The whole of the deep Atlantic is fed from a patch of sea you could cross in a few days' sailing.
Level and unhurried, pitched at the volume of figures being read aloud in a room with the sea nowhere near it. Nothing about this circulation makes a noise anybody has ever heard; the water is inferred, and the voice doing the inferring stays flat, patient and exact. One proportion is stated and then left standing, with a long pause behind it. The tempo is that of a measurement being checked rather than a discovery being announced.
Ocean circulation is described with far more confidence than it is watched: two independent engines that are routinely mistaken for one, a circuit of roughly a thousand years, and — as the whole empirical basis for saying it is changing — twenty-two years of continuous measurement at a single latitude.
Where the ocean puts its water decides a fishery off Peru, the deep water filling every basin, and the terms of a long-running argument about European winters. It is also a system described with far more confidence than it is watched, and nearly every time an array has actually gone into the water something has moved. The engine room of the North Atlantic turned out to be on the other side of Greenland. The largest current on the planet came out about thirty per cent above the value climate models had been tuned against. The most repeated claim about what the Gulf Stream does for northwest Europe did not survive its first serious test, nearly a century and a half after it was made. None of that is scepticism about the physics. It is the difference between a mechanism and a measurement, and here that difference is visible while it operates.
Everything contested about this ocean lives in the distance between a mechanism and a measurement. Two figures that disagree here are usually two definitions rather than two oceans.
This subject runs on more than one time axis. The chapters are not one sequence.
The reader stops picturing one circulation and starts sorting every later fact into one of two drives that cannot do each other's work.
Ekman Transport and the Ekman Spiral · Thermohaline Circulation · North Atlantic Deep Water (NADW) · Antarctic Bottom Water (AABW) · Coastal and Equatorial Upwelling
The shape of every ocean map stops looking like a consequence of coastlines and becomes a consequence of the planet being round.
North Atlantic Subtropical Gyre · Western Intensification of Wind-Driven Gyres · Gulf Stream · Kuroshio · Vagn Walfrid Ekman · Henry Stommel · Sverdrup (Sv)
The reader sees that the images of this ocean arrived before the measurements did, and that each one went on setting expectations long after it should have stopped.
Matthew Fontaine Maury · Wallace Broecker · Gulf Stream · Thermohaline Circulation · Meridional Ocean Heat Transport
The mechanism survives and the geography does not. The reader learns that a textbook location can be wrong for decades without the physics behind it being wrong at all.
North Atlantic Deep Water (NADW) · Labrador Sea · Denmark Strait Sill · Overturning in the Subpolar North Atlantic Program (OSNAP) · Atlantic Meridional Overturning Circulation (AMOC)
The reader stops asking how fast the overturning is weakening and starts asking how much of the record is signal — and how a figure that nobody measures directly is put together.
RAPID-MOCHA-WBTS Array at 26.5°N · Argo Profiling Float Array · Sverdrup (Sv) · Atlantic Meridional Overturning Circulation (AMOC) · Gulf Stream
Runs alongside another chapter, not after
Productivity stops being a property of sunlight and becomes a property of wind direction — and the reader meets the one place in the Journey where the circulation has an address people live at.
Coastal and Equatorial Upwelling · Humboldt (Peru–Chile) Current · Peruvian Anchoveta (Engraulis ringens) · El Niño–Southern Oscillation (ENSO) · Ekman Transport and the Ekman Spiral
Runs alongside another chapter, not after
The Atlantic stops being the subject. The reader sees that the circuit is closed, and the abyss filled, in the ocean furthest from anyone watching.
Antarctic Circumpolar Current (ACC) · Southern Ocean · Drake Passage · Weddell Sea · Antarctic Bottom Water (AABW) · Sverdrup (Sv)
The reader leaves able to hold two things at once: that a decline is assessed as very likely, and that the popular account of what a decline would do to Europe was already wrong before anyone worried about it.
Atlantic Meridional Overturning Circulation (AMOC) · Thermohaline Circulation · Meridional Ocean Heat Transport · Matthew Fontaine Maury · Gulf Stream · RAPID-MOCHA-WBTS Array at 26.5°N
28 records in this Journey.
Showing 8 of 17 featured records. Atlas does not choose which of the rest matter.
The continuous estimate of Atlantic overturning is not built purely from oceanographic instruments. One of its three terms comes from voltages induced on a submarine telecommunications cable — communications infrastructure, laid for other reasons, doing duty as the longest-running component of a climate observing system.
