Northern ice sheets may have helped drive ice ages by cooling the Southern Ocean, model finds

That single number is the core of a proposal about how ice ages start. The researchers suggest that spreading Northern Hemisphere ice sheets chill North Atlantic water. That water travels south and changes circulation around Antarctica. The Southern Ocean then takes up and stores more carbon dioxide.

ScienceDaily reported the work on 28 October, citing the University of St Andrews. The paper appeared in the Journal of Climate. The question it addresses is old and awkward. The Southern Ocean is known to have pulled down atmospheric CO2 during glacial periods. But the trigger for that drawdown has been hard to tie to the north. ScienceDaily's summary calls this a possible explanation. The study offers a candidate mechanism for a link that researchers had struggled to explain: how processes at opposite ends of the planet could be so closely connected.

What the simulation actually did

The team did not lower temperatures everywhere. It cooled only the northernmost latitudes, mimicking a large ice sheet, and then watched how the Southern Ocean responded. According to the ScienceDaily account, the result was 74% of the CO2 drawdown seen in a separate run that cooled the whole planet.

That is a modelling result, not a measurement from ice cores or sediment. The study used computer simulations of ocean circulation. The mechanism it proposes runs through water movement between hemispheres, not through anything measured directly in the modern ocean.

The chain the authors describe has several links. Northern ice sheets grow in response to natural variations in Earth's orbit. They cool nearby North Atlantic waters. Those colder waters circulate southward over time and eventually reach the Southern Ocean. If the cooling is strong enough, it alters circulation around Antarctica and lets more CO2 be stored beneath the surface. Lower atmospheric CO2 then cools the planet further, which could intensify an ice age.

Why the Southern Ocean carries the argument

The Southern Ocean does most of the work in the story. It is thought to control a large share of the ocean's uptake of atmospheric carbon. ScienceDaily states that changes in the Southern Ocean surrounding Antarctica controlled much of the process behind past CO2 reductions. The unresolved part was the link to the north.

The researchers propose that the north sends a cold signal south through the ocean, not through the atmosphere. That distinction matters for how to read the result. It is a statement about ocean circulation in a model, and the strength of the effect depends on how faithfully that circulation is represented. The study is also not the only recent work pointing at polar ocean carbon. Other headlines include work on Antarctic sea ice loss and ocean salinity, on glacier-driven tsunamis and Antarctic ocean mixing, and on whether the Southern Ocean is a safe place for carbon removal. The St Andrews paper sits in that broader effort to pin down how the polar oceans move carbon.

What the number does and does not say

Seventy-four per cent is a comparison between two model runs. It is not a measured share of any real glacial cycle. It says that in this model, northern cooling alone gets most of the way toward the CO2 drawdown produced by global cooling. It does not say that northern ice sheets caused any particular ice age. It also does not attach a timescale to the southward transit of cold water.

ScienceDaily relays the authors' framing: the simulations provide strong support for the proposed connection. That is a claim about model behaviour. Confirmation would need to come from elsewhere, and the report does not present one.

One contrast helps readers judge how such results are presented. A separate study on lung cancer prediction, covered by HPCwire on 2 November, reports accuracies of 94.31% in simulation and 91.90% on quantum hardware. It states plainly that its extra samples were synthetic and did not represent newly recruited patients. Model and simulation results carry caveats, and the strongest reporting states them.

Where this fits in the wider research picture

The freshest items elsewhere are mostly unrelated: Google's Gemini 4 Carbon model, IBM Research's meter-scale quantum interconnects between two 120-qubit processors, and an arXiv paper on entanglement-assisted quantum codes. None of them bears on ocean carbon.

That is worth saying plainly. The St Andrews study is not breaking news in the sense of a product launch. It is a Journal of Climate paper whose public write-up appeared on 28 October. Its importance depends on whether other groups can reproduce the mechanism in their own models. For context on the scale of the problem, the Guardian reported on 27 October on international cancer research, a reminder that the news spans very different fields. On climate, the relevant background is the long-standing effort to explain glacial cycles through CO2 storage in the Southern Ocean. The St Andrews work tries to connect that to northern ice growth.

What to watch

The obvious next steps are replication and testing against paleoclimate records. If the mechanism holds, it should leave fingerprints in proxies for Southern Ocean circulation and sea ice when northern ice sheets expanded. The ScienceDaily report does not describe such a test, so the claim currently rests on the simulations.

Readers should also note what the study does not claim. It does not put a figure on how much CO2 the Southern Ocean stored during any past ice age. It does not forecast anything about future climate. Its subject is the past.

The simulations provide strong support for the proposed connection between expanding northern ice sheets and changes in the Southern Ocean, according to ScienceDaily's account of the study.

For a field that has spent years trying to explain why the two poles seem to act together during glacial cycles, a model result that links them through ocean circulation is a hypothesis worth testing. It is not yet a settled answer. The 74% figure should be read as a property of the simulation, not of the Earth.