AMOC weakening declared a national security threat
Iceland has declared the AMOC a national security threat and an existential risk in the coming decades, enabling its Government to prepare for worst-case scenarios. Should we be worried?
What is the AMOC?
I’m writing this as I sit in my home office at almost 52° North on the Welsh borders of the UK. It’s a similar latitude to Calgary, or Edmonton in Canada, and Irkutsk in Siberia. The climate is very different here in the UK though. For example last week I cut the grass wearing a tee-shirt. Meanwhile it’s -11°C in Calgary and -9°C in Irkutsk. What makes it so mild in northern Europe for it’s latitude is all thanks to the Atlantic Meridional Overturning Circulation or AMOC.1
Back in 1751 a sea captain called Henry Ellis discovered that the deep ocean water in the tropics was cold, very cold. It had been assumed that the sun would warm the water evenly so that over time, even the deepest waters would be warm. In 1797 Count Rumford wrote a supposition that these cold deep waters could only be coming from the polar oceans. 200 years on, we now understand this mechanism as part of a global inter-ocean system of currents that move heat about the planet.
The Spilhaus projection map of the world above is an easier way to visualise the oceans as a single interconnected system. The great currents carry warm surface waters and deep cold waters in a constant circulation, transporting heat, salt, nutrients and carbon around the planet.
The AMOC is one of these interconnected currents but is unique in one important regard. Most currents transport heat away from the tropics to higher latitudes. The AMOC carries warm salty water from the Southern hemisphere up and through the tropics then on to the north Atlantic and Arctic. The warm salty water is constrained to the top 1,000 meters of the ocean. As it moves through the tropics it becomes more saline as the warm conditions evaporate water from the surface. When it reaches the north it is therefore both warm and salty, both aspects of which are important for what happens next.
The amount of heat transported to the north Atlantic is huge. It’s over a petawatt or 1015 Watts. That’s about 50 times more than all the energy used by humans today. It’s enough to warm the entire region by 4.5°C on average which means that I’m cutting the grass in November rather than shovelling snow. It also means the northern hemisphere is about 1.4°C warmer than the southern hemisphere and why the thermal equator is at 10°N which positions the tropical rain belts further north than they would otherwise be.
So that covers the Atlantic, Meridional (south and north) and Circulation aspects of AMOC, what about the Overturning bit? The engine that drives the current is what happens when the warm salty water reaches the north Atlantic. The winds strip the heat from the surface waters, cooling them and taking the heat downwind to warm northern Europe. As the water cools it becomes denser than the waters beneath it since it is now both cold and salty. Being denser, it then sinks to the abyss down to between 2,000 and 3,000 meters deep through a process of convection. From there the deep water current flows south, down the entire length of the Atlantic.
The image above shows that the overturning doesn’t happen in just one place, there are various sites and gyres which together make up the AMOC system. Incidentally the Gulf Stream, sometimes confused with the AMOC, is a wind driven current that passes Florida and flows with the AMOC and contributes to the water flow but not much to the heat transport. It’s part of the system but not the whole picture since the returning surface water that flows back from the East is of a similar temperature, meaning not much heat has been lost.
Sea ice formation in the Arctic also contributes to deep water formation. When salt water freezes, the salt is expelled as the ice consists of just freshwater. The cold highly saline and therefore dense brine sinks to the abyss to join the southern flow.
This NASA video is an excellent animation of the Ocean currents and worth watching by clicking the link in the caption.
How could it be weakened?
Back in 1961, a US oceanographer Henry Stommel, realised how the salinity controls the stability of the AMOC and discovered it had two stable states and a tipping point between them. The AMOC flows because it is salty and its salty because it flows. This is a self-sustaining feedback. The problem is it works both ways. If the water becomes less salty, the flow reduces, which brings less salt in a self-reinforcing negative feedback.
The AMOC is therefore a classic bistable system. It can be on or off, but the region in-between is unstable. It also exhibits hysteresis since in either of the two stable states, it is hard to get it to flip. Just like a chair that has fallen over, the last little nudge to tip it over is far less than the effort to right it again. Paleo-evidence from previous shut-downs shows that full collapse may take 50-100 years, but once off it stays off for thousands of years. Switching back on however can be more rapid once the right conditions are reached.
