How will media report on this new AMOC study?

I’ve been getting a lot of media queries about a new paper on the AMOC, which has just been published. In my view this large media interest is perhaps due to confusing messages conveyed in the title of the paper and in press releases about it by the journal Nature and by the Met Office. Whether intended or not, these give the impression that new model results suggest that the AMOC is more resilient than previously thought. That’s (unfortunately!) not the case.

This new paper does not (and does not claim to) contradict earlier modeling studies about future AMOC changes and their climatic impact, as one of the authors (Richard Wood) has confirmed to me (we are presently both attending an AMOC workshop in Utrecht). It’s the same models, showing the same things – just the wording is different. What previous studies have labelled an ‘AMOC collapse’ is now called ‘no collapse’. It’s essentially a discussion about semantics, not physics. Do you call it an AMOC collapse if a weak and shallow wind-driven overturning persists after the thermohaline part has collapsed? Or not?

That is not about any difference in climate impact. The AMOC’s climate impact in these model simulations is the same as in previous studies, which have indeed used the same models and often in fact the same model experiments, as this study has analysed existing model runs.

A typical example of those is the paper by Bellomo et al. 2023 using the EC-Earth3 model – that same model is also included in the new paper by Baker et al. If you try to kill the AMOC by adding a lot of freshwater to the northern Atlantic (no greenhouse gas increase), you get the following change on Atlantic overturning (Fig. 1).

It is clear that – as usual – some overturning remains. The climate impact looks similar to other models (Figure 2): massive cooling in the Northern Hemisphere.

Figure 2 Surface temperature change in response to a near-shutdown of the AMOC.

Another example is the analysis of CMIP6 results which I showed in my presentation last October at the Arctic Circle Assembly, when presenting the open letter by 44 experts to the Council of Nordic Ministers. The preprint of this work has been online since last September. CMIP6 is the current model generation, also used in Baker et al.

It presents a selection of standard climate scenarios with those CMIP6 models, as shown in the last IPCC report, in which the AMOC largely grinds to a halt in the next century, as shown in Figure 3.

Figure 3 AMOC strength and ocean heat transport at 26°N in a selection of CMIP6 models in which the AMOC nearly collapses.

It is clear that some overturning remains in all of them – as is indeed expected, as it is the density-driven (i.e. thermohaline) part of the AMOC which has a well-known tipping point, due to Stommel’s (1961) famous salt transport feedback. It has been well-established since the 1990s (Toggweiler and Samuels 1995) that the AMOC also has a component driven by the winds, which will persist as long as the winds won’t stop blowing (except possibly if it switches to the North Pacific rather than North Atlantic). We actually pointed out in our preprint mentioned above that we focus on whether “the deeper thermohaline part of the AMOC becomes weak and/or collapses, since it is only this part of the AMOC that possesses a tipping point.” Now the new paper focuses on the wind-driven part, which thus complements our study, but it does in no way contradict it.

It does not change the assessment of the risk and impact of future AMOC changes in response to human-caused global warming.

(Please post a link in the comments if you see media coverage of the Baker study.)

Post script: As an aside, together with my Australian colleague Matt England I published a systematic study on the wind-driven part of the AMOC already in 1997.