Retreat is not a steady process, but happens in short bursts
Twenty thousand years ago, retreating ice sheets carved ridges into the Norwegian seafloor at speeds of up to 600 metres per day — a geological memory that scientists are now reading as a warning. New research published in Nature reveals that ice collapse can move twenty times faster than anything satellites have captured in half a century of observation, and that Antarctica's flat-bedded glaciers, including the imperilled Thwaites, sit on terrain where the same physics could apply. The distance between what we have measured and what the Earth has already done is, it turns out, vast — and the coastlines of the world's great cities stand in that gap.
- Ancient ridges etched into the Norwegian seafloor reveal ice sheets once retreated at 600 metres per day — a speed twenty times beyond the maximum ever recorded by modern satellites.
- The urgency is geographic: Antarctica's Thwaites glacier rests on the same kind of flat terrain that enabled Norway's catastrophic ancient collapse, and current melting rates are already sufficient to trigger similar rapid pulses.
- Fifty years of satellite data, once considered the benchmark for understanding ice loss, now appears dangerously incomplete — a blink against the millennia-scale record preserved in 7,600 seafloor ridges.
- The most alarming unknown is duration: pulses lasting days are serious, but pulses lasting months could compress centuries of projected sea level rise into decades, outpacing every coastal defence on Earth.
- Hundreds of major cities — New York, London, Shanghai, Mumbai — face a future where storm surge and flooding arrive not in the distant future of models, but within living memory, if the ice behaves as its ancestors did.
Geologists studying the seafloor off Norway have uncovered a record that should unsettle anyone living near a coast. Etched into the seabed 20,000 years ago, as massive ice sheets collapsed at the end of the last ice age, are 7,600 parallel ridges stretching across 30,000 square kilometres. Each ridge formed at the grounding line — where the base of an ice sheet met the ocean — as tidal movements twice daily compressed sediments beneath the ice. As the ice melted and the grounding line retreated inland, it left behind these marks like growth rings. Measuring the distance between them, scientists could calculate exactly how fast the ice had moved. The answer: up to 600 metres per day, twenty times faster than anything satellites have recorded in fifty years of observation.
The study, led by Dr Christine Batchelor at Newcastle University and published in Nature, carries immediate implications for Antarctica. Thwaites glacier and other West Antarctic ice sheets rest on similarly flat terrain — the very geometry that enabled Norway's ancient rapid collapse. Batchelor's team found that present-day melting rates are already sufficient to trigger the same kind of pulses there. If a sea level rise projected to unfold over 200 years were instead compressed into 20, coastal defences worldwide would be catastrophically overwhelmed. Hundreds of cities — New York, London, Shanghai, Mumbai — sit on coastlines already growing more vulnerable to flooding and storm surge.
The critical question is not whether such pulses can happen, but how long they last. The Norwegian evidence suggests some ancient pulses continued for up to eleven days; Batchelor suspects they could extend for months. At Pope Glacier in West Antarctica, satellites recorded a sustained retreat of 30 metres per day for three and a half months — a fraction of the speeds now revealed as physically possible. If rapid pulses in Antarctica sustain for months rather than days, current sea level models may be dangerously conservative.
What the geological record offers is not a prediction but a demonstration: this is what ice sheets are capable of doing. The physics is straightforward — on flat seafloor, modest melting at the base of an ice sheet can lift and shift the grounding line far inland in short bursts. Once retreat begins at this scale, researchers at the Potsdam Institute note, it is generally irreversible. The ridges on the Norwegian seafloor are not a relic of a distant world. They are a mirror held up to the present, and what they reflect is the range of futures still available to us.
Geologists studying the seafloor off Norway have discovered something that should unsettle anyone living near a coast: ice sheets can vanish into the ocean far faster than we thought possible. The evidence comes not from satellites or recent measurements, but from the geological record—from ridges etched into the seabed 20,000 years ago when massive ice sheets collapsed at the end of the last ice age. Those ancient patterns reveal retreat rates of up to 600 metres in a single day, speeds that are twenty times faster than anything satellites have recorded in the past fifty years.