So what could weaken the AMOC and lead to it tipping to the off, or collapsed state? If the surface waters become fresher, this reduces their density so they remain on the surface which prevents the warm waters underneath from losing their heat and becoming more dense. This happens in two ways. Heavy rainfall can create a fresher surface layer. As the climate warms, more water can be held in the air and released as rainstorms in the cooler air of the north Atlantic. This is sufficient to temporarily halt local convection in the water column, stopping local deep water formation. Increased rainfall over Arctic land masses also leads to increased river runoff, also providing freshwater to the ocean.
Greenland ice melt is the major source of ocean freshening however. Global warming is driving accelerated melt of Greenland’s ice sheet. On average 200 Gigatonnes of ice mass are lost each year - that’s 22 million tonnes of freshwater running off the continent into the ocean every hour! Since it is freshwater, it is less dense and flows over the salty current from the south inhibiting heat loss, convection and deep water formation, slowing the AMOC.
Sea ice formation is also declining in the Arctic. 2025 saw the lowest sea ice maximum area on record, showing less ice was formed last winter than ever recorded before. Ice volumes are also declining. This means less salty brine is expelled to join the current south. Reduced deep southern flow means less ‘pull’ in the system which can also reduce the convection and surface flow northwards.
The freshening acts to slow down the AMOC circulation, which in turn brings less salty water to the region, further weakening the system in the familiar self-reinforcing feedback loop. Eventually the tipping point of no return is reached where collapse is inevitable and unavoidable. The trick is telling where on the curve the AMOC sits today, how far is it away from tipping, and what happens if it does?
How could we tell?
Since elements of the system are weather driven, they have a huge amount of natural variability. Rain in the North Atlantic is not a constant, melt water and river run-off from Greenland and the Arctic are seasonal and there are decadal patterns in ocean circulations too.
Away from the tipping point, in either the on or off state, the system is stable, meaning it can respond to local changes and recover without approaching any critical limits. All these parameters make direct measurements prone to high levels of noise, but they are being attempted.
The RAPID2 Array consists of two lines of sensors tethered to moorings in the East and West Atlantic at 26°N. When combined with other measurements of flow at 26°N, namely the Gulf Stream through the Florida Straits and surface wind-driven flow (Ekman transport), they provided an estimate of the strength of the current every hour. It was established in 2004 and has been providing continuous measurements of deep water flow since then. Because the flow is measured at 26°N it is referred to as simply MOC. The research has shown that the MOC varies on timescales of days to a decade. This is useful since long term trends are now starting to emerge. Wind-forcing is playing the dominant role in the high variability, which makes the signal very noisy. Coherence between the northward and southward flows is also more complex than expected.
Despite the large variability discovered on short time scales, statistical analysis shows that the array revealed a major downturn of the MOC in 2010–11 that marked the beginning of a sustained decline since that time. This is therefore direct observational data that shows the AMOC is weakening. It’s a bit noisy though and easy to argue against, so what other evidence is there?
Before the RAPID array detected the signal from the noise, a fingerprint was detected in the Atlantic. A “Cold Blob” has appeared which is the only part of the globe that has not warmed through global warming. Related to the cold blob, a warming front now exists along the northwest Atlantic coast as the Gulf Stream is pushing in that direction. Both of these indicate a slowing down of the northbound surface current and AMOC.
Model simulations of the ocean where the AMOC has been forced to slow down show these fingerprints are expected with a weakening AMOC. Models create these fingerprints both through being freshened and when running elevated atmospheric CO2 concentrations. They match the observations very closely. This is another key piece of evidence that a long term slowdown has been underway for over 50 years.
A new paper published last month has identified another compelling, if not conclusive, indicator. Qiuping Ren at el.3 have identified a mid-depth warming signal in the equatorial Atlantic between 1,000 and 2,000 meters deep. Perturbations in the AMOC trigger oceanic Kelvin waves that propagate rapidly (within months) along the western boundary and equator, making the equatorial Atlantic a key crossroads for the dynamical signals of AMOC change.
Due to the depth, the signal is less variable than sea surface signals such as the cold blob. Observations reveal a robust mid-depth warming since 1960 that emerged from natural variability in the early 2000s with 99.9% certainty, suggesting a slowdown started in the late 20th century and is gaining momentum. Observational data is available for far longer than the RAPID array through the WOA hydrographic dataset and, more recently the Argo float project which provides a more basin wide picture. The WOA shows mid-depth warming in the Atlantic that is eight times faster than the Pacific, and aligns well with ocean computer models of AMOC slowdown.
When taken together, the evidence shows that observations are matching the best models. Those models show that even with intermediate greenhouse gas emissions, there is now a more than 50% chance of the tipping point being reached this century and that deep mixing through convection is on track to stop between 2030 and 2050.
