Connect with us

NEWS

Seafloor Clay Shows Thwaites Failed in Every Warm Period

China clay in Amundsen Sea cores vanished in every 800,000-year warm period, fingerprinting Pacific deep water as the switch already acting on Thwaites.

Published

on

Seafloor clay from 4,180 metres down on Antarctica’s Pacific margin shows Thwaites Glacier lost its marine ice basin in every warm period of the last 800,000 years. A British Antarctic Survey team read that history in kaolinite, the mineral sold as china clay, in mud cores from the Amundsen Sea.

The clay stopped arriving whenever Pacific deep water intensified against the ice. That is the same class of current already melting Thwaites and Pine Island glaciers from below.

China Clay as a Glacier Attendance Record

Kaolinite does not form under today’s Antarctic ice. It needs warm, wet, acidic soils, the kind that covered parts of the continent more than 34 million years ago, before a lasting ice sheet took hold. Those old beds still sit under West Antarctica, and the only way the mineral reaches the deep ocean in bulk is if grounded ice grinds it out and sends the dust to sea.

Dr. Claus-Dieter Hillenbrand, a marine geologist at the British Antarctic Survey who led the study, said the tap is either on or off. When the kaolinite signal at the drill sites suddenly collapses, he said, marine-based ice in the Amundsen Sea sector of West Antarctica must have melted and retreated far back into the continent’s interior.

When the kaolinite signal at our drill-core sites suddenly collapses, it tells us the marine-based ice in the Amundsen Sea sector of the West Antarctic must have melted and retreated far back into the continent’s interior.

Claus-Dieter Hillenbrand, marine geologist, British Antarctic Survey

The paper, published 9 November 2026 in Nature Communications, finds that subglacial erosion under Thwaites is the main offshore source of that clay. Pine Island Glacier sends some too, though its mix is richer in illite. About 80% of the West Antarctic Ice Sheet rests below sea level on a bed that slopes inland, and those two outlets drain one third of it into the eastern Amundsen Sea Embayment.

WHAT KAOLINITE IS RECORDING

  • The mineral: Kaolinite, sold as china clay, forms in low-pH soils in tropical or swampy climates, not under modern Antarctic ice.
  • The factory: Grounded ice on old sedimentary beds under Thwaites, and to a lesser extent Pine Island, grinds the clay and carries it offshore.
  • The off switch: When that ice ungrounds or pulls far inland, delivery of kaolinite-rich debris to the rise almost stops.

The china clay locked beneath the ice sheet is still being ground today, which is why modern mud near the drill sites is not empty of it. The warm-period layers are.

Pacific Mud Rode South on Deep Water

Retreat is only half of the record. The team needed a carrier for the heat, and they found it in a second clay, smectite. Local smectites weathered from Antarctic rock and volcanoes are relatively poor in magnesium. A magnesium-rich variety in the warm-period layers matches clay from the deep Pacific floor far to the north, formed as seawater alters basalt.

Something had to move that Pacific clay south and drop it in the nearshore muds at the same time the kaolinite vanished. Hillenbrand’s group points to stronger flow of Circumpolar Deep Water toward the West Antarctic shelf. That current would bring the foreign smectite with it and, at the same time, melt the grounded ice whose grinding produces the kaolinite signal.

THE TWO CLAYS AT SITE U1533

Signal Cold-period mud Warm-period mud
Kaolinite from Thwaites beds 12% in modern mud nearby; at least 8% within 400 km Below 5%, and in some layers below 2%
Magnesium-rich Pacific smectite Absent or drowned by local clay Present, with magnesium contents of 35 to 57%
Summer sea surface versus today No warm spikes of this kind 0.6 to 4.1 C warmer, depending on the period

Local Type-1 smectites carry 5 to 35% magnesium and lack the richest saponite varieties. Type-2 grains from the sub-Antarctic southeast Pacific include those high-magnesium forms. Glacial samples from the rise hold the local mix. Warm-period samples hold both, which is the chemical mark of deep water arriving from the north.

How Warm Were Waters Off West Antarctica?

Shells of the diatom Fragilariopsis kerguelensis change shape with water temperature. Applied to Site U1533, that proxy puts summer sea-surface temperatures during the last major warm period, 130,000 to 115,000 years ago, at 1.3 to 2.2 C warmer than now. Across warm intervals since about 621,000 years ago, the binned estimates run from 0.6 to 1.4 C warmer in one stage to 3.3 to 4.1 C warmer in another.

Global mean air in those warm periods was only 1.0 to 1.5 C above pre-industrial levels, or even cooler, the authors write. The local ocean ran hotter than that global figure. The paper treats that mismatch as thermal amplification at the ice edge, not a simple replay of worldwide air temperature.