The same rotational turn that piles water into the middle of a basin and sets a gyre spinning also drags water off an eastern margin and pulls nutrients up behind it. One piece of physics produces both the largest features on any ocean map and the fertility of a strip of coast, and the corpus records the dependency in both directions.
Ekman Transport and the Ekman Spiral · Coastal and Equatorial Upwelling · Peruvian Anchoveta (Engraulis ringens)
Two of this Journey's headline figures are quoted in the same unit and are routinely set against each other, and they are not the same kind of quantity: one is horizontal circulation around a continent, the other a vertical cell. The unit is what makes the comparison look legitimate. This is the Journey's lens in a single relation.
Antarctic Circumpolar Current (ACC) · Atlantic Meridional Overturning Circulation (AMOC) · Sverdrup (Sv)
The author of the claim this Journey ends by overturning is also the organiser of the conference that agreed a uniform way of recording observations at sea. The habit of standardised measurement he set up is what any long marine record rests on — including the records used to test him. Being wrong about the mechanism and right about the method are separate achievements, and he managed both.
Matthew Fontaine Maury · Argo Profiling Float Array · Meridional Ocean Heat Transport
The deepest water in the world ocean is made by freezing: sea ice forms over Antarctic shelves and rejects salt into the water beneath, which then sinks. Which means the threat to the abyssal circulation is not heat arriving at depth but fresh water arriving at the surface — the mechanism is undone at its source rather than along its path.
Antarctic Bottom Water (AABW) · Weddell Sea · Southern Ocean
How long does one circuit actually take?
The thousand-year figure is a convention rather than a measurement of any particular parcel of water. Radiocarbon gives the time since a water mass last touched the atmosphere, mixing along the way blurs that, and no parcel has ever been followed round. The corpus states the differences between basins and declines to convert them into a transit time for anything specific.
What does the Labrador Sea contribute over decades?
The array measured twenty-one months and found the eastern section dominant by roughly a factor of seven. Twenty-one months is short for a system with strong interannual variability, and later work reports changes in the Labrador contribution linked to subsurface freshening. The result relocated the dominant term; it did not close the question.
How fast is the Atlantic overturning weakening?
Two published analyses of the identical observing period give decadal trends differing by more than a factor of two. Before the array went in there is no continuous record at all, so every statement about the twentieth century rests on proxies with their own assumptions. The disagreement is largely about which quantity is being called the overturning.
How close is a threshold, and would anyone see it coming?
The assessed position is that decline through this century is very likely and that an abrupt collapse before 2100 is unlikely, at medium confidence. Two lines of work push against the comfortable reading: a statistical early-warning analysis of a proxy reconstruction, and a physics-based indicator. Both are contested, and the direct record is described in its own literature as consistent with model projections rather than with a mid-century collapse. This Journey does not resolve it and does not quote a single year.
When did the Antarctic Circumpolar Current begin?
Unresolved, and the estimates are not close together. They run from the Eocene–Oligocene transition through several intermediate arguments from foraminifera, gateway alignment and flow speed, to a recent case that the modern current is a much later development. The corpus states the disagreement rather than choosing among them.
Will the winds that drive coastal upwelling strengthen or weaken?
Actively disputed, with studies disagreeing on both sign and region. The widely repeated share of global catch attributed to upwelling systems also rests on fisheries accounting whose boundaries are not standardised, so the proportion is softer than its repetition suggests.
Where, exactly, is any of this?
Canonical knowledge locates the constrictions and the basins — a passage, a sill, two seas, one coastal current — and declines to give a coordinate to the overturning, the circumpolar current, the upwelling or the thermohaline circulation itself. That is not an oversight to be worked around. A circulation is not a point, and the absence is the most accurate thing the map can say about it.
The gap between how confidently this ocean is described and how thinly it is watched is not a scandal. It is the ordinary condition of a system that takes roughly a thousand years to turn over and has been instrumented, at one latitude, for twenty-two. Two careful analyses of that same record disagree by more than a factor of two, and most of the disagreement is about which quantity each of them agreed to call the overturning. Which leaves the harder question open, and the evidence genuinely does not close it: can twenty-two years at one section establish that the thing being measured there has changed — or only that the people measuring it were not measuring the same thing?
This is unfinished in the world, not only in the telling.
Ask the world another question. Dynamic keeps this Journey as its context and never changes what is written above.
Open Dynamic Atlas →