Transcription
Britain sits at the same line of latitude as Siberia, and Norway lines up with the southern tip of Greenland. By all rights, Northern Europe should be a frozen tundra, in the nicest way possible, for 6 months a year, largely uninhabitable and home to about 1/10th of its current population.
It isn't, because of a current in the Atlantic Ocean that has been quietly pumping tropical heat northward for thousands of years. But right now, that current is the weakest it has ever been in 16 centuries. And this year, for the very first time, scientists have confirmed that we are living through the early stage of its pre-collapse warning system. This is the Atlantic Meridional Overturning Circulation, or the AMOC. And today, I want to look at what exactly it is, why it is so fragile, and what the new research is really telling us about how close to losing it we really are.
The system starts in the Gulf of Mexico, where shallow tropical water sits at around 27° C. From here, wind stress on the surface drives the warm water northward along the US East Coast, through the Straits of Florida, up past North Carolina, at about 2 to 4 miles per hour. By the time it reaches the open Atlantic, it has more than doubled in size, moving 80 million cubic meters of water every second in a flow 150 km wide. This is the Gulf Stream, the part that most people have heard of, but it's only the surface layer of something that runs the full depth of the ocean and spans the entire planet.
As it travels, it releases its heat into the atmosphere above it. And this oceanic underfloor heating carries enough thermal energy to keep European winters mild.
"It is well off."
As this water cools, it evaporates and becomes saltier and denser. And somewhere off the Norwegian coast, it grows so cold and so dense that it can no longer stay at the surface. And so it sinks, descending into the deep ocean, where it begins flowing southward along the seafloor. This slow, dark river moves a few cm per second towards Antarctica.
When it finally reaches the Southern Ocean, the powerful westerly winds that circle Antarctica drag the surface water northward away from the continent. And as that surface water is pulled away, this draws the deep water upwards in a process called upwelling. This powerful circulation drags the water behind it, but also pushes the water in front of it, keeping the whole feedback loop running. And the amazing part is that it takes a single molecule roughly 1,000 years to complete this circuit, which is kind of wild to think about.
The scale of this energy transfer in this system is almost incomprehensible. The AMOC moves roughly one petawatt of thermal energy every second, which is about 50 times the total of all human civilization's energy consumption, and is all maintained by this perpetual and elegant self-reinforcing feedback loop. The problem is that feedback loops are only stable until suddenly they aren't. And that is what the evidence suggests that we are heading towards. I want to talk about that.
But first, we have to thank the sponsor who makes these sorts of videos possible, ODU. ODU is the all-in-one business management platform, the self-reinforcing loop your business has been missing. It has over 45 apps such as invoicing, accounting, project management, and inventory. The bit I want to highlight specifically today is the website builder that lets you build your own custom site in just four simple steps.
First, tell it what kind of site you're building, your industry, and what you want it to achieve. Maybe say something to inform the public about the imminent collapse of the AMOC. Step two, choose a color palette or upload your logo. ODU builds a visual identity around you rather than the other way around. Step three, pick the pages that you need, and then all that is left to do is choose a theme. I'm going with something clean, minimal, and slightly foreboding. And then you have a fully functioning website. It's customizable just by dragging and dropping blocks. And there's no technical skills required. The text, colors, images, and animations all stay where you put them. And unlike the AMOC, the position of every element is entirely within your control and stays where you need it to. The first app is free for life, including unlimited hosting, unlimited support, and a free custom domain name for a year. Click the link in the description to try your first app for free. And thank you to ODU for supporting the channel. Now, back to the video.
"What keeps a system like the AMOC flowing is the ability for the cooling Northern Atlantic water to be so dense that it sinks relative to the water around it."
Water can be dense for two different reasons: either its temperature or its salinity. Warm water is less dense than cold water because heat makes the molecules vibrate faster, pushing their neighboring molecules further apart, making the system take up more space and lowering its density. Cooling them down, and that vibration slows, and the molecules pack slowly closer together. And that water gets heavier per unit volume and so sinks. Fresh water is less dense than salt water for the simple reason that salt water has dissolved salt ions physically packed into the water, adding mass without adding much volume. The same liter of water simply weighs more with salt added to it.
