Warm water from the deep Southern Ocean has been steadily creeping closer to Antarctica’s ice shelves over the past four decades, according to a Cambridge-led study that combined ship-based measurements, autonomous float data, and machine-learning analysis. Lead author Joshua Lanham and colleagues reconstructed the poleward shift of Circumpolar Deep Water, a warm, salty water mass that sits hundreds of meters below the surface and can accelerate ice-shelf melting when it reaches the continental shelf. The finding sharpens concern that heat delivery to Antarctic ice may already be increasing in ways that existing climate projections have not fully captured.
Why the poleward march of warm water matters right now
Circumpolar Deep Water, often abbreviated CDW, circles Antarctica at intermediate depths and carries heat absorbed from lower latitudes. When CDW shifts closer to the continent, it can reach the grounding zones where ice shelves meet bedrock, thinning them from below. That thinning allows glaciers on land to flow faster toward the sea, raising global sea levels. The new study, published in the journal Communications Earth & Environment, documents a clear poleward redistribution of CDW around the continent, not just in isolated sectors but as a circumpolar pattern.
Separate research published in Nature has already shown that Antarctic meltwater is slowing abyssal ocean overturning and warming deep layers. That modeling work, which explored how freshwater from ice-sheet melt alters dense water formation, found that increased melt can weaken the deep circulation that ventilates the abyssal ocean and helps regulate global heat storage. If CDW continues to push southward while deep circulation weakens, a feedback loop could form: more melt produces more freshwater, which further disrupts circulation, which in turn allows still more warm water to reach the ice. The two lines of evidence, one observational and one modeling-based, now point in the same direction, and that convergence is what gives the latest findings their urgency.
Ship transects, Argo floats, and machine learning built the record
The reconstruction rests on two distinct data streams stitched together by a machine-learning classifier. The first stream comes from decades of repeat hydrographic ship transects collected under the GO-SHIP and CLIVAR programs, whose cruise-level temperature, salinity, and nutrient profiles are distributed through NOAA’s repeat-section archive. These ship sections provide high-vertical-resolution snapshots along fixed lines but are spaced years apart and cover limited geographic tracks, especially in winter and in regions with heavy sea ice.
The second stream comes from the global Argo array of autonomous profiling floats, which has supplied sustained upper-ocean temperature and salinity observations since 2004. Argo floats deliver broad spatial coverage on monthly timescales but typically do not reach the seafloor and are sparse close to the continent where sea ice persists. According to the University of Cambridge, the research team applied a machine-learning classification to a monthly gridded Argo climatology starting in 2004 and then extended the record backward using the older ship data to span roughly four decades.
That hybrid approach introduces a tension the authors had to manage. The Argo-era portion of the record offers dense, regular sampling, while the pre-2004 portion relies on sparser ship transects with irregular timing. The study’s strength is that it bridges both eras into a single consistent framework, but readers should recognize that the earlier decades carry wider observational gaps than the later ones. The accepted manuscript, available through the University of Cambridge repository, includes the central quantified finding with a confidence interval describing the poleward redistribution rate and discusses how sampling uncertainties vary by region.
What the CDW shift could mean for Antarctic ice loss
The practical consequence is direct. Ice shelves act as buttresses, holding back the vast ice sheets on the Antarctic continent. When warm CDW reaches the cavities beneath those shelves, it erodes them. The West Antarctic Ice Sheet is especially vulnerable because much of its base sits below sea level, meaning warm water can penetrate deep inland along reverse-sloping bedrock. If CDW access to those cavities increases over the coming decades at the rate the new study documents, the pace of ice-shelf thinning and glacier acceleration could outstrip current projections that assume a more static ocean circulation.
The Nature study on abyssal overturning slowdown provides a mechanistic explanation for why CDW might keep pushing south. As Antarctic meltwater freshens the surface ocean near the continent, it disrupts the formation of cold, dense bottom water that normally sinks and flows northward. With less bottom water pushing outward, the deep circulation weakens, and the thermal barrier that once kept warm CDW at a distance erodes. Joshua Lanham’s observational record now supplies field evidence that this theoretical pathway is already playing out in the real ocean, linking surface freshening, deep circulation changes, and the lateral migration of warm water toward the ice margins.
The implications for sea-level rise are global. Even modest increases in basal melt rates can destabilize ice shelves that are already thinning from the atmosphere above. Once a shelf loses enough thickness, it can fracture or collapse, as seen in previous disintegration events along the Antarctic Peninsula. In West Antarctica, where key outlet glaciers rest on retrograde beds, sustained erosion at the grounding line could commit the region to long-term retreat, adding meters of potential sea-level rise over centuries. The new evidence that CDW is systematically moving closer to the continent suggests that ocean-driven contributions to this process may be intensifying.
Gaps in the evidence and what to watch next
Several questions remain open. The study does not isolate a single cause for the poleward CDW shift. Wind-driven changes in the Antarctic Circumpolar Current, shifts in large-scale atmospheric patterns, and meltwater-driven circulation changes could all contribute, and their relative importance likely varies by region and decade. The machine-learning classifier, while effective at identifying water-mass boundaries, still relies on the underlying observations; where data are sparse, especially before the Argo era and close to the ice edge, the inferred trends carry larger uncertainty.
Another gap lies beneath the ice shelves themselves. Most of the cavities where CDW would actually contact ice remain poorly observed, accessible only through occasional ship-based surveys, moorings, or instrumented seals. The Lanham study tracks where CDW sits in the open Southern Ocean and on the continental slope, but translating that shift into precise melt rates under individual shelves will require more direct measurements and high-resolution modeling that can resolve narrow troughs and sills on the seafloor.
Future work is likely to focus on three fronts. First, expanding autonomous observing systems into the ice-covered coastal zone, including ice-capable floats and gliders, could capture how CDW actually crosses the continental shelf break. Second, coupled ocean–ice-sheet models will need to incorporate the observed poleward migration of warm water and test how sensitive different Antarctic basins are to further shifts. Third, researchers will be watching for continued changes in abyssal overturning, as described in the Nature analysis of Antarctic-driven circulation slowdown, to see whether the feedback loop between meltwater, circulation, and CDW access strengthens.
For now, the message from the Southern Ocean is that the conditions controlling Antarctic ice loss are not standing still. Warm water that once stayed farther offshore is edging closer to the continent, and the deep circulation patterns that helped insulate the ice are showing signs of strain. How quickly policymakers and coastal planners respond to this evolving picture may help determine how societies cope with the sea-level changes that follow.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
