"Antarctic Intermediate Water Reversal: Sea-Ice Loss Is Rewriting a Southern Ocean Layer"

A vast layer of the Southern Ocean has changed direction. Antarctic Intermediate Water, the water mass that carries heat, carbon and oxygen from near the surface down into the ocean interior and then spreads northward across the Atlantic, Indian and Pacific oceans, spent two decades growing fresher. Then, in less than a decade, it reversed: it is now becoming saltier, denser and deeper. The finding, published on 28 October 2026 in Nature Climate Change, is the first direct observational evidence of the switch, and it arrives with a warning that our models of deep-ocean warming and sea-level rise may be missing something.

The driver is Antarctic sea ice, or rather its absence. From 1995 to 2015, the seasonal summer melt around Antarctica fed a steady supply of freshwater into the ocean, keeping this intermediate layer relatively fresh. After 2016, Antarctic sea ice fell to record lows, and that freshwater engine sputtered. Lead researcher Bishakhdatta Gayen, of the University of Melbourne and the Indian Institute of Science, puts the scale in terms that are hard to ignore: until 2016, the summer melt delivered roughly three times the Amazon River's entire annual flow, about 6,600 cubic kilometres of water a year. After 2016, it dropped to roughly twice that flow. A third of the engine's output, gone.

The team, led by Gayen with PhD student Kaushik Mishra and Alberto Naveira Garabato of the University of Southampton, traced the change through the global Argo network of around 4,000 robotic floats, combined with satellite measurements and reanalysis data. What makes the result striking is not just the speed but the precedent. Gayen notes that the last time this water mass switched in this way was at the end of the last interglacial period, around 130,000 years ago, when Antarctica held very little ice. That transition took thousands of years. This one took under a decade.

Why should anyone who does not study oceanography care? Because Antarctic Intermediate Water is part of the machinery that keeps the planet's climate stable. The oceans absorb more than 90% of the excess heat released by human activity, and this layer is one of the routes by which surface water, and the heat and carbon it carries, is drawn downward. If a weakened, saltier layer changes how that transport works, two things follow. First, more excess heat could be pushed deeper, meaning existing models may be underestimating deep-ocean warming and therefore miscalculating regional sea-level rise. Second, the altered water mass travels north across three ocean basins and feeds, over years to decades, into the Atlantic overturning circulation, the system linked to climate stability in the North Atlantic. The authors are careful about what is not yet known: the magnitude and timing of these knock-on effects are uncertain, and the study did not measure environmental impacts directly. But the direction of travel is now documented, and it is the wrong direction for comfort.

There is a useful discipline in this story for anyone who follows climate news. We tend to watch the dramatic quantities: ice extent, temperature records, storm counts. This paper is about a quieter variable, the saltiness of a layer of water most people will never see, measured by robots drifting in the dark. It changed first, and it changed fast. The ocean does not always announce its turning points with a spectacle. Sometimes it just gets saltier, a little at a time, until the trend line bends and does not bend back.

The honest question this leaves is whether the reversal is a blip or a new regime. Sea ice can recover in a good year, and a single study covering 2016 to 2024 cannot settle what the next decade holds. But the burden of proof has shifted. For twenty years the freshening trend was the background condition; now the burden is on anyone who claims the old pattern will simply resume. I am watching the Argo data the way you would watch a patient after a first warning sign: not in panic, but with the understanding that the first sign is rarely the last. My prediction, recorded alongside this article, is that the saltening will still be there when the next round of measurements comes in.