The chart above from the 2025 State of the Cryosphere Report4 overlays the observations in cyan over the model outputs, indicating with very little doubt that the AMOC is in serious decline and we, therefore could be in serious trouble.
What effects can be expected?
The most recent modelling work suggest deep convection collapse between 2030 and 2050 with the tipping point being crossed between 2055 and 2070.
Not surprisingly a lot of work is focussing on what the implications of AMOC weakening and collapse are. One important thing to note though is that it is a transitional process. Although it will take 50-100 years to unleash the full effects, every fraction of decline brings less heat to the region meaning a steady progression of impacts to the end point, rather than a step change in the far future.
Temperature
Temperature change is the most obvious implication. René M. van Westen’s group at Utrecht University is doing a lot of work in this area with the support of European and Dutch Government research funding. They have been looking at the full effect change with and without climate change, since continued global warming will temper the changes to some extent.
The plot above from van Weston and Baatsen5 shows the change in January temperatures expected if the moderate emissions scenario RCP4.5 is followed and AMOC has collapsed. The shading shows the average January temperature difference from pre-industrial, while the coloured contour lines trace areas that could experience 1 in 10 year minimum temperature extremes. Scandinavia and Eastern Europe experience extremes 40 to 50°C cooler. The UK would be 10 to 20°C cooler on average and experience extremes of 20 to 30°C cooler.
Note that central Europe and northern Italy show little average change but exposure to extreme events 10 to 20°C colder than before. Iceland drops by 20°C on average with extremes of -50°C, explaining their alarm.
These strong temperature gradients will drive increasing jet stream strength driving very severe and damaging winter storms across the whole continent.
Summer temperatures however will continue to rise along with emissions, at least for mainland Europe and the UK. The cold water influences the air pressure distribution in a way that draws warmer air from the south across Europe. These extreme season swings will result in a more Canadian climate, increasingly hot in the summer, extremely cold in the winter.
The cooler winters are driven by an expansion of winter sea ice pack as shown in the chart below. Very low emissions coupled with AMOC collapse show lower latitude sea-ice and drive even colder temperatures, but RCP 4.5 is a good approximation of current policy trajectory.
Rainfall
Rainfall patterns will also change dramatically. In general, drier hydro-climatic conditions are expected under an AMOC collapse since less evaporation will occur across the cooler ocean surface. The chart below from a van Westen pre-print paper6 shows the change expected for a collapsed AMOC with RCP4.5 emissions scenario. Overall a 20% drop in annual rainfall can be expected. This will be significant for agriculture as well as water availability for human consumption, energy, industry and nature.
Sea level rise
Sea level is not the same around the globe. There are high parts and low parts in different regions caused by gravitational variations, the background ocean circulation and water density. For example the Atlantic side of the Panama Canal is 20cm lower than the Pacific side.
The present flow of water in the Atlantic and the coupled Coriolis affect lower sea levels compared to the global mean along the American northeast coast and in the north Atlantic and Arctic oceans. With a collapsed AMOC, sea levels could rise by up to a meter on the American coast through this effect alone. 0.5 to 1 meter rises are also projected for the north European coastline which would be extremely problematic for low lying countries such as Belgium and the Netherlands and many coastal cities including London. This would be on top of global sea level rise from global warming and ice loss.
Steady AMOC weakening will continue to increase the frequency of flooding events. In a paper published this year7, the link between AMOC variability and tidal gauge measurements on the US shoreline was demonstrated. Periods of weaker AMOC link to higher tidal gauge readings.
The rate of change is an important consideration from an adaptation perspective. Changes due to water temperature (steric change) also need to be taken into account. In another pre-print, van Weston et al.8 identified a 6.6mm/yr rise in sea level in the north Atlantic which is twice the current rate through climate change alone.
For the North Sea region including the UK east coast, Belgium to north Denmark and southern Scandinavia, current rates of sea level rise would double putting huge pressure on coastal protection to prevent flooding and erosion.
Global implications
In addition to the effects on European weather system and Atlantic basin sea levels, since the AMOC is part of the global ocean system, the implications are also felt globally.
As described earlier the warming that the AMOC brings to the northern hemisphere shifts the thermal equator of the globe 10° north of the equator. The position of this line determines the position of the tropical rain belts and the Hadley circulation cells. Without the AMOC, the southern hemisphere warms more rapidly as heat is not passing through to the north. This lowers the thermal equator and takes the rain belts with them. Areas used to significant rainfall such as the northern part of the Amazon and southern Asia would see significant less rain fall affecting the crop growing potential for billions of people. Currently arid areas would experience significantly more rainfall leading to flooding.
