The ice sheet is poised on a knife's edge, vulnerable to warming.
Millions of years before human civilization, the West Antarctic Ice Sheet repeatedly collapsed and reformed as Earth's temperatures rose and fell — a cycle now made newly urgent by the pace of human-driven warming. Scientists studying ancient seafloor sediments from the Amundsen Sea have found that the ice sheet retreated deep into the Antarctic interior at least five times during the Pliocene Epoch, when temperatures were only 3 to 4 degrees Celsius warmer than today. The past, in this case, is not merely history — it is a mirror held up to a future that coastal populations worldwide may be forced to inhabit.
- Thwaites and Pine Island glaciers are already shedding ice faster than anywhere else on the continent, and the question scientists are racing to answer is not if they will retreat further, but how catastrophically and how soon.
- Ancient sediment cores from the Amundsen Sea seafloor reveal that the West Antarctic Ice Sheet did not hold steady during past warm periods — it collapsed dramatically, repeatedly, and far inland, leaving behind a geological record of its own instability.
- Chemical fingerprints in iceberg debris — isotopes of strontium, neodymium, and lead — trace the origins of ancient meltwater back to Antarctica's deep interior, confirming that past retreats were not marginal but wholesale.
- The ice sheet's behavior followed a four-stage cycle of basal melting, inland retreat, massive iceberg calving, and eventual readvance — a pattern that repeated at least five times and could be triggered again by continued emissions.
- If the WAIS retreats as far as it did millions of years ago, sea levels could rise by several meters, redrawing the coastlines of cities from Miami to Shanghai and displacing millions of people with little historical precedent for the speed of change.
Two of Antarctica's most troubled glaciers — Thwaites and Pine Island — are losing ice faster than anywhere else on the continent, together shedding more mass than the rest of Antarctica combined. The question haunting climate scientists is not whether they will retreat further, but how far and how fast.
To find answers, a team led by Keiji Horikawa at the University of Toyama turned to the deep past. During the Pliocene Epoch, between 5.3 and 2.58 million years ago, global temperatures ran 3 to 4 degrees Celsius warmer than today and sea levels stood more than 15 meters higher. Sediment cores pulled from the Amundsen Sea continental rise — a submarine archive of ice sheet behavior — held the evidence they were looking for.
The cores revealed a repeating story. Cold periods left thick gray clays on the seafloor; warmer intervals produced thinner greenish layers colored by algae that thrived in open water. Scattered through those warm-period layers were fragments of rock carried from Antarctica's interior by icebergs and dropped as the ice melted. Chemical analysis of these fragments — measuring isotopes of strontium, neodymium, and lead — traced their origins to the Ellsworth-Whitmore Mountains and other deep interior regions, confirming that the ice sheet had not merely thinned at its edges but had retreated far inland.
Between 4.65 and 3.33 million years ago, researchers identified 14 distinct melt intervals and at least five major retreat cycles, each following the same four-stage pattern: basal melting beneath the ice, inland retreat, massive iceberg calving, and eventual readvance as temperatures fell again.
The warning embedded in this ancient record is precise. The West Antarctic Ice Sheet did not permanently collapse during the Pliocene — but it came apart episodically, suddenly, and repeatedly under conditions not far beyond what current emissions trajectories are projected to produce. If it retreats as far inland as it did millions of years ago, sea levels will rise by meters, and the coastlines of cities from Miami to Shanghai to London will be fundamentally redrawn. The sediment cores are not simply a record of deep time. They are a map of where the present path leads.
Two of Antarctica's most unstable glaciers sit in the Amundsen Sea, grinding toward the ocean faster than any ice on the continent. Thwaites and Pine Island are hemorrhaging ice at a rate that keeps climate scientists awake at night. Together they are shedding more mass than the rest of Antarctica combined, and the question that haunts researchers is not whether they will retreat further, but how far, and how fast.
To understand what might happen next, a team of scientists led by Keiji Horikawa at the University of Toyama looked backward—to a time when the Earth was warmer than it is today. The Pliocene Epoch, stretching from 5.3 to 2.58 million years ago, saw global temperatures running 3 to 4 degrees Celsius higher than the present day. Sea levels then stood more than 15 meters above current levels, much of that rise driven by melting Antarctic ice. If the West Antarctic Ice Sheet could collapse then, it can collapse now. The question was: how often did it happen, and what did it look like?
