push back against the cliff, suppressing further collapse.

 

6. More Land Ice Enters the Sea

 

Whatever the mechanism, the consequence of accelerated flow is the same and this part is unambiguous physics. Floating ice, whether sea ice or an ice shelf, has already displaced its own volume of water, so its melting does not raise sea level. Ice resting on land has not, so every tonne of it that reaches the ocean adds directly to global sea level.

This distinction is why attention concentrates so heavily on the marine-based sectors of the West Antarctic Ice Sheet, where the bed lies below sea level and slopes downward inland. Thwaites Glacier alone, roughly 120 kilometres wide and nicknamed the Doomsday Glacier, holds enough ice to raise sea level by around two feet on its own, and its position guards a much larger interior basin whose eventual loss would be measured in metres.

 

7. Scientists Model the Risk

 

Because MICI has never been observed operating anywhere, the debate must be settled through modelling, laboratory work on ice fracture, satellite observation, and the geological record of past warm periods. This is exactly where the last decade of research has landed, and the results have shifted substantially against the original hypothesis.

Three lines of evidence stand out. A 2019 study in Nature reassessed the uncertainties and found that MICI was not required to reproduce sea level during the mid-Pliocene, the last interglacial, or the satellite era, and that projections without it fell back into line with previous estimates. A 2024 study led by Mathieu Morlighem, published in Science Advances, implemented a more physically grounded representation of cliff failure across three independent high-resolution ice sheet models and simulated the collapse of Thwaites' ice shelf. All three models found that Thwaites would not undergo runaway retreat during this century, and that even when the grounding line was forced back into deeper basins to expose taller cliffs, the resulting thinning and acceleration reduced the calving rate and stabilised the cliff. By their calculation, the calving rate would need to be roughly twenty-five times higher than their parameterisation to trigger MICI-style retreat. It should be said that the authors were careful about what they were and were not claiming, noting explicitly that their results do not suggest West Antarctica is stable, since the separate and better-established process of marine ice sheet instability, driven by grounding line retreat into deepening bedrock, remains a genuine long-term threat.

 

8. Coastal Risks Increase

 

If MICI does occur at scale, the consequences would be severe. The IPCC's central projections put global mean sea level rise at roughly half a metre to one metre by 2100 depending on emissions, and MICI could plausibly double that, with the difference between those two futures measured in the fate of coastal cities, ports, deltas, and low-lying island nations, and in the timescales available for adaptation.

The honest scientific position today is one of reduced but not eliminated concern. MICI entered the IPCC's most recent assessment not as a central expectation but as part of a low-likelihood, high-impact storyline, precisely the category reserved for outcomes that cannot be ruled out and would be catastrophic if realised. The evidence of the past few years has pushed the hypothesis further toward the unlikely end of that range, at least for this century, and researchers have pointed out that this matters practically, because planners deciding whether to build sea walls or relocate communities need projections grounded in defensible physics rather than in worst cases. Yet uncertainty runs in both directions. Nobody has watched a hundred-metre ice cliff stand unsupported for long, the models still simplify fracture mechanics considerably, and the timescale beyond 2100 remains far more open than the timescale within it. The most defensible summary is that marine ice cliff instability now looks less likely to drive catastrophic sea level rise this century than it appeared a decade ago, that Antarctic ice loss remains the dominant uncertainty in sea level projections regardless, and that the underlying driver of all of it, ocean and atmospheric warming, is not in doubt at all.

 

Did You Know?

 

Marine Ice Cliff Instability remains an active area of scientific research rather than a settled conclusion, and its status has genuinely shifted. While the ice sheet models that first introduced it projected dramatically accelerated Antarctic ice loss, subsequent work using multiple independent models and more realistic fracture physics has found that the mechanism may be far less likely to operate this century than originally feared, in part because glaciers that accelerate also thin, lowering the very cliffs that were supposed to collapse. What has not changed is the underlying vulnerability. No one has ever observed a runaway ice cliff collapse in nature, which is both the reason for scepticism and the reason for caution: the process is unprecedented in the observational record, and so is the rate at which the ocean around Antarctica is now warming.

 

Note: This article reflects the state of ice sheet science as of mid-2026, drawing on sources including studies published in Nature, Science Advances, and Nature Communications, together with NASA, NSIDC, and the IPCC Sixth Assessment Report. Marine ice cliff instability is a contested hypothesis, and the balance of evidence continues to evolve.