As the world careens past our hoped-for target of 1.5 degrees Celsius warming, scientists are growing increasingly alarmed that we may be nearing a dramatic, long-feared “tipping point” — a moment when the main ocean current in the Atlantic Ocean becomes destined to shut down, clamping off the primary source of warmth for northern Europe and playing havoc with the global climate.

Such a scenario has been a concern for many decades, but the issue is now heating up. “I have personally researched this for 35 years,” says Stefan Rahmstorf, a physical oceanographer at the Potsdam Institute for Climate Impact Research in Germany. “For the first 30 years we considered this a low likelihood event — I would have said a 5 percent chance of occurring. It’s more like 50/50 now. I would even say more likely than not.”

The main ocean current in the Atlantic — the Atlantic Meridional Overturning Circulation — is one of the most powerful drivers of heat around the globe. The AMOC is the reason that northern Europe is so temperate, a couple of degrees warmer than it ought to be given its latitude. 

In November, Iceland made the unusual move of designating the risk of an AMOC shutdown a national security threat.

The new worries come from multiple lines of evidence. Researchers have now amassed two decades of solid data on ocean flow, just enough to show a hint of a trend, which suggests the AMOC has begun to slow. To predict the current’s future, researchers turn to models: While there hasn’t been a dramatic improvement in the ability of models to simulate the AMOC, studies focusing on models that best match the real-world data, or that look further into the future, put a pessimistic spin on results. 

The science is still far from certain, with a lot hanging on still-limited data and models that disagree with each other wildly. That leaves many scientists unwilling to bet whether we’re close to a tipping point or not. “I absolutely take seriously the existential threat of an AMOC shutdown, and it’s not outside the realm of possibility in the next century,” says Fraser. “I couldn’t pin it down more than that.”

In October 2024, more than 40 researchers, including Rahmstorf, issued an open letter to the Nordic council of ministers drawing their attention to the “serious risk” of AMOC collapse. Some politicians are sitting up and paying attention. In November 2025, Iceland’s government made the unusual move of designating the risk of an AMOC shutdown a national security threat. “There are some now waking up to the threat,” says Rahmstorf.

The two decades of ocean data that researchers have now amassed is just enough to begin to spot possible trends.

The AMOC is driven by the formation of deep water near Greenland: Cold, salty water sinks for the simple reason that cold water is more dense than warm water, and salty water is more dense than fresh. This sinking pulls surface waters toward northern Europe from the east coast of the Americas and further south, powering a massive global conveyor belt of water that extends all the way down to the southern tip of Africa. But as the north warms thanks to climate change, its waters become less dense and therefore less inclined to sink. The current slows. This pulls less salty water up from Africa, making the water in the north even less dense, and even less inclined to sink. The current slows more. Melting ice from Greenland freshens the waters further, exacerbating the problem. Eventually, this feedback loop shuts the current down.

Something similar has happened before: An AMOC shutdown played havoc with the planet 12,000 years ago, when the end of the last ice age dumped huge amounts of fresh water into the northern ocean. This plunged Europe back into ice-age-like conditions, with Greenland, for example, cooling by a stunning 18 degrees F (10 degrees C). “The effects are amongst the most drastic we have seen in the paleorecord,” says Rahmstorf. “We really don’t want this to happen.”

Add enough fresh water, and at some point, researchers expect the system to pass a “tipping point” after which collapse becomes inevitable, though it might still take decades to a century to play out. Once the AMOC collapses, the current doesn’t just spontaneously start up again if planetary warming reverses. Instead, the Atlantic gets locked into a new stable “off” state for centuries.

Researchers have been on the hunt for signs of a slowing or weakening AMOC for a long time, tracking ocean temperatures, pressure, salinity, and the speed of flow. It has been tricky: The ocean’s properties vary so much from year to year that it makes long-term trends hard to spot. 

A consistent dataset from an array of moored buoys looking at the AMOC only dates back as far as 2004 (coincidentally the same year the Hollywood blockbuster “The Day After Tomorrow” threw the North Atlantic current into the spotlight). The two decades’ worth of data that researchers have now amassed is just enough to begin to spot possible trends. It shows plenty of wiggles, including roughly a decade where the current strengthened rather than weakening. “AMOC is much more variable than we had previously realized. There’s lots of ups and downs,” says Fraser.