This increased southern hemisphere warming would also accelerate Antarctic ice sheet melt and sea ice reduction. This sea ice reduction would more than match the increase in Atlantic sea ice area so cancelling out any increased albedo effect in the north.
It’s also worth noting that although warming reduces in the northern hemisphere when the AMOC collapses, this does not effect the trend of global warming through continued greenhouse gas emissions. The AMOC is a distributer of heat not a generator. If we continue emitting, even after an AMOC collapse, the planet as a whole continues to warm.
Models also suggest that AMOC collapse increases the Earth’s Energy Imbalance further with most of the additional heat entering the ocean. Steric sea level rise would increase globally by a further 20cm as a result. All in all an weakening AMOC is bad news.
Is there anything we can do?
Prior to passing the tipping point, emissions reduction is the only way of delaying or potentially preventing AMOC collapse. The latest data presented at the recent Nordic Tipping Week workshop suggests a probability of passing the tipping point under an intermediate emissions scenario this century is 57%. High emission scenarios carry a 70% risk. Lower ‘Paris aligned’ scenarios drop his chance to 25%.
There is no longer a choice available to us that drops the chance below 10%, but the faster we can decarbonise and then ramp up negative emissions, or carbon capture, the better the odds become. Given the severity of the risk and the, at best 25% chance of it happening, this has to be treated extremely seriously. Any sensible precautionary principle approach would demand rapid urgent action to reduce emissions immediately.
Adaptation and preparation also now becomes a priority. Last year the Nordic Council was handed an open letter by climate scientists warning of the serious implications of AMOC weakening and urging urgent action9. Now the Icelandic Government have announced they are treating the issue as a national security threat10.
The adaptation challenge is monumental. If we just look at the UK, where I am, none of the current housing for 70 million people is suitable for a post AMOC Collapse climate. Insulation and heating for winter temperatures of -20°C would be required, coupled with cooling for increasingly warmer summers. Water management struggles currently so with droughts projected to increase through a further 20% reduction in rainfall would be extremely challenging and take decades to prepare for, involving huge infrastructure projects. Much of the water supply and waste treatment facilities would freeze in -20°C winters so will need to be re-engineered. Transportation is already a joke with the slightest snowfall bringing the country to a complete halt. Road surfaces, rail infrastructure even airport runways would need almost complete replacement. Agriculture as it is today would effectively cease, driving increased reliance on imports, but set against global rain belt shifts, that may not even be possible. Sea defence schemes would probably be unaffordable, leading to high inundation and retreat along the coastlines and inland through Lincolnshire and Cambridgeshire.
In short the UK would not remain a viable habitable state for even half the current population. Coastline retreat would also affect Belgium, the Netherlands and parts of Germany and Denmark. Coastal cities of the western US would also face huge challenges through increased rates of sea level rise.
Wildlife both flora and fauna would also be severely effected. The climate would not be suitable for the majority of the current woodlands, hedgerows or fruit orchards. Migration patterns would be affected and many animal species would simply die out. Looking out of my office window, I can see only a few species of tree that would survive and none of the current native birds or mammals. A new national forest is being planned not far from here, are they planning for what might survive in 50-100 years and planting accordingly? I bet they're not.
There is one controversial measure that could be used however. The Geoengineering technique of Stratospheric Aerosol Injection (SAI) is being discussed as a means of preventing further warming, once decarbonisation and carbon capture is underway. The technique involves injecting sulphate aerosols into the stratosphere which reflect some of the incoming solar radiation, lowering the heat accumulation. This could be used to cool the planet11.
A study published this year12 modelled different SAI scenarios to determine if it could be applied to slow AMOC weakening. They found that depending on the latitude of injection, SAI could be used to prevent further decline. The long term outcomes are more complex and unsure and there are a host of legal, moral and regulatory issues that would need to be addressed in addition to the technical challenges. However if, as seems likely, the world cannot end its addiction to fossil fuels, the risk-risk proposition of geoengineering versus AMOC collapse may need to be discussed. It may be the only hope.
Rahmstorf, S. 2024. Is the Atlantic overturning circulation approaching a tipping point? Oceanography 37(3):16–29, https://doi.org/10.5670/oceanog.2024.501.
