Collapse of AMOC ocean current may already be locked in

A potentially catastrophic collapse of the Atlantic Ocean currents that control Europe’s climate may already be inevitable. Based on model simulations, researchers estimate that there is a 10 to 23 per cent chance that such a collapse is locked in.

“There is a significant probability that we’re already committed to collapse, and we can’t change that even now,” says Phil Holden at the Open University, UK.

The Atlantic Meridional Overturning Circulation (AMOC) carries warm, salty water from the tropics into the North Atlantic Ocean, where it cools, sinks and then returns south. This circulation regulates the climate across Europe, Africa and the Americas.

Recently, there have been signs that this vital current system is weakening, including by slowing in some areas, partly because the melting of the Greenland ice sheet caused by climate change is making the salty water less dense, so that it sinks more slowly.

Some scientists have suggested that the AMOC could collapse, plunging Europe into near-Arctic conditions and weakening monsoon systems around the world. One recent study found the AMOC could cross a tipping point within decades, but it is difficult to say how likely this is.

“The AMOC collapse has just been so intangible,” says Holden. “So far there’s been no firm quantification of whether it is going to happen or when it is going to happen.”

“There’s a lot of uncertainty, and a lot of different opinions among the expert scientific community,” says Tim Lenton at the University of Exeter, UK.

To get a better idea of what might happen to AMOC, Holden, Lenton and their colleagues ran 21 computer simulations with varying rates of Greenland ice melt and emissions peaking at different dates, at 10-year intervals from 2005 through to 2135. The team assumed that, after the peak, greenhouse gas emissions would fall to net zero over 35 years while the melt rate of Greenland ice would stay constant. Each simulation ran for a total of 300 years.

The models indicate that under very conservative assumptions – emissions peaking in 2025 and the Greenland ice sheet adding just 54 millimetres to sea-level rise by 2100 – there is a 10 per cent chance that the collapse of the AMOC is already inevitable. The researchers defined this as when the circulation would only occur at lower latitudes, and when the overturning current that brings heat to the high latitudes has stopped.

If we don’t start on the path to net zero until 2100, the probability of collapse rises to 80 per cent, the model predicts.

Under less conservative assumptions, with melting ice from Greenland adding 274 mm to sea-level rise, as is projected by 2100, the probability that we are already committed to collapse is 23 per cent.

Even when a collapse becomes inevitable, it would take a long time to happen. In the simulations, the average delay between the year in which the world becomes committed to collapse and the collapse coming to pass was 84 years, with the earliest collapse occurring around 2060.

“This idea of talking about committed collapse, rather than when the collapse actually happens, frames it in a way that’s quite interesting for risk management,” says Till Wagner at the University of Wisconsin-Madison, but he is cautious about extrapolating to the real world. “I think there’s fairly good evidence that there’s going to be a weakening, but the actual larger-scale dynamical outcome is still very much up in the air.”

Jonathan Baker at the Met Office, the UK’s national weather service, says the simulations provide a valuable way to explore how the AMOC responds to different scenarios, but the low resolution of the model used means it isn’t as sensitive as some other climate models, which might influence its estimates of risk.

Many state-of-the-art climate simulations can compute the globe in grids of 1° of latitude and longitude, which takes huge amounts of computational resources and time to run long-term simulations. The model used in this study uses 5° grids, but this low resolution was a deliberate choice, says Lenton.

“There isn’t the compute power, or no one’s had the ability to do this exercise with a higher-resolution model,” he says. This does mean the probabilities of the risk estimates could be different at higher resolution, he says, but recent research on the AMOC using a higher-resolution model indicates that, if anything, this might raise rather than lower the estimates.

“Further work using multiple climate models and comparison with the wider body of evidence will be important before drawing reliable conclusions about the magnitude of future AMOC collapse risk,” says Baker.

If there is a chance the world is already committed to AMOC collapse, as the model suggests, this should provide extra incentive to cut emissions, says Lenton. This is because the model indicates that the probability of AMOC collapse rises sharply if net zero is delayed. If emissions continue unabated for 10 extra years beyond the point of commitment, the actual collapse would happen faster – after 57 years, on average, rather than 84.

“What the model is saying to me is ‘let’s do everything in our power to get to net zero as quickly as possible to try to keep this probability down at the 10 per cent level’,” says Lenton.

This chimes with research published last month hinting that the slowdown of the AMOC may be reversible – if carbon dioxide emissions come down enough.