Unfortunately for us, right now there is essentially a tap of ice-cold fresh water pouring into the North Atlantic from the Greenland ice sheet, which is currently melting at a rate faster than any point in recorded human history. This means that the cooling, salty water from the Gulf Stream is now competing against the icy freshwater glacial runoff. This dilutes the density contrast of the water that usually tells the cooling AMOC water to sink in the first place. What this means is that it sinks with a reduced force. So, it either doesn't reach as deep, or it doesn't move as quickly. And that means that it doesn't drive the overall current with as much power.
This whole system is pretty robust and can absorb a significant influx of fresh water, and the only consequence will be a gradual weakening of the current, but that is up to a certain point called a tipping point. And beyond that, the process collapses suddenly and entirely. We know that this is possible because it has happened before. 12,900 years ago, near the end of the last ice age, the climate was warming rapidly and the great ice sheets were retreating. Then, very suddenly, the North Atlantic region snapped back to near full glacial conditions. This event is known as the Younger Dryas and lasted for a little over a thousand years. We believe that it was caused by a massive influx of glacial meltwater in the North Atlantic, exactly the kind of freshwater disruption that we've been describing here. This ultimately triggered a major slowdown or full shutdown of the AMOC.
The important part now is understanding what the early stages of this collapse might actually look like.
"Don't push me 'cause I'm close to the edge. I'm trying not to lose my head. Uh-huh. Uh-huh."
For decades, the honest answer to "how close are we to the tipping point?" was: we don't really know. We could measure the AMOC strength directly. The RAPID array of moorings across the Atlantic at 26 degrees has been doing this since 2004, but strength alone doesn't really tell you proximity to collapse. A current can be gradually weakening, but still nowhere near its threshold. What we needed was less of a speedometer and more of a fuel gauge. And a team at Utrecht University have just built one, and it gives us a better understanding than anything we have ever had before.
Their insight came from an observation that has been sitting in the literature for years without anyone really fully appreciating just how important it is. The Gulf Stream's position isn't held in place by surface winds. It is anchored below by the deep western boundary current, making its return journey southward across the seafloor. As it flows past the point where the Gulf Stream detaches from the American coastline, it generates a rotational force that pins the Gulf Stream in place. If the AMOC is weakening, that deep current should weaken too, and the Gulf Stream should start to drift north. However, most of our models use grid squares roughly 100 km across, and at that resolution, we have never actually seen this behavior.
So that's when researchers Van Westen and Dijkstra decided to see if they could fix it. Their models use a 10 km grid square, 10 times the resolution of the standard climate models, and fine enough to see the Gulf Stream sit where it actually sits. The slightly disappointing reason why this work hasn't been done before is pretty straightforward. The relationship between grid resolution and computing cost means that this model is roughly 1,000 times more expensive to build and run than any of the others. If you try to run this simulation on a normal laptop, you would be waiting 14 million years for any results. And even using the Dutch national supercomputer called Snellius, which I always find kind of weird, which runs at 14 quadrillion calculations per second, it took 6 months. This time obviously needed paying for, and humanity, at the moment, and maybe for the future, you might have heard, doesn't really like funding climate research. Fair enough.
To that output, though, this is now a very high-resolution model. They then began to introduce fresh water into the North Atlantic at a slowly increasing rate, simulating the Greenland melt. And they basically just watched what happened. What then unfolded happened in two stages that were pretty distinct.
In the first stage, over roughly 400 simulated years, the Gulf Stream drifted gradually northward, about 133 km in total. Slow and easy to miss if you weren't looking for it. This is the pre-collapse signal that the AMOC is weakening, but it is still in its original, stable-ish state. Then, running the simulation forward over just two further simulated years, the Gulf Stream lurches northward by about 219 km. This is the tipping point. At this point, the deep circulation has changed so rapidly it has lost the ability to pin the Gulf Stream in place. And as a result, the AMOC disintegrates. What took 400 years in stage one took just 2 years in stage two.
