For more than two decades, thousands of autonomous robots have been quietly patrolling the depths of the Southern Ocean, diving as deep as 2,000 meters, recording temperature and salinity, then surfacing to beam their findings to satellites overhead. Now that long accumulation of data has helped scientists crack one of the most pressing puzzles in climate science: why Antarctic sea ice, after decades of slow expansion, abruptly began collapsing in 2016.
A study published in the Proceedings of the National Academy of Sciences, led by Stanford oceanographer Earle Wilson and colleagues, pins the reversal on a shift in how the Southern Ocean handles heat trapped in its depths. The findings, summarized by Stanford in March 2026, draw on the full observational record of the global Argo float network and offer the clearest physical explanation yet for a decline that has rewritten the satellite record in just a few years.
A freshwater lid, then a broken seal
Before 2016, increased snowfall and rainfall over the Southern Ocean freshened the surface layer around Antarctica. Fresh water is lighter than the salty water below it, so this created a buoyant cap that acted like insulation, trapping warmer, deeper water well below the ice. Sea ice grew slowly but steadily, even as the planet warmed.
Then the winds changed. Stronger and more persistent westerlies began driving upwelling, a process that pulls deep water toward the surface. That deep water is relatively warm. Once it breached the freshwater lid, it began melting ice from underneath, beyond the reach of satellites but well within the sensing range of ice-capable Argo floats drifting beneath the frozen surface.
“The ocean has been doing the heavy lifting,” the Stanford summary notes, describing how subsurface heat ventilation, not surface air temperatures alone, emerged as the dominant driver of the post-2016 decline. Wilson’s team paired float observations with idealized ocean models to test whether this two-phase mechanism, lid formation followed by lid destruction, could reproduce the observed ice loss. It could. The match between modeled and measured outcomes reinforced the conclusion that changes originating deep in the water column are steering what happens at the surface.
The numbers behind the collapse
The scale of the decline is stark. Antarctica’s summer sea-ice minimum in 2023 set an outright record low in the satellite record maintained by the National Snow and Ice Data Center, which stretches back to 1979. The February 2024 minimum tied for the second lowest. These are not isolated bad years. Every summer since 2016 has clustered near the bottom of the record, a sustained depression that has no precedent in the satellite era.
Separate research published in Nature examined what happened to the ocean and atmosphere during the 2023 record low. The collapse increased heat loss from the ocean to the atmosphere and intensified storm activity across the Southern Ocean. Those effects rippled outward, altering how the ocean and atmosphere exchange energy across the Southern Hemisphere. Antarctic ice loss, in other words, is not a problem confined to Antarctica. It reshapes circulation patterns that influence rainfall, storm tracks, and temperatures thousands of kilometers away.
What scientists still cannot answer
The PNAS study explains the mechanism, but important questions remain open as of May 2026.
Wind-driven upwelling and reduced precipitation both attack the same freshwater lid, and their relative contributions have not been fully separated. The balance likely varies by region. The Weddell Sea, with its own distinct wind and precipitation patterns, probably responds differently from the Ross Sea, and the study’s idealized models do not yet resolve those regional differences in fine detail.
Whether the current low-ice state is a permanent shift or a prolonged but reversible phase is also unclear. Antarctic sea ice is almost entirely seasonal, forming and melting each year, which makes it acutely sensitive to ocean conditions in ways that the multi-year ice pack of the Arctic is not. That sensitivity cuts both ways: it could allow a rapid recovery if wind patterns revert, or it could leave the system especially vulnerable to further warming.
Attribution remains contested, too. Wilson and colleagues argue that the mechanism they identified, stronger winds and changing precipitation disrupting ocean stratification, is consistent with a warming climate. But they acknowledge that the abruptness of the 2016 shift could partly reflect natural variability in the Southern Ocean superimposed on the long-term trend. Disentangling the two will require longer observational records and better climate models, many of which have historically struggled to reproduce Antarctic sea-ice behavior in either direction.
And the float network itself has limits. Argo instruments sample along sparse tracks at fixed intervals. They do not produce a continuous three-dimensional map of the ocean. Ice-capable floats, which can store months of data while trapped under the ice, are expensive to build and difficult to deploy. The Polar Argo pilot program aims to close that gap, but coverage remains thin in the regions where it matters most.
Why the deep ocean matters at the surface
For decades, the conventional focus in sea-ice science was on what happens at the surface: air temperature, snowfall, wind chill. The Argo record has shifted that focus downward. The Southern Ocean stores enormous quantities of heat at depth, carried there by global circulation patterns that take centuries to complete. When that heat finds a path to the surface, the consequences can be sudden and dramatic, even if the underlying warming has been building for a long time.
That is the core insight of the PNAS study, and it carries implications beyond sea ice. The same vertical mixing processes that melt ice from below also influence how much heat and carbon dioxide the Southern Ocean absorbs from the atmosphere. If upwelling intensifies and stratification weakens, the ocean’s capacity to buffer global warming could change in ways that current projections do not fully capture.
Multiple independent lines of evidence now point in the same direction. Satellites show the ice declining. Floats reveal warming and weakening stratification in the upper ocean. Atmospheric reanalyses document the wind shifts that favor upwelling. When datasets collected with different instruments, subject to different errors, converge on the same story, the overall picture gains weight.
New float deployments, expanded satellite missions, and higher-resolution models are all in the pipeline. They should sharpen the details. But the broad conclusion is already difficult to avoid: changes unfolding in the hidden depths of the Southern Ocean are reshaping the ice at its surface, and the consequences reach far beyond Antarctica’s frozen edge.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
