Southern Ocean Could Shift from Carbon Sink to Source

Estimated reading time: 5 minutes · Last updated: 11 September 2026

The Southern Ocean, which currently removes roughly 10% of human CO2 emissions, may flip from sink to source as atmospheric CO2 declines after net zero, the modelling indicates. The Science paper (DOI: 10.1126/sciadv.aee9228) examines scenarios where atmospheric CO2 falls from 700 ppm and projects the Southern Ocean moving from taking up around 3 grams of carbon per square metre each year today to releasing 8–9 grams per square metre annually by 2400. The paper warns this sink-to-source reversal would slow the fall in atmospheric CO2 and could lock in additional warming for centuries even if society deploys large-scale carbon dioxide removal.

That impact can sustain much, much longer than we expected,

Yechul Shin, Seoul National University

How the Southern Ocean stores carbon today

The Southern Ocean sits under the strong westerly winds known as the roaring forties and furious fifties. These winds agitate the surface, increase gas exchange and drive surface waters into cold Antarctic air, where they cool, sink and carry dissolved CO2 into the deep ocean. That vertical transfer stores carbon away from the atmosphere for centuries to millennia, making the region the single biggest oceanic carbon sink.

Physical circulation and biology both matter. Plankton build alkaline calcium carbonate shells; when they die the shells sink and help move alkalinity and carbon to depth. Ocean circulation then isolates that deep water, so carbon does not readily return to the surface on human timescales. That mechanism is why the Southern Ocean currently accounts for roughly 10 percent of anthropogenic CO2 uptake.

What the Science modelling experiment tested and found

A paper published in Science modelled scenarios in which atmospheric CO2 begins to fall from 700 parts per million after humanity reaches net zero. One scenario attains net zero in 2125; the paper also includes a run in which net zero is reached in 2025 followed by large-scale carbon dioxide removal. In both simulations the Southern Ocean shifts from acting as a net sink to becoming a net source of carbon.

Quantitatively, the models project a change from a present-day uptake close to 3 grams of carbon per square metre annually to emissions of 8–9 grams per square metre by the year 2400. Atmospheric CO2 declines in these scenarios but then levels off after about 2200 because carbon returning from the Southern Ocean counteracts further drops. In the carbon-removal scenario the simulated atmospheric CO2 even rises again for roughly 50 years before stabilising at 425 ppm; without active removal it stabilises at 590 ppm.

Two feedbacks that amplify the Southern Ocean’s release

The paper identifies at least two physical feedbacks that strengthen the release of stored carbon as the ocean adjusts. First, heat stored alongside carbon at depth slowly returns to the surface; warmer surface waters hold less dissolved gas, so a warming return flow forces more CO2 out of the ocean and into the atmosphere. Coverage quotes Nicolas Gruber at ETH Zurich using a piston analogy: if you pull the piston back the CO2 comes back out.

Second, changes in mixing and alkalinity matter. Planktonic calcium carbonate production transfers alkalinity to the deep ocean. If surface waters warm and become less dense, they mix less with alkaline deep water and less alkalinity returns to the surface. Reduced surface alkalinity lowers the water’s capacity to chemically absorb CO2, reinforcing a shift toward net outgassing as the system relaxes.

Implications for carbon removal and climate policy

The modelling shows that removing CO2 from the atmosphere is not a simple undo button for past emissions. Alberto Naveira Garabato at the University of Southampton warns that any change in ocean–atmosphere exchange is tantamount to adding a very large element to the global carbon budget, and the study finds the ocean as a whole may become a net source as stored carbon returns to the air.

That outcome raises a practical problem for strategies that rely heavily on future carbon dioxide removal. Even aggressive removal in the model leaves atmospheric CO2 near 425 ppm after a period of rebound, rather than pushing concentrations below present-day levels. The sooner emissions are halted and the Southern Ocean stops heating, the smaller the eventual outgassing will be; delayed net zero and continued warming make the sink-to-source transition more damaging and longer lasting.

Case for and against the study’s conclusions

The bottom line

The study in Science and commentary from ocean scientists underline a central difficulty: the ocean’s physical and chemical memory can oppose human attempts to lower atmospheric CO2. Heat and alkalinity already stored at depth will act over centuries to release carbon as surface concentrations fall, so delaying net zero increases long-term committed warming. The practical takeaway for policy is clear — faster emission cuts reduce the magnitude of this committed outgassing, and monitoring plus improved process understanding of the Southern Ocean should be priorities to narrow uncertainty about how large and how fast the sink-to-source transition will be.

What to watch

• Watch for follow-up observational studies of Southern Ocean air–sea CO2 fluxes; no date has been set.
• Watch for the next IPCC assessment to incorporate these ocean feedbacks into carbon-budget estimates; no date has been set.