Why in news?
A study published in Nature Communications on 12 October 2026 projects greater ice loss from West Antarctica's Thwaites Glacier than some earlier estimates. Researchers represented seawater entering a broad transition beneath the glacier, rather than treating grounded and floating ice as meeting along one sharp line. Their model was checked against satellite records of past changes before being extended to 2100. In its higher grounding-zone configuration, Thwaites contributed about 30 millimetres, or 3 centimetres, to global mean sea-level rise by that date. This was roughly double the earlier model-comparison mean under the same baseline conditions. That can affect longer-term planning for coasts exposed to rising seas. The result concerns one glacier's projected contribution, not a doubling of all global sea-level rise or a prediction of its immediate total collapse.
A large outlet from the West Antarctic Ice Sheet
Thwaites drains ice from the West Antarctic Ice Sheet into the Amundsen Sea, part of the Southern Ocean surrounding Antarctica. The Antarctic Place-names Committee locates the feature near 75°30′ south and 106°45′ west. It lies in the same broad coastal sector as Pine Island Glacier. These are outlets through which ice accumulated over the continent moves towards the ocean, rather than isolated blocks floating far from land.
The International Thwaites Glacier Collaboration describes its drainage basin as approximately 192,000 square kilometres. That is the area feeding the glacier, not the area of its floating shelf alone. Distinguishing the two is essential: losing a floating section and losing the grounded ice that feeds it are different events. A photograph of a broken ice front cannot, by itself, establish that the whole drainage basin has collapsed.
Why grounded ice and floating ice matter differently
Ice resting on the continent's bedrock can flow into the ocean and eventually become afloat. The transition is commonly described using a grounding line. A floating ice shelf already displaces seawater, so its loss has a different direct effect from adding previously grounded ice to the ocean. However, a shelf can provide resistance to ice flowing behind it. Changes to that resistance can affect the discharge of grounded ice.
Warm ocean water can also reach the underside of a glacier and melt it from below. Here, “warm” means warm enough relative to the ice and local freezing conditions, not water that would feel tropical. Melting can thin the ice and alter where it remains supported by the bed. The relationship between ice thickness, flotation and the shape of the seabed therefore helps determine how the glacier changes.
The new emphasis is a zone, not a single boundary
The University of California, Irvine's account explains that satellite observations show seawater entering kilometres beneath ice near the transition. Tides can lift ice temporarily and let water reach areas that are not permanently afloat. Treating this shifting contact as an abrupt line can miss part of the melting process. A grounding zone instead represents a region across which the ice's contact with the bed and exposure to seawater vary.
This is a change in how a physical process is represented, not the discovery that Antarctica is melting for the first time. The researchers used the Ice-sheet and Sea-level System Model to test different intrusion lengths and melt rates. They then compared simulations with independent records of ice speed, thickness change and grounding-line migration. The purpose was to find configurations consistent with the glacier's observed recent evolution before projecting its future.
What the three-centimetre projection does and does not say
The historical comparison covered 1996–2023. Forward simulations retained a baseline based on average climate conditions over that period, rather than imposing additional future warming. In the paper's wider-zone configuration, the projected contribution reached approximately 30 ± 6 millimetres by 2100. The uncertainty is part of the result, not a promise of an exact rise. Other tested configurations produced smaller contributions.
Thirty millimetres equals three centimetres. That figure is added to global mean sea level from this glacier in the stated experiment. It is not the total rise expected from thermal expansion, Greenland, other glaciers and all other contributors combined. The comparison with earlier projections also concerns a particular model-comparison baseline, not every previous scientific estimate. Earlier projections had a broad range; this study's estimates lie towards its upper end rather than outside all previous possibilities. Saying simply that “sea-level rise will double” would change the meaning of the finding.
The seafloor can temporarily restrain retreat
The university's explanation highlights the role of submarine ridges that currently help stabilise the main glacier trunk. Once ice loses contact with a stabilising ridge, it can retreat towards another bedrock high. This makes the geometry below the glacier important alongside the amount of heat reaching it. The simulations explore how seawater intrusion can affect that contact; they do not supply a single guaranteed date for the whole glacier to disappear.
Thwaites' total sea-level potential is a different quantity again. The collaboration's factsheet estimates about 65 centimetres if its grounded ice were lost completely. That describes the scale of ice available to contribute, not what the October study projects by 2100. Wider West Antarctic change could involve still more ice over much longer periods. Combining all these figures into one immediate “doomsday” event would erase the difference between a stock of ice and its rate of loss.
Model improvement does not remove every uncertainty
Checking a model against several observations is more informative than matching just one changing feature. It still does not make the future an observation. The paper holds its ice front fixed and uses a time-averaged intrusion length. It also simplifies ocean-driven melting rather than fully resolving the circulation of water through an evolving cavity. Those choices isolate a process while leaving other influences to be investigated more completely.
A separate British Antarctic Survey study, publicised on 13 October, asked what happened if ocean-driven melting was switched off in simulations. Thwaites continued losing ice during that experiment, although the loss slowed. That is evidence of a modelled long-lasting response to earlier change, not an actual cooling experiment in Antarctica. It must not be confused with the Nature Communications study or used to imply that reducing greenhouse-gas emissions no longer matters.
Why a global mean matters to coastal places
Sea-level contributions from distant ice loss become relevant to coasts because they add to the background water level. Local effects also depend on land movement, ocean conditions, tides and the particular coastline. The October projection therefore cannot be read as an identical three-centimetre increase at every harbour. Its practical significance is that representing an overlooked process can change the evidence used to assess longer-term coastal exposure.
Conclusion
The study strengthens the case for representing the grounding zone explicitly when modelling Thwaites Glacier. Its lesson is about a physical process and the quality of projections, not a newly announced date of catastrophe. Better observations below the ice and more complete ice–ocean models remain necessary. Coastal planning must take the risk seriously while keeping the glacier's total potential, this century's projections and local sea-level changes distinct.