Nature Claims Their Circulation Is Decreasing

Guest Post by Willis Eschenbach (@WEschenbach on X)

Let’s get something straight: this new study entitled Equatorial Atlantic mid-depth warming indicates Atlantic meridional overturning circulation slowdown, hot off the Nature presses and already getting the full catastrophe-trending treatment, claims that mysterious “mid-depth warming” in the equatorial Atlantic—between 1,000 and 2,000 meters down—is the long-sought “fingerprint” proving that the Atlantic Meridional Overturning Circulation (AMOC) has been slowing since the 1990s.

According to the authors, this subsurface warming is more “reliable” than surface measurements, cleaner than the noise-soaked sea surface temperatures everyone’s been arguing about for years, and—conveniently—it fits right into the climate alarm narrative that the AMOC is heading for catastrophic collapse. If you buy this story, you’re supposed to panic now, because the collapse might already be underway, the tipping point looms, and we’re all headed for Day After Tomorrow ice-age-in-a-week horror.

There’s just one problem: the whole edifice is built on model outputs, selective data interpretation, circular reasoning, and a studied refusal to acknowledge that the AMOC might be doing just fine.

Let’s start with the foundation. The authors run the MIT General Circulation Model (MITgcm) using forcing data pulled from CMIP6 climate model simulations under a quadrupled CO₂ scenario—because apparently we needed more hypotheticals stacked on hypotheticals. Of course, there’s not enough fossil fuels on the planet to lead to a quadrupled CO₂ level, but pay no attention to the man behind the curtain …

Then they impose abrupt changes in wind stress, sea surface temperature, and sea surface salinity, taken straight from models that already assume an AMOC slowdown, and—surprise!—the model dutifully spits out a slowdown and equatorial warming. It’s the scientific equivalent of asking a parrot to repeat what you just taught it, then claiming the parrot discovered language.

The “experiments” (BUOY, BUOY-NA, SSS-NA) are all variations on a theme: fiddle with surface forcing, watch the model respond in pre-cooked ways, then cherry-pick the results that support the thesis. Want mid-depth warming? Easy—force in North Atlantic freshening (which the CMIP6 models already baked in as part of their AMOC-weakening scenarios), sit back, and let “baroclinic Kelvin waves” do the propagation magic. It’s not an independent test of the hypothesis; it’s a self-fulfilling simulation.

The study’s logic runs in a perfect circle.

Step one: assume the CMIP6 models (which all project AMOC weakening under CO₂ forcing) are correct.

Step two: use those same models’ outputs as forcing for your ocean model.

Step three: when your ocean model shows equatorial warming, declare that this warming is the “fingerprint” of AMOC slowdown.

Step four: find warming in the observational data.

Step five: conclude the AMOC must be slowing because you found the “fingerprint.”

At no point do they test whether the warming could be driven by anything other than AMOC changes—such as, oh, I don’t know, local wind patterns, El Niño teleconnections, subsurface advection from other basins, or the natural variability that the equatorial Atlantic is famous for.

The authors triumphantly point to warming trends in gridded datasets (WOA, Argo, IAP, Ishii, EN4) showing 0.14°C of mid-depth warming from 1960 to 2020. Sounds impressive until you realize: (1) pre-1980 subsurface data in the equatorial Atlantic is sparse, riddled with gaps, and heavily interpolated; (2) Argo floats only started coverage in 2004, so the “robust” trend is mostly a post-millennial artifact; and (3) the “signal-to-noise” ratio they’re so proud of is high precisely because mid-depth waters have high thermal inertia—meaning the warming could be a slow, cumulative response to anything, not necessarily AMOC changes.

Moreover, the study conveniently ignores a major new finding published just this year in Nature: an independent reconstruction using air-sea heat flux data concluded that the AMOC has not declined over the past 60 years and is more stable than previously thought. That study, from Woods Hole, directly contradicts the premise here and suggests that relying on SST proxies (the very thing this new study is trying to replace) was always the problem—not because the AMOC is weakening, but because SSTs are lousy proxies in the first place.

Yes, baroclinic Kelvin waves are real, and yes, they can propagate warming signals from the subpolar North Atlantic down the western boundary and across the equator. The authors make a big deal about “rapid adjustment” and “dynamical signals,” invoking vertical mode decomposition and wave speeds to lend scientific gravitas. But here’s what they don’t tell you: the equatorial Atlantic is also subject to wind-driven upwelling, Atlantic Niño/Niña variability, off-equatorial Rossby wave adjustments from the northwestern tropical Atlantic (not the subpolar regions), and teleconnections from the Pacific. Multiple studies have shown that equatorial Atlantic warming is more consistently linked to local and regional processes—especially wind stress curl anomalies associated with the Atlantic Meridional Mode—than to remote AMOC forcing.

The fact that their passive tracer experiment (BUOY-NA-passive) shows minimal mid-depth warming from mean circulation alone doesn’t rule out these other mechanisms—it just means their model setup didn’t include them.

The authors claim the warming “emerged” from natural variability around 2001, with a signal-to-noise ratio exceeding their arbitrary threshold of 4 (99.9% confidence). But “time of emergence” is a model-dependent metric that depends entirely on how you define “signal” and “noise.” They define signal as the linear trend and noise as detrended variability—which guarantees that any long-term drift, regardless of cause, will eventually “emerge.” It’s a statistical trick that makes any monotonic change look significant if you wait long enough.

The study buries or ignores inconvenient facts. For instance:

• The subpolar North Atlantic “cold blob” (the supposed surface fingerprint of AMOC slowdown) has been highly variable and even disappeared in recent years.
• Direct AMOC measurements from the RAPID array show large interannual variability with no clear long-term trend since 2004.
• A 2025 Nature study entitled “Continued Atlantic overturning circulation even under climate extremes” using 34 CMIP6 models found that the AMOC does not collapse even under extreme greenhouse forcing and freshwater input—it just weakens and stabilizes at a lower level.
• Multiple lines of evidence (including paleo-proxies and modern reanalysis) suggest the AMOC has been relatively stable over the instrumental period, with most of the “weakening” claims based on noisy, short-term SST datasets.

This isn’t dispassionate science. It’s model-driven alarmism dressed up in the language of “fingerprints” and “dynamical mechanisms.” The authors want you to believe that a few tenths of a degree of subsurface warming—detectable only after heroic interpolation and smoothing of sparse data—is the smoking gun for an AMOC on the brink. They invoke “tipping points,” cite scary projections of 2 Sv decline since the 1950s, and darkly hint that we’re already past the point of no return.

But the data don’t support it, the models are unreliable, and the alternative explanations are everywhere if you’re willing to look. The equatorial Atlantic warms for lots of reasons—most of them having nothing to do with the AMOC. This study cherry-picks one mechanism, builds an entire edifice on model outputs that assume what they’re trying to prove, and then declares victory because gridded datasets (with all their uncertainties) show a trend.

If you want to monitor the AMOC, measure it directly. Don’t use proxy tea-leaves from a noisy, complex region, filtered through biased models, and call it a “superior fingerprint.” This is climate science at its worst: models all the way down, confirmation bias all the way up, and a conclusion tailor-made for the next round of climate crisis headlines. The AMOC is probably fine. The real collapse is in scientific skepticism.

Seems like as the climate alarmism grift collapses, the claims get more outrageous. Meanwhile, here’s the situation in the real world …

My very best to everyone,

w.

PS—As is my wont, I ask that when you comment you quote the EXACT WORDS you are discussing. I can defend my words. I can’t defend your interpretation of my words. Thanks.