The answers lay in sediment cores pulled from the seafloor during an international drilling expedition. Researchers from Japan, Britain, the United States, and elsewhere examined material recovered from the Amundsen Sea continental rise, a submarine archive that records millions of years of ice sheet behavior in its layered strata. What they found was a repeating pattern. During cold periods, thick gray clays accumulated on the seafloor, marking times when ice extended far across the continental shelf. During warmer intervals, thinner greenish layers formed—colored by microscopic algae that thrived in open water. But these warm-period layers held a crucial signature: tiny fragments of rock scattered throughout, pieces that had been plucked from the Antarctic interior by icebergs, carried across the sea, and dropped onto the bottom as the ice melted.
Between 4.65 and 3.33 million years ago, the researchers identified 14 distinct intervals rich in this iceberg-rafted debris, each one a fingerprint of a major melt event. To trace where the debris originated, the team analyzed the chemical composition of the sediments—measuring isotopes of strontium, neodymium, and lead that vary by rock type and age. The signatures pointed inland, to the Ellsworth-Whitmore Mountains and other interior regions of West Antarctica. The ice sheet had not simply thinned at its margins. It had retreated far into the continent itself, calving massive icebergs that carried evidence of that retreat across the Amundsen Sea.
The sediment record revealed a consistent four-stage cycle. As the climate warmed, basal melting began beneath the ice sheet, triggering retreat. At peak warmth, the ice margin had pulled back so far that enormous icebergs broke free and drifted seaward, laden with sediment from deep within Antarctica. Then, as temperatures fell again, the ice sheet surged forward, readvancing across the shelf and pushing previously deposited material downslope into deeper water. This cycle repeated at least five times during the Pliocene, each retreat taking the ice sheet far beyond where it sits today.
What makes this history urgent is not the past but the future it illuminates. The West Antarctic Ice Sheet did not permanently collapse during the Pliocene, despite conditions warmer than those projected for the coming centuries. Instead, it underwent episodic but rapid retreats—sudden, dramatic pullbacks followed by regrowth. The implication is stark: the ice sheet is not stable. It is poised on a knife's edge, vulnerable to the kind of warming that has happened before and will happen again if emissions continue unchecked. If it retreats as far inland as it did millions of years ago, sea levels will rise by meters. Coastal cities from Miami to Shanghai to London will be redrawn. The sediment cores from the Amundsen Sea are not just a record of deep time. They are a warning about what the present trajectory leads toward.
Notable Quotes
We wanted to investigate whether the WAIS fully disintegrated during the Pliocene, how often such events occurred, and what triggered them.— Professor Keiji Horikawa, University of Toyama
The Amundsen Sea sector of the WAIS persisted on the shelf throughout the Pliocene, punctuated by episodic but rapid retreat into the Byrd Subglacial Basin or farther inland, rather than undergoing permanent collapse.— Professor Keiji Horikawa
The Hearth Conversation Another angle on the story
Why look so far back? We have satellite data from the last few decades. Doesn't that tell us what we need to know?
Satellites show us the present crisis, but they can't show us the full range of what the ice sheet is capable of doing. The Pliocene gives us a natural experiment—a time when the Earth was as warm as it might be in a century or two, and we can see how the ice sheet responded. That's invaluable.
And it collapsed?
Not permanently. That's the unsettling part. It retreated far inland, repeatedly, but then it came back. The ice sheet is not a simple on-off switch. It's more like a system that can flip between states—stable and unstable—depending on temperature. We're pushing it toward instability.
How do they know the debris came from inland? Couldn't it have come from anywhere?
The isotope signatures are like a fingerprint. Different rocks have different chemical compositions depending on their age and origin. By measuring strontium, neodymium, and lead, they traced the debris back to specific mountain ranges in the interior. The ice sheet had to retreat far enough to expose those rocks and let icebergs carry them out to sea.
So this happened five times in the Pliocene?
At least five times, yes. Fourteen distinct melt events in a span of about 1.3 million years. That's not rare. That's a pattern. And each time, the ice sheet pulled back dramatically, then regrew. The question now is whether it will regrow if it retreats again under modern warming.
And the answer?
We don't know yet. But the sediments suggest the ice sheet is far more unstable than we thought. If it retreated that far before, it can do it again. And if it does, the world changes.