Editorial NexoScience
Europe’s mild winters have less to do with its latitude than with a conveyor belt of warm water called the Atlantic meridional overturning circulation, or AMOC. Without it, cities as far south as London and Paris would likely feel weather typical of much colder latitudes. A team at Utrecht University in the Netherlands ran one of the most ambitious climate simulations attempted so far to test how close that conveyor belt actually is to breaking down, and what the world would look like if it did.
The simulation pushed the AMOC past its breaking point for the first time inside a full-complexity global climate model. Using the Community Earth System Model, the same kind of simulation used in major international climate assessments, the researchers slowly added fresh water to the North Atlantic over more than a thousand simulated years, mimicking the effect of melting ice and heavier rainfall. For roughly thirteen centuries the circulation weakened gradually. Then, within about a hundred years of simulated time, it collapsed almost entirely, its strength dropping from around ten Sverdrups, a unit equal to a million cubic meters of water per second, to nearly zero.
The fallout inside the model was fast and severe. Northern Europe cooled by more than one degree Celsius per decade in the worst-affected areas, and in Bergen, Norway, February temperatures dropped by roughly three and a half degrees per decade, a pace far beyond anything current infrastructure or agriculture is built to absorb. Some coastal stretches of the Atlantic also saw dynamic sea level rise of more than seventy centimeters as the collapse reorganized ocean currents, while the Amazon’s wet and dry seasons flipped entirely, threatening to disrupt the rainforest’s ecosystem.
The real contribution of the paper is not the collapse itself, but a way to see it coming. Earlier attempts to forecast an AMOC tipping point relied on statistical patterns in sea surface temperature, like rising variance or growing autocorrelation, and the authors show these can give inconsistent answers depending on which time window gets analyzed. Instead, the team identified a more physical, directly measurable indicator: the minimum in the AMOC’s own freshwater transport at 34 degrees south, a value tied to the feedback loop that destabilizes the current. In the simulation, this indicator dipped roughly twenty five years before the full collapse.
That lead time matters because the same indicator can, in principle, be tracked today. When the researchers applied it to real-world reanalysis data spanning the past four decades, they found the trend already pointing in the same direction the model showed just before its own collapse: a steady, measurable weakening of that freshwater transport.
None of this means a collapse is imminent or guaranteed. The simulation needed an unusually large amount of added fresh water, close to eighty times the current melt rate of the Greenland ice sheet, to trigger the tipping point, largely because of known biases in how the model distributes rainfall and salinity. The authors are explicit that real-world observation records are still too short, at most about a century, to say how many years remain before the AMOC could reach that same threshold. What the paper does establish is that the collapse is physically possible in a full-complexity climate model, not just in the simplified conceptual models used until now, and that a concrete, trackable warning sign now exists.
Real-world signs are already pointing in a troubling direction. Long-term reconstructions suggest the AMOC’s strength has dropped by roughly three Sverdrups since 1950, and proxy records indicate it may be at its weakest point in over a thousand years. Continuous monitoring arrays across the Atlantic, including one at 34 degrees south running since 2009, are now the closest thing scientists have to an early warning system, precisely the kind of instrument this new indicator depends on.
The study, led by René M. van Westen, Michael Kliphuis and Henk A. Dijkstra at Utrecht University, was published in 2024 in the journal Science Advances under the title “Physics-based early warning signal shows that AMOC is on tipping course” (DOI: 10.1126/sciadv.adk1189).