The study, published in Nature and led by Dr Christine Batchelor at Newcastle University, examined 7,600 parallel ridges across 30,000 square kilometres of Norwegian seafloor. These ridges formed at the grounding line—the boundary where the base of an ice sheet met the ocean. Twice daily, tidal movements lifted and lowered the ice, compressing sediments into ridges. As the ice melted over days and weeks, the grounding line retreated inland, leaving behind sets of these parallel marks like growth rings. By measuring the distance between them, scientists could calculate precisely how fast the ice was moving. The results were startling: retreat speeds ranged from 50 to 600 metres per day, with the fastest rates occurring where ice sheets rested on relatively flat seafloor.
Why does this matter now? Because Antarctica's ice sheets, including the ominously named Thwaites glacier, sit on similarly flat terrain. If present-day melting rates are sufficient to trigger the same kind of rapid pulses in Antarctica, the implications are severe. Batchelor explained the stakes plainly: if a sea level rise expected to unfold over 200 years instead compressed into 20 years, coastal defences worldwide would be catastrophically inadequate. Hundreds of major cities—New York, London, Shanghai, Mumbai, and countless others—sit on coastlines increasingly vulnerable to storm surge and flooding. The West Antarctic ice sheet may have already crossed a threshold beyond which major losses are unstoppable, eventually leading to metres of sea level rise.
Previous estimates of ice sheet collapse have relied almost entirely on satellite data, which covers only about fifty years. That's a blink in geological time. The Norwegian ridges stretch back thousands of years, capturing a far wider range of conditions and revealing patterns satellites simply cannot detect. As Andrew Shepherd, a glaciologist at Northumbria University not involved in the study, noted, satellites typically track changes only once per year at most. The rapid pulses happen in bursts—the Norwegian evidence suggests some lasted up to eleven days, though Batchelor suspects they could extend for months. At Pope Glacier in West Antarctica, satellites did record a sustained retreat of 30 metres per day for three and a half months, but 600 metres per day could not persist for a year without exhausting the ice entirely.
The mechanism is straightforward physics. On flat seafloor, a relatively small amount of melting at the base of an ice sheet can lift a large section of it, shifting the grounding line far inland. On steeper slopes, the same melting produces slower retreat. This is why the Norwegian data matters so acutely: it shows what ice sheets are physically capable of doing under the right—or rather, wrong—conditions. Batchelor's team found that present-day melting rates are already sufficient to trigger these rapid pulses across flat-bedded areas of the Antarctic ice sheet, including at Thwaites.
The critical unknown is duration. If rapid retreat pulses last only a week or two, the long-term impact on sea level, while significant, remains within certain bounds. But if they sustain for months or longer, the acceleration of sea level rise could be far more dramatic than current models predict. Johannes Feldmann at the Potsdam Institute for Climate Impact Research emphasized the gravity: ice sheet retreat, once begun on this scale, is generally irreversible. The geological record is not a prediction—it is a demonstration of what the physics allows. What happens next depends on whether Antarctica's ice sheets begin to behave like their Norwegian ancestors did twenty millennia ago.
Notable Quotes
Our research provides a warning from the past about the speeds that ice sheets are physically capable of retreating at. Pulses of rapid retreat can be far quicker than anything we've seen so far.— Dr Christine Batchelor, Newcastle University
Retreat is not a steady process, but happens in short bursts. We didn't spot that from space because we tend to track changes once per year at most.— Prof Andrew Shepherd, Northumbria University
The Hearth Conversation Another angle on the story
So these ridges in the seafloor—they're basically a record of how fast ice melted?
Exactly. The tides lifted and lowered the ice twice a day, squeezing sediment into ridges. As the ice retreated, it left behind these parallel lines. The spacing tells you how far the grounding line moved each day.
And the speeds they found—600 metres a day—that's genuinely new information?
It's not that ice couldn't do this before. It's that we didn't know it could. Satellites only go back fifty years. These ridges go back thousands. We were looking at a tiny slice of time and missing the full range of what's possible.
Why does flat seafloor matter so much?
Because melting at the base of the ice doesn't have to do as much work. On a slope, the ice stays grounded longer. On flat ground, a little melting lifts a huge section and the whole thing shifts inland rapidly. Antarctica has a lot of flat seafloor.
Is this saying sea level rise is about to accelerate?
It's saying it could. The conditions that allowed these rapid pulses in Norway exist in Antarctica now. Whether they actually trigger depends on whether the melting sustains long enough. A few days of 600-metre retreat is one thing. Months of it is another entirely.
And if it does sustain?
Then timelines compress. What we thought would take two centuries could happen in two decades. Coastal cities aren't built for that kind of change.