Sluggish Atlantic circulation could cause global temperatures to surge

The circulation system of the North Atlantic Ocean has weakened and is predicted to weaken further in the coming decades. An analysis suggests that this decline could lead to accelerated global surface warming.
Gerard D. McCarthy is at the Irish Climate Analysis and Research Units (ICARUS), Department of Geography, Maynooth University, Maynooth, Ireland.

Search for this author in:

Peter W. Thorne is at the Irish Climate Analysis and Research Units (ICARUS), Department of Geography, Maynooth University, Maynooth, Ireland.

Search for this author in:

Global surface temperatures rose steadily from 1975 to 1998, but this growth then slowed somewhat for about 15 years — an event that gained popular attention1 as a ‘hiatus’. Since then, we have experienced the four warmest years on record, which has served to dampen popular interest in the event. However, because climate change is a complex response to slowly varying external drivers, it is important to fully understand past climate behaviour and the underlying causes. In a paper in Nature, Chen and Tung2 report that the system of ocean currents known as the Atlantic Meridional Overturning Circulation (AMOC) can explain changes in rates of global surface warming. Rather than the conventional picture of a vigorous AMOC associated with elevated surface temperatures in the Atlantic Ocean, the authors emphasize the role of the AMOC in taking heat from the surface and storing it in the deep ocean.

The connection between the AMOC and variations in the heat content of the subpolar North Atlantic Ocean has long been acknowledged. The AMOC transports heat northwards to the subpolar North Atlantic and to the Greenland, Iceland and Norwegian Seas. There, through a range of processes, deep water is formed that moves as a southward cold flow. This conveyor belt of northward-flowing, warm, shallow water and southward-flowing, cold, deep water defines the AMOC.

Relative to latitudinal averages, surface temperatures could be 5 °C cooler in the subpolar North Atlantic Ocean and up to 10 °C cooler in the Norwegian Sea if the AMOC were absent3. Consequently, a strong AMOC is typically associated with warming in the Northern Hemisphere. This association is consistent with evidence from palaeoclimatology that suggests that, during the most recent ice age, warmer periods coincided with a vigorous AMOC and colder periods coincided with a weak AMOC4.

Chen and Tung’s study emphasizes a different role for the AMOC in the modern climate. Atmospheric concentrations of greenhouse gases are currently being increased at a rate that is unprecedented in millennia and most likely millions of years. As a result, the role that climate mechanisms might have had in the past might not be a good guide to their current or future role. The authors contend that half of the heat arising from ever-increasing greenhouse-gas concentrations is stored in the deep waters of the North Atlantic when the AMOC is increasing, thereby reducing overall global surface warming (Fig. 1).

Figure 1 | The role of Atlantic circulation in the modern climate. Increasing atmospheric concentrations of greenhouse gases mean that more incoming solar radiation is trapped in the atmosphere, a consequence of which is surface warming. Chen and Tung2 report that the ocean circulation system known as the Atlantic Meridional Overturning Circulation (AMOC) can offset this warming. The AMOC transports heat northwards (red arrows) in the North Atlantic Ocean; the light- and dark-orange colours indicate regions that are 5 °C and 10 °C warmer, respectively, than latitudinal averages. The authors emphasize the role of the AMOC in transporting heat from the surface to the deep ocean (black arrows); the pink and blue colours indicate regions where heat increases and decreases, respectively, when the AMOC is increasing (based on Fig. 2a of ref. 2).

The authors show that a cycle of increasing and then decreasing AMOC from the 1940s to the mid-1970s coincided with a period of global-warming slowdown; a quiescent period of weak AMOC from the mid-1970s to the late 1990s coincided with rapid global warming; and an increase in AMOC strength from the late 1990s to 2005 and a decrease thereafter coincided with the ‘hiatus’ in global warming (see Fig. 3 of the paper2).

When the causes of the ‘hiatus’ were first being investigated, the Atlantic was not an obvious place to look. The focus was on the Pacific Ocean because the tropical Pacific was one of the only places where surface temperatures did not rise during this period5. Understanding of the event developed as several factors were taken into account, including the effect of changes in ocean heat content across multiple ocean basins6. Chen and Tung now bring focus to the North Atlantic. Their work suggests that the warm surface temperatures there were indicative of an increasing AMOC and that the associated increase in ocean heat uptake played a key part in the ‘hiatus’.

One of the main caveats of Chen and Tung’s study is that, by necessity, the authors used proxies for AMOC strength because no direct observations of sufficient length exist. There are only four observatories that measure the AMOC across the full width of the Atlantic: SAMBA at 34.5°S, RAPID at 26°N, NOAC at 47°N and OSNAP between 53°N and 60°N. The longest-running, RAPID, was deployed in only 2004. These observatories need to be maintained for many decades if we are to fully understand the role of the AMOC in our changing climate.

There is much to be done to determine how the AMOC affects surface temperature in different regions and on different timescales. For instance, Chen and Tung highlight the potential role of the Southern Ocean in heat uptake in the period since 2005. Such a feature could be part of a see-saw pattern of alternating heat uptake by the North Atlantic and Southern Ocean.

There is also a distinct difference between the effects of decadal AMOC variability and of an AMOC collapse on global temperatures. Although the prospect of the AMOC passing a tipping point and collapsing is considered unlikely, it is not impossible, and an event this dramatic could lead to global surface cooling7. The threshold between a weak AMOC that reduces ocean heat uptake, allowing global surface temperatures to rise unabated, and a very weak or collapsed AMOC that causes cooling in the North Atlantic and global surface warming to slow or stop will be a key point of debate.

The AMOC is deemed “very likely” to weaken in the coming decades1. Indeed, the Atlantic has seen muted rises in surface temperature relative to the global ocean over the past few decades. This relative lack of warming has been interpreted as a fingerprint of AMOC decline, potentially linked to anthropogenic climate change8. Whether the AMOC observatories will document the predicted decline remains to be seen, but they have already observed that the AMOC is in a weakened state9. Chen and Tung predict that such a weak AMOC will result in a period of rapid global surface warming that could last for more than two decades.

Nature 559, 340-341 (2018)

doi: 10.1038/d41586-018-05712-x
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Intergovernmental Panel on Climate Change (IPCC). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC, 2014).

  2. 2.

    Chen, X. & Tung, K.-K. Nature 559, 387–391 (2018).

  3. 3.

    Jackson, L. C. et al. Clim. Dyn. 45, 3299–3316 (2015).

  4. 4.

    Kosaka, Y. & Xie, S.-P. Nature 501, 403–407 (2013).

  5. 5.

    Medhaug, I., Stolpe, M. B., Fischer, E. M., & Knutti, R. Nature 545, 41–47 (2017).

  6. 6.

    Broecker, W. S. Oceanography 4, 79–89 (1991).

  7. 7.

    Drijfhout, S. Sci. Rep. 5, 14877 (2015).

  8. 8.

    Caesar, L., Rahmstorf, S., Robinson, A., Feulner, G. & Saba, V. Nature 556, 191–196 (2018).

  9. 9.

    Smeed, D. A. et al. Geophys. Res. Lett. 45, 1527–1533 (2018).

Download references