Home Climate The Atlantic Circulation Is Probably Slowing. That Still Does Not Mean Collapse...

The Atlantic Circulation Is Probably Slowing. That Still Does Not Mean Collapse Is Near.

A new synthesis argues that the available evidence now supports a past and continuing slowdown of the Atlantic Meridional Overturning Circulation, or AMOC. That conclusion is serious. It is also narrower than the claim that the Atlantic circulation is about to collapse.

The distinction matters because the AMOC is not measured by one instrument, one temperature map or one dramatic indicator. It is a basin-scale system whose behavior is reconstructed from direct observations, deep-ocean measurements, model comparisons and historical proxies. Those lines of evidence increasingly point in the same direction, but they do not yet provide a complete historical record or a reliable timetable for any threshold event.

The most defensible conclusion is therefore this: the AMOC is probably weakening, and human-caused warming is part of the explanation. The magnitude, timing and future rate of that weakening remain uncertain, and an imminent collapse has not been demonstrated.

What the new claim says

In a November 17 preprint, climate scientists Stefan Rahmstorf and Levke Caesar reviewed several strands of AMOC research and concluded that the balance of evidence strongly supports past and ongoing slowing in response to global warming. The authors also emphasize that continuous monitoring covers only roughly the past two decades and that historical reconstructions have different limitations.

That makes the paper a synthesis, not a new measurement. Its contribution is to ask what the evidence says when different methods are considered together.

The result is more consequential than the claim that the AMOC is stable because one portion of the record temporarily improved. It is also less dramatic than saying the circulation is nearing a sudden shutdown. Those are different propositions, and the evidence for one should not be used as proof of another.

What the AMOC is—and is not

The AMOC is a large Atlantic system that transports heat, salt, carbon and nutrients. Warm surface water moves northward, while colder, denser water returns southward at depth. The system influences regional sea level, precipitation, climate patterns and marine ecosystems. NOAA’s explanation of the AMOC describes it as part of the broader global overturning circulation.

It is not identical to the Gulf Stream. The Gulf Stream is an important component of the Atlantic circulation, but the AMOC includes multiple layers and pathways across the basin. A change in the Gulf Stream, North Atlantic temperature or coastal sea level may provide information about the larger system without directly measuring the entire overturning circulation.

That distinction has been relevant to WBC coverage before. In Norman Hargrove’s November 6 report on AMOC-related coastal-flood prediction, the question was whether circulation variability could help forecast Southeast Coast flood frequency several years ahead. The present question is different: what do scientists know about the circulation itself?

The direct record is important—and short

The strongest evidence comes from the RAPID observing system, which has measured the AMOC near 26.5 degrees north since 2004. A 2025 study in Geophysical Research Letters estimated a decline of approximately 1.0 Sverdrup per decade from 2004 through 2023, with a reported 90 percent confidence interval of roughly 0.4 to 1.6 Sverdrups per decade. A Sverdrup is a unit of ocean transport equal to one million cubic meters of water per second.

This is important because the estimate is based on direct ocean observations rather than on a historical reconstruction alone. But the record is only about twenty years long, which is brief compared with the timescales of ocean circulation and natural climate variability.

The direct record is also not a straight line downward. The circulation weakened sharply early in the RAPID record, partly recovered from approximately 2009–10 through 2018, and then returned to a weakening pattern. The same data can therefore support two observations at once: a statistically detectable decline over the full period and substantial shorter-term variation within it.

That is not a contradiction. It is the central measurement problem.

Why a temporary recovery does not settle the question

A 2024 Nature Communications study by NOAA and collaborators examined the apparent pause in AMOC weakening after the early 2010s. Using climate-model simulations, surface-forced ocean models and direct observations, the researchers concluded that anthropogenic forcing contributed to weakening in the 2000s while naturally driven variability temporarily opposed part of that trend.

The study’s conclusion is model-dependent. It does not provide a perfect separation between human influence and natural variability. It does, however, offer a physically plausible explanation for why a short direct record might appear less alarming than longer-term climate projections.

Climate systems do not respond to greenhouse warming as a single uninterrupted curve. Winds, atmospheric patterns, ocean mixing and other processes can temporarily strengthen or weaken parts of the circulation. A pause can therefore obscure a forced trend without disproving it.

More than one line of evidence

The case for weakening does not rest on RAPID alone. Researchers are also examining deep-ocean temperature and salinity, surface-temperature patterns, historical reconstructions and comparisons between observations and climate models.

