The Atlantic Meridional Overturning Circulation (AMOC) - part of the global ocean thermohaline circulation - plays a crucial role in the circulation patterns that influence global temperature and precipitation. While there is uncertainty 9,10, some observational evidence suggests that the AMOC has weakened by around 15 percent in recent decades, is at its weakest point in the last 1,600 years and that this weakening may be accelerating 11–14. Ongoing freshwater influx in the North Atlantic may inhibit deep water formation resulting in a continuing weakening - or even collapse - of the AMOC 15–17. This would represent a fundamental reorganization of ocean circulation, causing a redistribution of heat around the planet and a corresponding coupled response from the atmosphere 18.
The IPCC consensus is that while the AMOC will “very likely” (90-100%) weaken later this century, it is “very unlikely” (<10%) to collapse within this century 19. However, others consider the risk of AMOC collapse to become significant between 2-3°C global warming 16,20,21. Recent research estimated that AMOC is on course to collapse 22 and this could occur between 2025 and 2095, with a central estimate of 2050, if global carbon emissions are not reduced 20. Here we assess the impacts of a collapse of the AMOC at 2-3°C global warming above pre-industrial temperature. A world in which the global mean temperature is 2.5°C warmer than pre-industrial times is representative of where we will end up later this century under current legally-binding policies 23. Crossing an AMOC tipping point would have major global ramifications. These include a drastic cooling of western Europe, a reduction of rainfall in the Atlantic region and a shift in the intertropical convergence zone shifting rainfall southwards. The change in ocean circulation patterns would affect marine ecosystems and give faster sea level rise along the Northeast seaboard of North America and parts of Europe. Stronger hurricanes in the Southeastern United States and the Caribbean, and reduced rainfall across the Sahel have also been predicted 1,24–27. It could also trigger tipping points in the Amazon and have significant impacts on tropical monsoon systems 28–32. Recent work indicates that, without the AMOC, the tropical Pacific Ocean cools, and trade winds intensify and shift south, putting the Earth in a climate state that resembles a permanent La Niña, which could trigger catastrophic monsoons and flooding in the South Pacific 33.
An AMOC collapse may thus represent an existential threat to humanity 34. Some economic analyses have hypothesized that AMOC collapse would cause a 25-30% loss to GDP comparable with the Great Depression but irreversible 35. However, other economic analyses have suggested that despite the possibility of wide-reaching economic and human impacts, the impact on GDP may be small or even positive 2–4. A key weakness across these studies, which rely on simple Integrated Assessment Models (IAMs), is the overlooking of precipitation changes and within-country climate variations in favor of a narrow focus on temperature impacts on GDP. This fails to account for substantially different climate change effects in subregions. Further, this approach does not account for the impacts of potentially divergent directions of climate change due to diverse processes. There are many recognised problems with simple integrated assessment model representation of the impacts of climate change and especially tipping points 5,36–40. Most fundamentally, they tend to conflate weather variability with long-term climate changes 5,41 and apply spatial analogues across different geographic and temporal scales. This false equivalence ignores the fact that climate change involves systematic shifts in baseline weather patterns over multi-decadal time spans, with profound economic implications relating to infrastructure, supply chains, and more. Much of the existing work often assumes that human economies are highly adaptable to climate change, implying minimal economic impacts, but this method does not fully account for the possibility of large-scale, nonlinear changes like tipping points, and the potential for successive climate regime shifts in different directions may challenge the presumption of high adaptability, testing the limits of social and economic resilience even with humanity's capacity to adjust.
As an alternative to the use of IAMs, the human climate niche 7 takes an ecological approach to quantify how human population density depends on mean annual temperature (MAT) and mean annual precipitation (MAP). This concept, analogous to the ecological niche occupied by a species 42, has been used to characterize the climatic conditions suitable for human societies 7,8. The niche is found to be conserved over centuries 8, with similar relationships between climate and human population density, crop yields, livestock distributions, and economic productivity observed across different historical periods. As such, the climate niche provides a simple, integrated way of starting to consider the impacts of temperature and precipitation on human habitability. However it does not capture the impact of sea level rise and lacks explicit representation of the effects of seasonality or extremes.
We examine the impacts of 2.5°C global warming and AMOC collapse independently and then in combination. We show how these successive large-scale shifts in global climate conditions in conflicting directions would challenge humanity's capacity to adapt in many regions.
Projected Change
Our study examines HadGEM3-GC2 model scenarios portraying the impacts of 2.5°C global warming under SSP1-2.6, isolated AMOC collapse, and AMOC collapse following 2.5°C global warming. The latter scenario is produced by a linear combination of the previous ones and as such misses some key non-linear effects. For instance, the simulated cooling in high northern latitudes will be overstated due to more strongly amplified sea-ice growth in the control climate than in a warmer world where much ice has been lost. The scenario of AMOC collapse alone remains theoretical, requiring additional warming or freshwater forcing for a full collapse. Our previous work considered the 2015 population distribution under the 1960–1990 mean climate as a baseline 8 and the 1980 population distribution under the 1960–1990 mean climate as the reference state 7. Here our analysis of changes in the human climate niche is based on deviations from a control scenario representative of the 1990-2020 climatological mean and uses the 2020 population. The 1990-2020 mean better represents recent climatological conditions that human systems are adapted to and is more relevant for assessing deviations caused by an AMOC collapse that may occur in the upcoming decades.