RAPID Project, https://rapid.ac.uk/
Ren, Q., Xie, SP., Peng, Q. et al. Equatorial Atlantic mid-depth warming indicates Atlantic meridional overturning circulation slowdown. Commun Earth Environ 6, 819 (2025). https://doi.org/10.1038/s43247-025-02793-1
René M. van Westen, Michiel L. J. Baatsen, European Temperature Extremes Under Different AMOC Scenarios in the Community Earth System Model, 2025, Geographical Research Letters, https://doi.org/10.1029/2025GL114611
René M. van Westenet al. Changing European Hydroclimate under a Collapsed AMOC in the Community Earth System Model, Pre-print https://doi.org/10.5194/egusphere-2025-1440
Liping Zhang et al. ,Skillful multiyear prediction of flood frequency along the US Northeast Coast using a high-resolution modeling system.Sci. Adv.11,eads4419(2025).DOI:10.1126/sciadv.ads4419
René M. van Westen et al. Dynamic and Steric Sea-level Changes due to a Collapsing AMOC in the Community Earth System Model, Pre-print https://egusphere.copernicus.org/preprints/2025/egusphere-2025-5102/
Open Letter by Climate Scientists to the Nordic Council of Ministers, Reykjavik, October 2024, https://en.vedur.is/media/ads_in_header/AMOC-letter_Final.pdf
Reuters, Iceland deems possible Atlantic current collapse a security risk By Alison Withers and Stine Jacobsen, https://www.reuters.com/sustainability/cop/iceland-sees-security-risk-existential-threat-atlantic-ocean-currents-possible-2025-11-12/
Ewa M. Bednarz et al. Stratospheric Aerosol Injection Could Prevent Future Atlantic Meridional Overturning Circulation Decline, But Injection Location is Key, 2025, Earth’s Future, https://doi.org/10.1029/2025EF005919
Where the public conversation becomes dangerously narrow is in the proposed response. The typical conclusion is that the only way to prevent AMOC collapse is through rapid emissions cuts.
That’s essential — but it is not enough.
Emissions reduction works, but slowly. Even an aggressive decarbonization effort will not meaningfully cool the North Atlantic or slow Greenland’s meltwater discharge within the next twenty to thirty years. The AMOC, however, is responding to current heat and freshwater imbalances. We cannot rely on long-term tools alone to stabilize a system that may tip within decades.
If AMOC collapse is a national security risk — and it is — we need a broader response.
There are additional ways to reduce AMOC risk that deserve immediate research and serious public discussion.
Marine Cloud Brightening (MCB) is one of the most promising. By spraying fine sea-salt particles into marine clouds, MCB increases reflectivity and cools ocean surface waters. Unlike stratospheric aerosol injection, it is regional and adjustable. Applied over the subpolar North Atlantic, it could cool the critical region where AMOC deep-water formation is weakening, strengthen density contrasts, and even slow Greenland melt.
Greenland and Arctic ice-shielding is another option. Reflective covers, winter seawater pumping, fjord shading, and ice-mélange stabilization could reduce the freshwater plume that is destabilizing the AMOC. Every tonne of meltwater kept out of the Labrador and Nordic Seas reduces risk.
Targeted downwelling support, not global ocean manipulation, could help break the freshwater “lid” that now prevents winter convection in parts of the subpolar gyre. Localised ocean-mixing systems may restore the sinking that drives the AMOC.
Stratospheric Aerosol Injection (SAI) is powerful but high-risk — not the only hope, but potentially part of a carefully governed portfolio only to be seriously considered when all other options have failed (for which we are running out of time as our window for climate cooling is closing). A Northern Hemisphere-weighted SAI deployment could cool the Arctic and slow cryosphere loss. But this approach demands global agreement, caution, and contingency planning as well as broad political alignment.
Finally, carbon removal remains essential for long-term stabilization but is too slow to alter near-term tipping risk on its own.
The correct framing is not “mitigation or geoengineering.”
It is risk-risk management — the same logic used in national defence, public health, and disaster planning. Doing nothing new carries risks. Expanding our toolbox carries different risks. The responsible path is to evaluate all options, transparently and scientifically.
We cannot negotiate with the melting point of ice.
But we can choose whether to engage every available tool to keep the AMOC — and the societies it stabilizes — from crossing a point of no return.
Oxford (2024): Addressing the urgent need for direct climate cooling: Rationale and options
https://academic.oup.com/oocc/article/4/1/kgae014/7731760
Brilliant Tom, just brilliant. Great writing on a very complex set of issues.