The Atlantic Meridional Overturning Circulation is one of the great drivers of Earth’s climate. It moves heat and salt between hemispheres, stabilizing temperatures, and anchoring rainfall belts that feed continents. The Amazon rainforest is another giant system, storing carbon, recycling water, and generating its own climate stability through evaporation. These two systems are now shown to be directly linked in collapse risk. A new study has quantified the probability that if the Atlantic circulation fails, the Amazon will also fail. This is not a scenario of isolated damage. It is a chain reaction.
The study uses a coupled model of the AMOC and the Amazon and applies a rare-event algorithm that allows collapses to be sampled efficiently. In a normal simulation the collapse of either system might occur only once in millions of years of virtual time. By biasing trajectories, the algorithm reproduces enough collapses to calculate meaningful probabilities. The result is clear. In the central Amazon basin, collapse into a savannah state does not occur unless the AMOC collapses first. The drying effect of circulation failure is a necessary condition. Without it the rainforest holds. With it, collapse follows.
In regions already stressed by drought, particularly in the southern Amazon, the model shows the forest tips regardless of what happens to the ocean. Fire and rainfall shortages are sufficient. But in the wetter core regions the system is stable until the AMOC weakens. Once that occurs, rainfall declines, dry seasons lengthen, and fires become able to burn deep into forest that was previously resistant. When that happens, the rainforest moves from stability into an irreversible savannah state. The probabilities are not speculative. They are calculated.
This quantification is critical because it shows the Amazon cannot be considered in isolation. Its future is conditional on the Atlantic circulation. In the model, probabilities of collapse for the central basin are close to zero without an AMOC failure. With AMOC collapse, they rise sharply. The transition from rainforest to savannah takes place in less than two centuries, with fires acting as accelerants once rainfall has been reduced. The shift is abrupt and irreversible in the model’s framework.
The AMOC itself is not stable. Observations already suggest a weakening trend. Its flow is maintained by contrasts of temperature and salinity between the tropics and the poles. Melting ice in Greenland and the Arctic adds freshwater that disrupts this balance. Models have long shown that once thresholds are crossed, the circulation can stop. In paleoclimate records, such stoppages are associated with abrupt shifts in global weather, droughts in the tropics, and sudden cooling events in the North Atlantic. The new study connects this ocean collapse directly to the fate of the Amazon. It shows the probability chain: if one fails, the other fails.
The Amazon is not simply a forest at risk from chainsaws and wildfires. It is a dependent system, tied into the circulation of the Atlantic. The collapse of one leads to the collapse of the other, and with the Amazon gone, billions of tons of carbon are released into the atmosphere. That in turn heats the planet further, intensifying the very processes that strain the AMOC. A loop is created, where one collapse feeds another. The study is not describing vague threats. It is mapping a chain of cause and effect, backed by numbers.
Fire dynamics are central to this process. In the model, wildfires are represented by heavy-tailed distributions, capturing the fact that most years see limited burning but rare events wipe out vast areas at once. In the dry south these fires alone can drive collapse. In the wetter north they normally cannot. When the AMOC weakens and rainfall declines, however, fire jumps into new power. Once tree cover is reduced below a threshold, open grasslands spread, fires feed on them, and recovery is blocked. The model shows that this threshold is reached quickly once circulation weakens.
There is no comfort in the timescales. In the simulations, once AMOC weakening begins, rainforest collapse follows within two centuries. In some southern regions the transition to savannah occurs in decades. This is a timeframe within human history, not geological epochs.
This is not the first time Earth has faced linked system failures. The paleoclimate record shows abrupt shifts where multiple subsystems changed together. One of the most striking examples is the Laschamp event about 42,000 years ago, when Earth’s magnetic field collapsed to a fraction of its normal strength. The poles drifted, radiation levels at the surface rose, and climate instability followed. Ice cores and sediments record circulation shifts, droughts, and vegetation changes. The mechanism is still debated, but the association is clear. Weakening of the field coincided with instability in atmosphere and ocean.
Today the magnetic poles are moving faster than ever recorded, and the field itself is weakening. While this was not part of the new study, the historical precedent raises concern. If a magnetic weakening once aligned with circulation collapse and ecological disruption, then current drift could act as another destabilizer in the present. Increased radiation and altered atmospheric chemistry from a weaker field can shift cloud cover and precipitation. For a system like the AMOC already under stress, that may be enough to move it closer to failure.
Seen this way, Earth’s systems cannot be separated into silos. The AMOC, the Amazon, and the magnetic field may seem unrelated, but history suggests they interact. The new study confirms a cascade between ocean and forest. The magnetic record shows that when the field weakens, circulation and climate are disturbed. If these align again, the chain of destabilization could be faster and more severe than expected.
The risk is magnified because the Amazon’s collapse does not stop at South America. Loss of the forest changes rainfall globally. Croplands in Africa, Asia, and North America depend on the planetary circulation patterns that the Amazon helps sustain. Collapse also means a surge of carbon into the atmosphere, accelerating warming everywhere. When coupled with an AMOC shutdown, which would bring European winters, tropical drought, and monsoon disruption, the picture is global destabilization on multiple fronts.
The study also shows that resilience is not unlimited. The central Amazon appears stable under current rainfall, but that stability vanishes once AMOC weakness removes moisture. Fire thresholds are crossed, feedbacks lock in, and the system shifts state. The probability analysis shows that what seems unlikely at first becomes inevitable once one condition is met. This is how cascades operate. One failure forces another.
The use of rare-event algorithms matters because it replaces speculation with quantification. Traditional models cannot simulate enough collapses to give reliable probabilities. The algorithm used here does. It tracks observables across thousands of biased trajectories and builds statistical pictures of collapse likelihoods. The numbers show that collapse of the Amazon in the central basin is conditional. That conditionality makes the relationship precise. The Amazon holds until the AMOC fails. After that, it falls.
This is not activist rhetoric. It is a numerical result from a peer-reviewed model. The conclusion is not a call to action but a map of risk. It demonstrates that the systems of Earth are coupled, that collapse can be calculated, and that probabilities are high once conditions align.
Past records already show how quickly Earth can change. Ocean circulation has collapsed before, producing sudden droughts and abrupt climate swings. The Amazon has shifted between forest and savannah states in earlier climates. The magnetic field has weakened dramatically in the past, with disturbances aligning with global instability. The present moment brings all three of these stresses together at once, raising the risk that a breakdown in one system will accelerate the others.
The convergence of these conditions is what makes the new study so important. It confirms that collapse does not occur in isolation. It spreads. It quantifies how one system drags another down. And it arrives at a moment when signals of instability are already visible in real observations. The AMOC is weakening. The Amazon is burning. The poles are shifting.
The study calculates a cascade from ocean to forest. The broader record shows that Earth’s systems rarely fail alone. Together they create compound risks that magnify one another. This is the picture now emerging. Not a single tipping point, but a chain of them. Once one slips, the rest follow.
Source:
Jacques-Dumas, V. & Dijkstra, H. A. (2025). Quantification of the cascading tipping probability from the AMOC to the Amazon rainforest. arXiv preprint arXiv:2508.13383. https://doi.org/10.48550/arXiv.2508.13383