Under SSP1-2.6, our model projects a 2.5°C warming by the end of the century. As shown by Xu et al. (2020)8 and Lenton et al. (2023)7 this warming will disproportionately impact the tropics. The projected geographical shift of favorable MAT-MAP conditions over the rest of the century is substantial, with optimal conditions for humans projected to move poleward and with the worst effects in the tropics and some extra-tropical regions (Figure 1A).
While AMOC collapse in isolation remains a hypothetical scenario, it provides valuable insights into the isolated impacts of this tipping point. Our simulations show that AMOC collapse would cause striking geographical shifts in the human climate niche (Figure 1B), disproportionately affecting the Northern Hemisphere, particularly Europe. In these simulations, Europe and higher northern latitudes experience significant cooling of up to 8°C, while the Southern Hemisphere experiences little to no warming (Figure S1A). Furthermore, most of the Northern Hemisphere experiences drying, except for North America, which becomes slightly wetter on average (Figure S1B). Notably, there are disruptions to the Indian summer monsoon and significant drying in the Amazon basin, highlighting potential cascading effects. On a global scale, regions in the tropics and south of the equator become more habitable on average, while habitability decreases in the Global North, particularly in higher latitudes (Figure 1B). Sub-Saharan Africa, as well as Central and South America, see the largest gain in habitability.
In the combined scenario of 2.5°C global warming followed by AMOC collapse (Figure 1C), the average change in population-weighted human experienced temperature is +0.7°C. This scenario exhibits contrasting temperature responses between higher northern latitudes and the tropics and Southern Hemisphere (Figure S1A). Precipitation patterns also differ, with reduced precipitation in the Northern Hemisphere and increased precipitation in the Southern Hemisphere. Europe emerges as the most negatively impacted region, experiencing both cooling and reduced precipitation. North America becomes mostly more suitable due to a negligible change in temperature from the combined effects of 2.5°C and AMOC collapse, accompanied by a general increase in precipitation. Large swaths of South America, particularly Brazil, become less suitable due to amplification of two factors under the combined 2.5°C warming and AMOC collapse scenario - a reduction in precipitation and increase in temperature making the region hotter and dryer. Relative to the warming only scenario, suitability in most of sub-saharan Africa increases due to increases in rainfall, while equatorial Africa and Northern Africa - where the impact is dominated by the temperature increase in the warming scenario - see a marked decrease in suitability.
Serial Shifts
Our analysis shows that while a 2.5°C global warming leads to widespread temperature increases across both hemispheres, the shutdown of the AMOC triggers divergent temperature responses (Figure S1A). Initially, extratropical latitudes in the Northern Hemisphere warm, followed by widespread cooling due to AMOC shutdown. In contrast, the tropics and the Southern Hemisphere experience widespread warming. These serial shifts in the climate niche highlight the complexity of the impacts, especially in regions like the Northern Hemisphere extratropics, where opposing temperature changes occur.
While a few regions may experience amplified effects of AMOC collapse on the human climate niche compared to warming alone (Figure 2), overall, AMOC collapse predominantly shifts the climate niche in the opposite direction of warming. This means that where warming has a positive impact on the MAT-MAP niche, AMOC collapse will have a negative impact, and vice versa. While some regions like northern Europe exhibit amplified cooling under the combined scenario compared to AMOC collapse alone, the predominant pattern is one of divergence rather than amplification of effects. For instance, parts of Canada experience warming of 2-4°C under 2.5°C, but this is almost entirely negated by the 2-6°C cooling from an AMOC collapse. This realignment of temperature regimes from one extreme to the other within a relatively short period poses significant challenges for adaptation.
The serial reversals are even starker when examining the spatial context of precipitation changes. The US Southwest is projected to experience over 20% drying under 2.5°C warming, potentially straining water resources. However, an AMOC collapse could increase rainfall by 10-20% in this region, rendering any incremental adaptations for drought conditions maladaptive. Globally, changes in suitability are compounding for 27% of the global population, while changes in suitability oppose each other for 73% of population.
Human Exposure
To illustrate climatic shifts and human exposure changes, we analyze the MAP-MAT niche relative to the 2020 population distribution43 (Figure S3). Under a scenario of 2.5°C global warming combined with AMOC collapse (Figure S3A), reduced precipitation decreases suitability in densely populated regions like western Europe, eastern North America, the Amazon, and western Africa (Figure 3). Conversely, suitability increases in less populated regions like southern Africa, parts of South America, and south Asia. Similar trends occur with MAT changes (Figure S3B), showing suitability increases in sparsely inhabited higher latitudes but decreases in densely populated temperate and tropical zones. Overlaying niche shifts with population maps reveals significant human exposures to declining suitability conditions in major population clusters, such as urban zones in Southeast Asia and India, due to the compounding impacts of warming and AMOC collapse.
Projections of climate niche shifts under a global warming followed by AMOC collapse scenario unveil winners and losers (Figure 4). Under 2.5°C global warming alone, the tropics and Southern Hemisphere face substantial declines in suitability compared to higher northern latitudes. In a combined scenario, countries across all regions experience both increases (Figure 4A) and decreases (Figure 4B) in suitability, with Europe emerging as a significant loser. Some countries, like Zimbabwe and Botswana, experience compounding gains initially due to favorable shifts following prior warming, reinforced by AMOC collapse. However, this may be short-lived if subsequent warming exceeds habitability thresholds, particularly affecting regions in the Global South.