HasTheAMOCCollapsedYet.com
No, not yet.*
* "Collapse" here doesn't mean a sudden crash. If it happens, it would unfold over decades. And the science does not say it's inevitable. How likely it is depends a lot on how much (and how fast) the world keeps warming.
| Question | Answer |
|---|---|
| Has it collapsed? | No |
| Is it still moving warm water north? | Yes |
| Is it weaker than it used to be? | Probably. Scientists are still arguing about how much. |
| Will it collapse this century? | Nobody knows. That's sort of the problem. |
| Did it collapse in 2025, like the headline said? | No (see FAQ) |
| Should you check back later? | Yes |
Counting from the end of the Younger Dryas, a cold snap about 11,700 years ago that's usually blamed on the last big AMOC shutdown.
The Atlantic Meridional Overturning Circulation is a giant loop of ocean current. Warm, salty water flows north near the surface (the Gulf Stream is part of it). Up near Greenland and Iceland it gives off its heat, gets cold and dense, and sinks. Then it creeps back south along the bottom of the ocean, and the whole thing goes around again.
NORTH (near Greenland) SOUTH (tropics)
[cools off, sinks] <~~~ warm salty water ~~~ [warmed by sun]
| (near surface) ^
| |
v |
============ cold deep water (way down) =============>
It moves something like 17 million cubic meters of water per second. Oceanographers call that 17 Sverdrups, a unit that exists mostly for this purpose. It's about 80 times the flow of the Amazon River.
That heat is part of why London's winters hover a few degrees above freezing, while Labrador, at about the same latitude, averages around -15°C in January.
The loop depends on water near Greenland being heavy enough to sink. Two things make it lighter: warming, and fresh water. Climate change is supplying both, with Greenland's ice sheet melting fresh water right into the spot where the sinking happens.
If the sinking slows enough, the loop could stall. Climate models suggest that once it passes a tipping point it wouldn't come back on any timescale humans care about.
If it did, models predict much colder winters in northwestern Europe, shifted monsoons in Africa and Asia, extra sea level rise on the US East Coast, and a lot of very busy climate scientists. It would play out over decades, not over a long weekend.
Artist's impression. Not to scale. Not to speed.
Yes, but only since 2004, when the RAPID array of moorings started measuring it across the Atlantic at 26.5°N. Twenty-odd years is not very long for something this slow and noisy, which is why there's so much disagreement. Older estimates come from indirect clues like sea surface temperatures and sediment cores.
The best-known clue is the "warming hole" (sometimes called the "cold blob"). Over the last century almost the whole ocean has warmed, but one patch south of Greenland has barely budged, and in some years it has actually been colder than it was in 1900. That patch is right where the AMOC delivers its heat.
How much the sea surface warmed, comparing the average for 1996–2025 with the average for 1900–1929. Each square is 2° of latitude by 2° of longitude; gray is land or sea ice. Data: NOAA ERSSTv5.
Both the hole south of Greenland and the extra-warm water off New England and Atlantic Canada are what a slowing AMOC would leave behind (Caesar et al. 2018). To be fair to the doubters: some of the hole is natural swings that last decades, and it isn't proof of collapse on its own. But it is the kind of map that makes oceanographers frown.
Loading headlines...
Loading papers...