Danish model has a weakened AMOC regaining strength within decades of cooling
Danish Meteorological Institute researchers report that the AMOC starts strengthening within decades of global cooling in one climate model. Getting that cooling at 0.2C a decade would take carbon removal equal to current annual emissions.
The Scientist · Science desk
Carbon removal goes from 0.1 per cent of annual emissions today, around 2 million tonnes a year, to the 100 per cent of current annual emissions, 45 petagrams a year, needed to cool the world at 0.2C per decade.
| Measure | Value | As of | Claim |
|---|---|---|---|
| Removed per year today | 0.1% of emissions | Today | 9 |
| Needed to cool 0.2C a decade | 100% of emissions | Required | 10 |
What happened
- A collapse of the AMOC would bring "ice age" winters to Europe, higher sea levels in North America and disrupted monsoons in Africa and Asia, according to New Scientist.
- The simulation counted melting of the Greenland ice sheet as well as higher air temperatures.
- With warming at 0.2C per decade, slightly slower than the current rate, the AMOC weakened significantly over centuries but did not shut down entirely.
- The world now removes around 2 million tonnes of carbon a year, equal to 0.1 per cent of annual emissions, through direct air capture and enhanced weathering.
Why it matters
- contradiction The recovery result comes from one model that may overstate the AMOC's stability, yet a study published in July 2026 also found slowdown could be reversed by cutting atmospheric CO2.
- decision Cooling on this scale may force a choice on solar radiation management, an option Utrecht's Rene van Westen says might be necessary.
- exposure Europe is the region exposed to a restart: Van Westen warns of accelerated warming after a period of dramatic cooling.
Earlier tipping studies pushed the forcing one way. Chuncheng Guo of the Danish Meteorological Institute said "most of the previous studies are only looking at warming scenarios" [4]. In the cooling case, the current responded along the same line as in the warming case. "There's a very linear relationship between the global mean temperature and the AMOC response," Guo said [3].
In our view, that linearity is both the result and the open question. A response that moves in step with temperature leaves no room for the current to stay weak after the forcing reverses. A model with no sharp threshold would produce the same behaviour.
The same run also limits what "collapse" means here. Guo said the AMOC would not be totally extinguished even with up to 10C of warming [7], yet he framed the test as whether the current would "stay in that collapsed state" [4]. Because the current never fully stops in this model, the run cannot test recovery from a full shutdown [6][7]. The size of the weakening is not given.
The cooling side needs the same care with denominators. Cooling at 0.2C per decade would take 45 petagrams of carbon removal a year, which the account equates to 100 per cent of current annual emissions [10]. It puts today's removal at 0.1 per cent of emissions [9]. Those two figures imply a 1,000-fold increase [18]. But a petagram is a billion tonnes, and 45 billion tonnes set against the account's stated tonnage for current removal gives a 22,500-fold increase [17]. The two comparisons do not reconcile. Either way the scale-up is at least a thousandfold.
Guo was direct about what that means. "AMOC reversibility does not mean safety, because we really need a huge, huge amount of carbon dioxide removal to achieve that cooling rate," he said [11]. "That is not something we can achieve with today's technology," he said [11].
What to watch
- The German rerun of the same experiment across 10 climate models. It will show whether the linear response and the recovery hold outside one model.
- Any model run in which the AMOC stays weak after the forcing reverses. A result like that would break the near-linear picture.
- Whether the authors publish carbon removal figures that reconcile the tonnage and the percentages.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence40
- Adoption
- Insufficient
- Hype gap+10
- Incentives
- Insufficient
- Confidence45
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
A collapse in the AMOC would be catastrophic, bringing 'ice age' winters to Europe, raising sea levels in North America and disrupting monsoons in Africa and Asia.
- [2]
Guo and colleagues used a climate model to simulate the AMOC's strength as global temperatures rise with carbon emissions, and found the AMOC could weaken significantly but would recover just as quickly if global cooling occurs.
- [3]
"There's a very linear relationship between the global mean temperature and the AMOC response," said Chuncheng Guo at the Danish Meteorological Institute.
- [4]
Guo said "most of the previous studies are only looking at warming scenarios" and asked whether the AMOC would "stay in that collapsed state" if the forcing were reversed.
- [5]
The simulation includes melting of the Greenland ice sheet as a factor, as well as higher air temperatures.
- [6]
The team modelled a temperature increase of 0.2C per decade, slightly slower than the current rate of warming, and found the AMOC would weaken significantly over the coming centuries but would not shut down entirely.
- [7]
The AMOC would not be totally extinguished even with up to 10C of warming, Guo says.
- [8]
If the world reached net zero emissions and then started to cool, for example by removing huge amounts of carbon and storing it long term, the AMOC would start to strengthen within decades.
- [9]
The world currently removes around 2 million tonnes of carbon per year, equivalent to 0.1 per cent of humanity's annual emissions, via technologies such as direct air capture and enhanced weathering.
- [10]
Removal would need to increase to 45 petagrams per year, equivalent to 100 per cent of current total annual emissions, to deliver global cooling at a rate of 0.2C per decade.
- [11]
"AMOC reversibility does not mean safety, because we really need a huge, huge amount of carbon dioxide removal to achieve that cooling rate," Guo said. "That is not something we can achieve with today's technology."
- [12]
More radical cooling options, such as dimming the sun using solar radiation management, might be necessary to deliver cooling on this scale, says Rene van Westen at Utrecht University.
- [13]
Van Westen warns that restarting the AMOC after it has reached a collapsed state would plunge Europe's climate further into chaos, exposing it to accelerated warming after a period of dramatic cooling.
- [14]
The results are based on just one climate model, which might be overestimating the stability of the AMOC.
- [15]
The findings tally with a study published in July 2026, which also found that AMOC slowdown could be reversed by reducing levels of CO2 in the atmosphere.
- [16]
The findings are part of an international project to study tipping-point risks, and researchers in Germany are rerunning the same experiment using 10 different climate models.
- [17]
Required removal of 45 petagrams a year is about 22,500 times the stated current removal of 2 million tonnes a year.
- [18]
The source's percentages (100 per cent of emissions needed versus 0.1 per cent removed today) imply a 1,000-fold increase.
Get this ledger inside ChatGPT. Ask about any of our stories and the answer brings its figure, each claim’s verdict and the sources. Free, no account.
Sources
1 independent publisher whose own reporting we read for this story.
- newscientist.comWe can rescue the AMOC after its collapse – but it won’t be easy
1 article · October 14, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- AMOC tipping pointFollow
- Climate modellingFollow
- Carbon Dioxide RemovalFollow
- Solar geoengineeringFollow
Entities
- Atlantic Meridional Overturning CirculationFollow
- Danish Meteorological InstituteFollow
- Chuncheng GuoFollow
- Rene van WestenFollow
- Utrecht UniversityFollow
- New ScientistFollow