That is why the clay record is an ocean story first. Air warmth set the stage. The current that carried Pacific smectite onto the shelf did the work on the ice.

Every Warm Interval Left the Same Gap

Site U1533 sits at 68° 44.1′ S, 109° 01.5′ W, in 4,180 metres of water on the lower western flank of Resolution Drift. Site U1532 sits 62 km east-northeast, in 3,960 metres, nearer the drift crest. Both were drilled from the JOIDES Resolution during IODP Expedition 379 in the Amundsen Sea, which ran from 18 January to 20 March 2019 between Punta Arenas, Chile, and Punta Arenas, with Karsten Gohl and Julia Wellner as co-chief scientists.

Icebergs kept forcing the ship off station, which is why U1533 needed four holes at Site U1533. The deepest reached 383 metres, with 76% recovery. U1532 went to 787 metres, with 90% recovery. Late Pleistocene to Holocene mud at U1533 is only a few tens of metres thick; the Brunhes-Matuyama magnetic reversal, about 773,000 years ago, was picked in holes U1533A and U1533D.

THE WARM PERIODS IN THE MUD

  1. Every warm interval after 800,000 years ago: Kaolinite at U1533 falls below 5%.
  2. Some warm layers: Kaolinite is nearly gone, below 2%.
  3. 563,000 to 533,000 years ago (MIS-14): The same kaolinite crash occurs in a stage that, on the Antarctic margin and globally, does not even look like a classic ice age.
  4. 130,000 to 115,000 years ago (MIS-5e): Maximum ice retreat arrives late in the warm period, after the first global sea-level peak.
  5. The Holocene, the last 11,700 years: Nearby modern mud still holds 12% kaolinite, so the Thwaites factory is still running.

U1532 shows the same pattern, though the warm-period drops are milder, likely because those layers there are condensed at about 1 cm per thousand years. A gap between cores at U1532 means mud from 243,000 to 300,000 years ago was not recovered at that site. U1533 carries the cleaner on-off record.

The Cores Stop Short of a Full Collapse

The loss of the Amundsen Sea drainage basins would have added at least 1.5 metres to global sea level, the authors write. They do not claim the entire West Antarctic Ice Sheet vanished. They cannot prove, and they do not dismiss, model runs that put a West Antarctic share of about 3 to 4 metres into the last interglacial highstand. What the mud requires is ungrounding of Thwaites and a large pullback of Pine Island.

WHERE THE EVIDENCE STILL SPLITS

  • Amundsen sector: Hillenbrand’s clay, smectite and diatom records show those basins were lost in every pre-Holocene warm period since 800,000 years ago.
  • A trans-Antarctic seaway: A 2023 genomic study of the octopus Pareledone turqueti found gene flow that would need open water between the Weddell, Amundsen and Ross seas, most plausibly in the last interglacial.
  • Ice that stayed: An ice core at Skytrain Ice Rise has been read as evidence that the Ronne Ice Shelf survived that warm period near its present size, which fits a partial collapse focused on the Amundsen Sea rather than a wipe-out of all marine West Antarctic ice.

Even if all marine-based ice in the Amundsen catchment went, a land-based ice cap would likely have remained on the Marie Byrd Land dome. A core to the west, PS2551-1, kept kaolinite above 10% through warm periods, against 17% in modern mud there, which is the signature of continued grinding on that dome. Illite chemistry at U1533 in those same warm layers points to leftover ice on the Thurston Island block, not a bare coast.

Raised-beach records in Pine Island Bay show relative sea level falling steadily for 5,500 years with no jump that would mark a huge re-growth of ice. That Holocene stability is why the new clay crash in older warm periods stands out. The present glacier front has not, in this interglacial, yet gone as far as the mud says it went in the ones before.

Under Thwaites the Historic Switch Is Already On

Modern melting along this coast is already driven by Circumpolar Deep Water coming onto the shelf, the paper notes, the same water mass fingerprinted by the Pacific smectite. Greenhouse gases since the 1960s and ozone loss since the 1980s have warmed Southern Ocean water and shifted winds over the Amundsen shelf break, increasing that upwelling. Hold those ocean conditions constant, the authors write, and the Thwaites and Pine Island drainage basins still most likely empty over the coming centuries.

That is a different clock from a complete melt of Thwaites under past natural warmth, which Hillenbrand said might have taken several thousand years. The clay is still arriving, at 12% in modern nearby mud, because the ice is still grounded and still grinding. The warning in the paper is about the state after that signal goes dark.

West Antarctic volcanoes cannot explain the pattern. They have sat under the ice longer than the ice has existed, so they cannot turn a kaolinite tap off in every warm period and on in every cold one. The on-off match is with ocean heat, not with a geothermal background that does not come and go with ice ages.

