For two decades, instruments bolted to the Atlantic seafloor have recorded the pulse of one of Earth’s most important ocean currents. Now, a peer-reviewed study published in April 2026 in Science Advances confirms what those instruments have been signaling: the deep Atlantic circulation has weakened at every monitoring station where scientists keep continuous records, across four sites stretching from the subtropics to the mid-latitudes along the ocean’s western boundary.
The finding, based on direct physical measurements rather than climate model projections, sharpens a long-running scientific debate about whether the Atlantic Meridional Overturning Circulation, commonly called the AMOC, is losing strength in ways that could alter weather patterns across two continents and push sea levels higher along the U.S. East Coast.
What the instruments recorded
The study in Science Advances drew on data from four deep-ocean monitoring arrays positioned between 16.5 degrees North and 42.5 degrees North latitude. Each array uses pressure-equipped inverted echo sounders, known as PIES, anchored to the ocean floor to track changes in the Deep Western Boundary Current, the cold, dense river of water that flows southward far below the surface as part of the global ocean conveyor.
What makes the result striking is its consistency. The decline showed up not at one location but at all four, each with its own instrumentation and local oceanographic quirks. That coherence across roughly 2,900 kilometers of ocean makes it difficult to dismiss the trend as a local anomaly or a quirk of a single instrument.
Much of the underlying data comes from NOAA’s Western Boundary Time Series project, operated by the Atlantic Oceanographic and Meteorological Laboratory. The program tracks deep currents with PIES arrays and monitors the Florida Current through a submarine cable stretched across the Florida Straits. Shipboard surveys and pressure gauges at the Abaco section near 26.5 degrees North round out one of the longest continuous records of deep Atlantic flow in existence.
A separate study points to the deepest layer
A closely related analysis, published in Nature Geoscience, focused on the very bottom of the water column: the abyssal limb fed by Antarctic Bottom Water, which forms around Antarctica and creeps northward along the seafloor. That study found a 12 percent decrease (plus or minus 5 percent) in northward Antarctic Bottom Water transport at 16 degrees North between 2000 and 2020. In concrete terms, that translates to roughly 0.35 sverdrups lost from a baseline of about 2.40 sverdrups. (A sverdrup equals one million cubic meters of water per second, a unit that conveys the staggering volume these currents move.)
NOAA’s summary of that research also flagged a deep warming rate of approximately 1 millidegree Celsius per year at those depths. That sounds vanishingly small, but compounded over decades across vast stretches of the abyss, it represents a significant accumulation of heat in a part of the ocean that was long assumed to be nearly unchanging.
Taken together, the two studies paint a coherent picture: the deep Atlantic is losing strength not at a single point but across a wide band of the western boundary, from the tropics to the mid-latitudes.
Why this does not equal an AMOC collapse
A weaker deep current is not the same thing as a collapsing AMOC, and scientists are careful to draw that distinction. The AMOC is a three-dimensional system. Cold, dense water sinks in the subpolar North Atlantic and flows south at depth while warm surface water travels north, carrying heat from the tropics toward Europe. Measuring only the deep western leg captures part of the story. Interior pathways and recirculation gyres can redistribute water in ways that either amplify or offset what the boundary current shows.
A 2021 analysis using data from the Overturning in the Subpolar North Atlantic Program (OSNAP) mooring array, published in Nature Communications (vol. 12, article 3408, doi:10.1038/s41467-021-23350-2), cautioned against treating deep western boundary density changes as a direct stand-in for total overturning strength. On the time scales currently observed, the relationship between deep boundary signals and the full subpolar overturning is more complicated than a simple one-to-one proxy.
Scientists have not yet produced a unified synthesis that stitches together all four Western Boundary Time Series sites with OSNAP observations, the RAPID array at 26.5 degrees North (the longest-running full-depth AMOC monitor), and upper-ocean data into a single, long-term AMOC estimate. Until that integration happens, the deep-current decline is best understood as a critical piece of a larger puzzle, not the whole picture.
What it could mean for coastlines and climate
For communities along the U.S. East Coast, the practical question is whether a weakening deep current translates into higher water levels at the shore. The short answer: it could, but the connection is not straightforward.
Regional sea-level rise is shaped by a mix of forces: thermal expansion of warming seawater, melting land ice from Greenland and Antarctica, vertical land motion (parts of the mid-Atlantic coast are still sinking from the rebound effects of the last ice age), and shifts in ocean circulation. A weaker Deep Western Boundary Current can alter pressure gradients and surface currents in ways that raise coastal water levels, but quantifying that contribution requires models that faithfully represent both the deep and upper ocean.
NOAA’s own summary of the abyssal research mentions a potential contribution to East Coast sea-level rise, though translating a 0.35-sverdrup decline in deep transport into centimeters of coastal flooding involves modeling steps that carry their own uncertainties.
On the climate side, a sustained AMOC slowdown could cool parts of the North Atlantic, shift tropical rainfall belts, and alter storm tracks that affect both North America and Europe. Many climate models project exactly this kind of weakening under continued greenhouse gas emissions, though they disagree on the pace and magnitude. The new observational data does not resolve those model disagreements, but it confirms that the deep Atlantic is changing on human time scales, not just geological ones.
Open questions on drivers and transmission speed
The 20-year observational window is long by oceanographic standards but short relative to the multi-decadal and centennial cycles that can shape deep ocean behavior. Whether the observed weakening continues, stabilizes, or reverses will depend on additional years of data and on new instruments that can fill gaps between existing arrays.
One key unknown is what is driving the decline. Freshwater pouring off the Greenland ice sheet could be making surface waters in the subpolar North Atlantic less dense, reducing the sinking that powers the deep limb. Changes in Antarctic Bottom Water formation near the Southern Ocean could be starving the abyssal layer from the other end. Disentangling these drivers is an active area of research, and the answer likely involves both hemispheres.
Another open question is speed of transmission. The deep western boundary current interacts with overlying water masses through mixing and shear, but how quickly a weakened abyssal flow affects intermediate and surface layers is not well constrained. Some modeling work suggests the signal could take decades to fully reach the upper ocean; other studies point to localized feedbacks on continental slopes that might transmit changes faster in certain regions.
For now, the most defensible conclusion is narrow but significant: the deep Atlantic circulation along the western boundary has slowed over the past two decades in a pattern that is consistent across multiple independent monitoring sites and difficult to attribute to natural variability alone. How far that change propagates through the rest of the ocean, and how strongly it shapes future climate and sea-level patterns, remains one of the central open questions in physical oceanography.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