Climate models that predict a shutdown disagree wildly with each other about when it might happen.

One way to check is to look for distinctive “fingerprint” patterns in sea surface temperatures, for which we have a longer record. In 2018, Rahmstorf and others showed that an unusual “cold blob” of water in the North Atlantic is consistent with climate models that feature an AMOC that has slowed from about 20 Sverdrups in 1950 down to around 17 Sverdrups in the 2000s. These fingerprint methods are powerful, says Fraser, but not perfect. “None of this stuff is controversial, but it’s all debated,” he says.

The next, more important questions are how much more the AMOC will slow, and if and when it will hit a tipping point and eventually collapse, reaching a sluggish flow of less than 6 Sverdrups. 

One approach to this question is to look for mathematical signs of tipping. All systems that have two stable states, from economics to the environment, exhibit the same sorts of statistical behaviors when they are close to their tipping point, such as wild swings in the data. Researchers including brother-and-sister team Peter and Susanne Ditlevsen at the Niels Bohr Institute in Copenhagen went hunting for such statistical “early warning signs” of tipping in more than a century of Atlantic sea surface temperatures. In 2023, the Ditlevsens used this method to predict that, if global emissions continue to grow, the AMOC will hit its tipping point between 2037 and 2109. The result corresponds to a slowdown of about 5 Sverdrups per decade, which matches the rate of decline seen over the last several years. Plenty of researchers caution, though, that these statistical methods have a lot of uncertainty. 

A complementary approach is to try to whittle down the best of the climate models. Not all global climate models support the idea that the AMOC is close to a tipping point. Some don’t show a collapse no matter how much fresh water is poured into the north; instead, the flow slows a bit but persists in its “on” state. The ones that do predict a shutdown disagree wildly with each other about how much fresh water is needed to trigger the switch, or when it might happen. 

“Even without a collapse, a weakening of the AMOC could have serious climate impacts,” says a scientist.

Rahmstorf and others argue that these disagreements reveal flaws in the models. First, says Rahmstorf, it’s hard to model the AMOC because it relies on subtle changes in water density, which requires modeling tricky things such as ice melt, precipitation, and clouds. “It’s a tough problem to get this right,” he says. 

On top of that, Rahmstorf contends, there has been a one-way bias in correcting these models. Models that initially showed that an AMOC collapse should have already happened would have been tweaked to fix this obvious error. On the other hand, models that show that an AMOC collapse lies far, far in the future (or never) aren’t flagged for correction. 

Some of the models that show the AMOC recovering from huge amounts of freshwater inflow, says Rahmstorf, just haven’t been run far enough into the future. Last summer, Rahmstorf and colleagues looked specifically at the few models that have been run all the way out to 2300 or later. From this subset, they estimate the chance of an AMOC collapse to be startlingly high: from 25 percent likely under low-emissions scenarios, to 70 percent under high-emissions scenarios, with the collapse happening most often in the early 2200s. “That was quite a shock to me when I saw those results,” says Rahmstorf.

In April, another paper was published with even more worrying numbers. All the IPCC models used in the 2021 report together predict a reduction in AMOC strength by about 30 percent below the preindustrial average by 2100; but when climatologist Valentin Portmann at the University of Bordeaux in France and colleagues constrained the models using real-world Atlantic Ocean temperature and salinity data, the models tightened to show an AMOC weakening of 50 percent by 2100. “Our work says: Warning, the AMOC decline may be stronger than expected,” says Portmann. This study didn’t try to pin down how close this puts us to a tipping point, but Rahmstorf says that a 50 percent decline would “very likely” put us past it. 

There’s a lot left to pin down — including the situation at the other end of the globe. Warming of the South Pole and the melt of Antarctic ice is similarly expected to muddle with the formation of deep waters there, which will also disrupt ocean currents. On the other hand, says Jonathan Baker of the U.K. Met Office, persistent winds around the Antarctic seem to be helping the AMOC to resist collapse. Regardless, everyone agrees: Whether the chances are 5 percent or 50 percent or 70 percent, the serious consequences of an AMOC shutdown means the risk is another, urgent reason to cut our greenhouse gas emissions as quickly as possible.