The 2025 RAPID analysis found that deep North Atlantic water properties may provide a less noisy indicator of low-frequency AMOC change than the full circulation estimate. A simple model using deep hydrographic measurements reproduced a substantial share of the low-frequency signal. That does not independently establish the complete historical trajectory, but it gives scientists another way to distinguish longer-term change from short-term, wind-driven fluctuations.

Climate models also generally project AMOC weakening under continued greenhouse-gas warming. The precise rate varies among models and emissions scenarios, and the possibility of abrupt change is more difficult to evaluate than the expectation of gradual weakening.

The evidence streams are not fully independent. Several rely on models, assumptions about how surface patterns reflect deep circulation or relatively short observational records. But their convergence makes the case stronger than any single proxy or reconstruction.

What the evidence establishes

Taken together, current research supports several conclusions:

  • The AMOC varies substantially from month to month and year to year.
  • The RAPID record contains a statistically detectable decline from 2004 through 2023.
  • The direct record includes reversals and periods of apparent recovery.
  • Natural variability can temporarily offset anthropogenic weakening.
  • Deep-ocean observations may help identify the longer-term signal.
  • Continued greenhouse warming is expected to weaken the AMOC.
  • No abrupt collapse has been observed in the current record.

The 2025 signal-to-noise analysis estimated that, if the observed trend continued, certain statistical thresholds described as “unfamiliar” and “unknown” might emerge around the 2040s and 2060s. Those are detection thresholds within a particular analytical framework. They are not predicted collapse dates.

This is precisely where public discussion often goes wrong. A statistically detectable trend is not the same thing as a tipping point. A projection of weakening is not a forecast that the Gulf Stream will suddenly stop. And a short-term rebound is not proof that the long-term risk has disappeared.

What remains uncertain

Scientists still cannot determine with high confidence how much the AMOC has weakened over the full industrial era. Continuous direct monitoring began only in the early 2000s. Earlier records depend on indirect indicators such as sea-surface temperatures, sediment data and ocean properties, and those indicators do not all measure the same part of the circulation.

Models also differ in their treatment of boundary currents, deep-water formation, ocean mixing, freshwater input from ice melt and small-scale eddies. These structural differences affect both historical attribution and future projections.

There is genuine disagreement here, but it is narrower than the public debate often implies. Researchers disagree over the size of the historical decline, the reliability of particular fingerprints, the contribution of natural variability and the probability of abrupt thresholds under realistic warming scenarios.

They do not disagree that the AMOC exists, that greenhouse warming affects ocean circulation or that the direct record is too short to answer every question.

Why continued observation is part of the science

The AMOC illustrates a broader problem in climate research: some of the most consequential variables are difficult to observe directly over the timescales that matter.

Climate models provide long simulations. Moorings provide a shorter but physically rich record. Reconstructions extend the record backward while adding assumptions. Satellite, coastal and hydrographic observations provide additional clues without measuring the entire circulation directly.

That means monitoring is not a task to complete after the science is settled. Monitoring is how the science becomes more settled. The value of the RAPID array, the related MOCHA and Western Boundary Time Series programs, and other observing systems lies not only in producing another annual number. It lies in creating a record long enough to separate forced change from natural variation.

For institutions, the question is therefore practical as well as scientific: are these observing systems funded and maintained long enough to detect a meaningful change before the consequences become obvious through coastal flooding, fisheries disruption or regional climate shifts?

The U.S. relevance is real—but not a city forecast

AMOC change matters to the United States through possible effects on regional sea level, precipitation, storm tracks, marine ecosystems and the background conditions that shape coastal flooding.

But the current research does not predict a particular flood in a particular American city. It does not establish that a specific hurricane season or winter was caused by AMOC change. Nor does it allow a basin-scale estimate to be placed directly into a local coastal forecast without accounting for tides, winds, land movement, geography and infrastructure.

The immediate implication is more modest and more useful: planners should not assume that the background ocean state is stationary. A circulation that is slowly changing—or changing in ways partly hidden by natural variability—complicates long-range projections and near-term risk planning.

Uncertainty does not make that information useless. It determines how carefully it should be used.

The conclusion the evidence has earned

The Atlantic overturning circulation is probably weakening, and human-caused warming is part of that explanation. The direct record, however, is short and noisy. It does not yet tell us exactly how much weakening has occurred, how quickly it will continue or whether the system is approaching a threshold.

That is enough to justify continued observation and cautious planning. It is not enough to justify declaring an imminent collapse.

The useful scientific position lies between complacency and catastrophe: a measured long-term risk deserves attention even when the dramatic claim has not yet been earned.

The November 2026 Rahmstorf and Caesar preprint, the 2025 RAPID analysis and the 2024 NOAA-led attribution study do not eliminate disagreement. They make the disagreement more precise—and that is progress.