Ice cliffs break off faster than models assumed, then stop, sea-level study finds

Picture a wall of ice 100 metres tall, as high as a 30-storey building, standing where a glacier meets the sea. For a decade, one idea has said such cliffs could keep collapsing, each fall exposing a taller face behind it. A new study of ice cliffs and sea level rise, published in Science on 15 October, finds the fast falls are real. The chain reaction has not shown up.

That idea is called marine ice cliff instability, or MICI. It sits behind some of the highest sea-level projections for this century. So a test built on real measurements matters to anyone who lives near a coast or plans for one.

More than 15,000 cliff measurements, mostly from a laser in orbit

The lead author is Caroline Needell, a PhD student in the MIT–WHOI Joint Program. She worked with glaciologist Catherine Walker of the Woods Hole Oceanographic Institution (WHOI) and Jeremy Bassis of the University of Michigan.

They studied 15,330 measurements of ice cliffs at the edges of the Greenland and Antarctic ice sheets. Most came from NASA’s ICESat-2 satellite, launched in 2018. It times laser pulses bouncing off the ice to measure surface height, and repeats its track every 91 days.

The team compared cliff height with ice thickness, water depth, cracks and how fast the front broke off. Three findings stand out:

• Faster than modelled. Tall cliffs that retreated broke off faster than the formulas ice-sheet models use.
• Short-lived. Every retreating glacier in the study settled down again.
• Self-limiting. As the glaciers thinned and pulled back, the front reached places where ice thickness changed little, so the cliff stopped growing taller.

The average cliff stood about 30 metres above the water, well below the roughly 100 metres older models treat as the point of failure.

Why height alone does not decide when a cliff gives way

What is marine ice cliff instability? It is the idea that once the floating ice shelf in front of a glacier breaks away, the exposed cliff is too tall to hold its weight. Each ice cliff collapse then leaves a taller one behind.

Models usually judge that risk by height alone. The study says other things matter. Taller cliffs had thicker mélange, a jam of icebergs and sea ice pressed against the front, which may act as a brake. Deeper water may help too.

Needell told MIT News the pattern was “rapid retreat, punctuated by periods of stabilization.”

This is not simply good news. The University of Michigan says some estimates of higher, faster sea-level rise, once dismissed by some scientists, “could be more likely than previously thought”. Bassis said the examples so far suggest glaciers restabilise, so a runaway, total break-up is not supported. Whether that holds as warming pushes ice sheets further, he added, “we do not yet know”.

The limits are real. Few glaciers had very tall cliffs. The rates cover months, not decades. Earth.com notes the team studied retreating glaciers, so the speeds are likely at the high end. We found no reaction from independent glaciologists yet.

Bursts, not a runaway: what ice cliffs and sea level rise mean for coasts

Cities plan sea walls, drains and flood maps around ranges like those of the IPCC, the UN’s climate science panel. Under very high emissions, its 2021 report puts the likely rise at 0.63 to 1.01 metres by 2100. It says close to 2 metres cannot be ruled out, partly because of processes like this one.

A 2016 Nature study used MICI to suggest Antarctica alone could add about 1 metre or more by 2100 under unchecked emissions. Later work scaled that back. This survey points to a middle path: short bursts of fast ice loss deserve a place in plans, while the most extreme chain reaction has less support. If your city has a sea-level plan, check that it covers the upper part of these ranges.

Earlier this month a separate study found seawater reaching further under Thwaites Glacier than models assumed. Different process, same lesson: measurements keep refining the models.

Satellites narrow the worst case without closing it

On balance, this is careful work. It swaps a long modelling argument for thousands of measurements, and it is honest in both directions. It is also a quiet tech story: a laser in orbit turned a hard question into something we can check.

The weak points are plain. Few very tall cliffs, a short window and a chosen sample make this evidence, not proof.

The likelier reading is calm. These cliffs can lose ice in sudden bursts, faster than expected, but none has been seen to collapse without end.

ICESat-2 keeps passing over the same cliffs every 91 days. The coming years of those passes will show whether the brakes hold as the oceans warm.