Now, we don't have direct measurements of the full AMOC because monitoring the deep ocean in real time is extraordinarily difficult and expensive. But we can monitor the Gulf Stream position from space using satellites continuously, in basically near real time. And so, with this insight, after establishing this two-stage pattern in the model, the team went looking for it in the real world. Using satellite altimetry data from 1993 to 2024 and subsurface temperature observations going back to 1965, they found both show a statistically significant northward drift of around 53 km in the Gulf Stream's position since just 1993. That 53 km of movement has happened in just 30 years.
We don't know how much total movement or how long this process has been occurring because we only have data that goes back to 1993. But the Gulf Stream's location is currently exactly the location that the model predicts. So we know that we are at least in stage one of this process. The question, though, is how far through it are we? And that question is much harder to answer.
This model is running under idealized pre-industrial conditions with only freshwater forcing added slowly. The real ocean is simultaneously experiencing warming, changing wind patterns, shifting atmospheric circulation, and freshwater input all at once. In the model, what they find is that stage one ends approximately 25 years before the full AMOC collapse. That figure carries real uncertainty, though. Ocean models aren't crystal balls, and the actual timeline depends heavily on how fast Greenland's ice sheet continues to melt, which in itself is kind of an open question. The windows could be much longer, but they could also be much shorter.
What the Utrecht research actually delivers, and it's important to be just precise about this, is not a countdown. It is finally a mechanism and a model. For the first time, there's a physically grounded, satellite-observable proxy for the health of the most important ocean circulation system on Earth. We have been watching the right data for 30 years without really knowing what we were looking at. We do now.
The question is just how bad could it get? There it is. Without the heat that the AMOC pumps northward, average winter temperatures across Europe drop by about 10°. Sea ice, currently confined to the Arctic, would creep southward, swallowing the coastlines of Britain and Northern Europe. London could see temperatures as low as minus 20°, and Scandinavia could be transformed into a tundra colder than Siberia. Growing seasons across Northern Europe, the agricultural epicenter for centuries, would simply stop working.
And as the North Atlantic cools, the great tropical rain belts would start to shift. The monsoon systems supplying seasonal rainfall to West Africa and South Asia would weaken and drift southward, disrupting the food supply of over a billion people. The AMOC also drives the upwelling of cold, nutrient-rich water from the deep ocean. If you remove it, then the surface waters begin to starve of the nutrients that feed phytoplankton, which are the base of basically the entire marine food web. This potentially collapses large parts of every food chain above it.
What I really think is interesting, and I'll borrow these ideas from a few other hats that I wear, is that humanity really gets and embraces the idea of asymmetric upside potential. The idea that a small bet might have a big payout, being discovered by a talent agent and being invited to the big time, or buying a lottery ticket and winning the jackpot, or making a small bet on a small company and finding the next Google. It is exciting and it is motivating, and it is what humans think about quite a lot.
What we are terrible at dealing with is asymmetric downside potential: that small spends or changes in our behavior might avert unbelievably negative consequences. I find it wild that one quarter of all homeowners in the UK do not buy home insurance. You've heard of things like house fires or gas main explosions or other reasons that people lose their houses, and still some people choose not to protect it. By contrast, we have never had to pay for the planet and the environment that we have, and none of us have ever seen a globally significant event like an AMOC collapse, 'cause it hasn't happened in recorded human history.
There is this concept in cosmology called the Great Filter, which I think about quite a lot. The idea that somewhere in the development of an intelligent civilization, there is a barrier that most civilizations simply do not survive. We tend to imagine it as something really cinematic: an asteroid, a war, a pathogen, something that arrives with enough drama and visibility to demand an immediate response. I think it is much more likely that it is a quieter version of the filter that does the real damage. Civilizations don't usually fail because they miss big, dramatic threats. They fail because they miss and normalize slow ones.
We have known the AMOC was weakening for over a century. We've had satellites watching the Gulf Stream drift for 30 years. The signal has been there the whole time, hiding in plain sight in the data that we were already collecting. Whether that information has come sufficiently early or already too late is the part that we, as a species, unfortunately don't get to simulate. We just have to live through. And it raises the kind of uncomfortable question: what else are we missing?