When researchers lowered sonar instruments into the frigid waters at the base of Alaska’s LeConte Glacier, they expected the underwater ice face to be melting. What they did not expect was how fast. Their measurements, published in the journal Science, showed the glacier’s submerged front dissolving at rates roughly 100 times greater than the leading theoretical models had predicted. That finding, now reinforced by a growing body of research from Greenland to Antarctica, is forcing glaciologists to rethink how quickly the world’s ice sheets could raise sea levels.
“The discrepancy was striking,” David Sutherland, an oceanographer at the University of Oregon who led the LeConte study, told reporters at the time. The team used repeat multibeam sonar surveys to build three-dimensional maps of the glacier’s underwater boundary with the ocean, capturing shape changes over days and weeks. It was the first time anyone had directly measured submarine melt rates at a tidewater glacier, and the results suggested that standard ocean-mixing theory was missing something fundamental about how warm seawater attacks ice.
A pattern across ice sheets
The LeConte findings did not stay isolated for long. A 2022 analysis published in Nature Geoscience linked regional variations in submarine melting to glacier retreat and acceleration across large sections of the Greenland Ice Sheet. Led by Donald Slater and Fiamma Straneo, the study distinguished which glacier sectors respond primarily to warming ocean water and which are driven more by rising air temperatures. The work connected a process happening meters below the waterline to ice loss visible from orbit, showing that underwater melt does not just thin glaciers – it changes how fast they slide toward the sea.
A separate line of research has uncovered a physical mechanism that helps explain why real-world melting so dramatically outpaces theory. In laboratory experiments published in Nature Geoscience in 2023, scientists demonstrated that tiny air bubbles trapped inside natural glacier ice, released as the ice melts, churn the thin boundary layer of water clinging to the ice face. That turbulence accelerates heat transfer and boosts melt rates compared with the bubble-free pure ice used in traditional models. Because real glacier ice is riddled with trapped air, the effect represents a systematic blind spot in older estimates.
Antarctica’s hidden plumbing
The implications extend well beyond Alaska and Greenland. At Thwaites Glacier in West Antarctica, often called the “Doomsday Glacier” because of its potential to reshape coastlines worldwide, researchers have identified another accelerant. A 2025 study in Nature Communications found that pulses of freshwater draining from lakes beneath the ice sheet alter ocean circulation patterns under the floating ice shelf, intensifying melting right at the grounding line – the critical boundary where the glacier lifts off bedrock and begins to float. That process creates a feedback loop: freshwater discharge drives melting, melting drives retreat, and retreat exposes thicker ice to warm water.
Observations from instruments deployed beneath Antarctic ice shelves have added texture to that picture. Researchers at Scripps Institution of Oceanography have described storm-like ocean circulation beneath the ice near Thwaites, with powerful currents funneling warm water against the glacier’s underbelly in bursts far more intense than steady-state models assume.
Modeling work at Bowdoin Glacier in northwest Greenland, published in Frontiers in Earth Science, adds yet another layer. Numerical simulations showed that when warm water carves an undercut into the base of a glacier’s vertical front, the overhanging ice fractures and eventually collapses in large calving events. The geometry of the undercut – its depth, its shape – directly controls when and where those colving blocks break free, meaning underwater melting effectively sets the schedule for above-water ice loss.
Thinning ice shelves compound these effects across both polar regions. When sustained basal melting reduces the thickness of floating ice shelves, those shelves lose their ability to buttress the glaciers behind them. A peer-reviewed analysis in Scientific Reports confirmed that changes in buttressing strongly influence how much ice glaciers discharge into the ocean, with the shape of the underlying bedrock determining how quickly each glacier responds. Glaciers resting on beds that slope downward inland are especially vulnerable, because every increment of retreat exposes a thicker cross-section of ice to warm water.
What scientists still cannot pin down
For all the progress, significant gaps remain. The sonar measurements that revealed extreme melt rates at LeConte Glacier were collected at a single site over a limited window. No comparable field campaign has yet replicated those measurements at other tidewater glaciers with different geometries, ocean temperatures, or seasonal ice conditions. Scaling site-specific snapshots to ice-sheet-wide projections requires assumptions about how often peak melt conditions occur and how they interact with calving cycles and sea ice cover – assumptions that remain largely untested.
The bubble-enhancement mechanism, while convincingly demonstrated in the lab, has not been quantified at the scale of a full glacier front in open ocean conditions. Sediment-laden water, varying salinity, and turbulent currents could amplify or dampen the effect in ways that controlled experiments cannot capture. How much bubble-driven turbulence contributes relative to other processes, such as buoyant freshwater plumes rising along the ice face, is still an open question.
At Thwaites, the connection between subglacial lake drainage and grounding-line retreat remains a single case study. Each Antarctic glacier has its own subglacial plumbing and ocean forcing, and the data needed to map those differences across the continent have not been collected. Deploying instruments beneath hundreds of meters of ice in some of the most remote waters on Earth is slow, expensive, and dangerous work.
Perhaps most consequentially, many of the large-scale ice sheet models used to project future sea level rise still rely on melt-rate formulas calibrated to older theory rather than direct measurements. If the hundred-fold discrepancy observed at LeConte applies broadly, projections of ice sheet contributions to sea level rise could require substantial upward revision. As of May 2026, that revision has not been formally incorporated into assessment-level reports like those from the Intergovernmental Panel on Climate Change, meaning the numbers most commonly cited in policy discussions may not fully reflect the latest process-level science.
What the evidence means for coastal risk
Across Greenland and Antarctica, independent research teams using different methods have converged on the same conclusion: submarine melting is more efficient and more dynamically important than the models long assumed. The direction of the evidence is consistent, even if the precise magnitude remains under debate.
The remaining uncertainties mostly concern how much worse the situation could be, not whether these processes are real. As new sonar campaigns, autonomous underwater vehicles, and improved ice-ocean models come online, projections of sea level rise are more likely to shift upward than downward.
For the roughly 900 million people living in low-lying coastal zones worldwide – a figure from the IPCC’s Sixth Assessment Report – the practical takeaway is straightforward. Coastal planning built on conservative, older projections may understate long-term exposure. That does not mean catastrophic sea level jumps are inevitable, but it does argue for flexible, adaptive strategies that can accommodate faster-than-expected ice loss. The glaciers are not waiting for the models to catch up.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