Dr. Rob Larter, a British Antarctic Survey marine geophysicist and co-author, put the same point in plainer words the day the paper appeared. Natural variations in past warm periods, he wrote, resulted in loss of most of the ice in Thwaites and Pine Island glaciers, at least, and climate change is now supercharging ocean forcing.

Human Warming Compresses a Millennial Collapse

If Thwaites Glacier melted entirely, global sea level would rise 65 centimetres. Pine Island Glacier holds another 51 centimetres. Those two figures sum to 116 centimetres and still sit below the paper’s at least 1.5 metres for the whole Amundsen sector, because other ice in that catchment would go with them. If wider marine parts of the West Antarctic Ice Sheet followed, the paper cites a rise of 3.3 to 5.3 metres. The cores do not show that full loss. They show the first phase that models have long treated as the opening move.

ICE AND SEA LEVEL IN THE NUMBERS

Ice Sea-level equivalent What the new cores show
Thwaites Glacier, complete melt 65 centimetres Its kaolinite factory shut down in every pre-Holocene warm period
Pine Island Glacier, complete melt 51 centimetres Its kaolinite-bearing debris also nearly vanished in those layers
Amundsen Sea drainage sector At least 1.5 metres The authors’ floor for the ice the mud says was lost
Entire marine West Antarctic Ice Sheet 3.3 to 5.3 metres Not proven by these cores; the potential if other sectors follow

Between 2003 and 2019, West Antarctic retreat and thinning added about 7.5 millimetres to global sea level. Model ranges cited in the paper put a likely Antarctic contribution at 8 to 24 centimetres by 2100, depending on emissions. Those century-scale figures are not a timetable for the 65-centimetre Thwaites reservoir. They are the near end of a process the mud says has already run to basin-scale loss, slowly, in every earlier warm period.

Hillenbrand said that even under natural conditions the team found at least a partial collapse of the West Antarctic Ice Sheet, and that what is being done now is making it much worse. Larter said he is convinced future generations will see substantial sea-level rise, and that how fast that rise arrives is really difficult to say. What people are living through now, he said, is an interglacial on steroids.

Frequently Asked Questions

Why does kaolinite in seafloor mud record ice retreat?

Kaolinite in this sector can only be sourced from pre-Oligocene sedimentary strata, because the mineral forms in low-pH soils that Antarctica has not hosted since widespread glaciation began about 34 million years ago. Those beds crop out from the Amundsen Sea mid-shelf inland toward the Byrd Subglacial Basin, so a crash in offshore kaolinite means ice has stopped eroding that inland source, not that the mineral dissolved in seawater.

How much sea level is stored in Thwaites and Pine Island glaciers?

Complete melt of Thwaites would raise global sea level 65 centimetres and complete melt of Pine Island would add 51 centimetres, a combined 116 centimetres from those two outlets alone. The new paper’s at least 1.5 metres applies to the wider Amundsen Sea drainage sector, which includes more ice than the two named glaciers.

Did the entire West Antarctic Ice Sheet collapse in past warm periods?

These cores document ungrounding of Thwaites and large retreat of Pine Island, not a proven wipe-out of the Weddell and Ross sectors. A land ice cap on the Marie Byrd Land dome likely survived, leftover ice on the Thurston Island block kept feeding a different clay mix to U1533, and a Skytrain Ice Rise record has been read as showing the Ronne Ice Shelf still in place near its present extent.

When were the Amundsen Sea sediment cores drilled?

They were recovered in austral summer 2019 on IODP Expedition 379, from 18 January to 20 March, by the JOIDES Resolution working out of Punta Arenas, Chile. Site U1533’s late Pleistocene section sits in the upper tens of metres of a hole that reached 383 metres into mud as old as about 6.4 to 6.75 million years at the base, so the 800,000-year clay story is only the youngest part of that archive.

The U1533 cores are still the attendance sheet. Kaolinite is still arriving in modern mud, which means Thwaites is still grinding its bed. The paper’s warning is about the stretch of time after that signal goes dark.

Harry is the editor and lead writer of STUDIO ONE NETWORKS, an independent title he owns and runs himself. Ten years in journalism, reporting first and editing later, taught him that entertainment and business are one beat seen from two sides: a box office figure is a company number, a streaming deal is a contract, a casting rumour is not a story until someone puts their name to it. He works from the record, whether that is a distributor's statement, a licensing agreement, an interview transcript or a set of published ratings, and checks every number against it before publication. The same rule holds for the rest of the site, which covers news, technology, science, sports, lifestyle, travel, auto and gaming for an audience spread across the world. When he gets something wrong, the article is corrected and the change is noted and dated, under a corrections policy anyone can read. Reader mail is answered by him at support@studioonenetworks.com.

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Trending