Ocean circulation

What keeps the ocean's great currents moving instead of settling into stillness?

Stir a cup of tea and the swirl dies within seconds; friction always wins. Yet the Gulf Stream has run for as long as anyone has measured it, and the deep ocean is not a still pool but a slow, organised traffic of water masses. Friction has not stopped working down there. Something must be paying the bill, continuously.

So the useful question is not why the water moves but what keeps putting energy in, and where that energy enters. Those have two quite different answers, and confusing them is the commonest error in this subject.

Start at the top. Winds blow steadily over the sea, and a wind dragging on water for years drags a great deal of water with it. But the sea does not simply flow downwind. On a rotating planet moving fluid is deflected, and that deflection plus the walls of the basins bends the flow into closed loops, the great gyres, whose western edges are squeezed into narrow fast currents like the Gulf Stream and the Kuroshio. This wind-driven circulation is fast and shallow: mostly the upper few hundred metres, with a parcel rounding a gyre in a matter of years.

Now the deep, where no wind reaches. What makes seawater heavy? Cold, and salt. Where do you get both at once? At high latitudes in winter, where the sea gives up heat to a freezing atmosphere and forming sea ice leaves its salt behind in the water below. Water made cold and salty in the Nordic and Labrador Seas, and around Antarctica, grows dense enough to sink and spread along the floor. That is the density-driven overturning: slow, deep, on a circuit measured in something like a thousand years.

Now the question that catches most explanations out. If dense water sinks, what lifts it back? The tidy story says it warms and rises — but notice where the heating and the cooling both happen: at the surface. Heating a fluid at the same level you cool it cannot by itself sustain a deep circulation, an argument going back to Sandstrom and still the reason oceanographers resist calling the overturning a simple heat engine. The lifting must be done mechanically: Southern Ocean winds draw deep water up along tilted density surfaces, and tides sloshing over rough sea floor generate internal waves that break and stir dense water upward. The pattern is set by density; much of the energy comes from wind and tides.

And why does none of it settle? Because the gradients are rebuilt faster than friction erases them. The Sun keeps heating the tropics, the poles keep radiating heat away, evaporation and rainfall keep redistributing salt, the winds keep blowing. A gradient continuously restored is a source that never runs out.

Three corrections. First, "ocean currents" and "thermohaline circulation" are not synonyms: the surface currents you can see are mostly wind-driven, and far faster than the density-driven overturning sharing the same water.

Second, the two are not separate machines. The Gulf Stream carries north the warm, salty water that later becomes the dense water that sinks; the overturning in turn shapes the surface temperatures that steer the winds.

Third, a caution about the Atlantic overturning, or AMOC, often discussed as though it were closely watched. Continuous measurement across the Atlantic at 26 degrees north began only in 2004, and the record swings widely from year to year — a short record against a circulation with a centuries-long memory. There is real evidence that warming could weaken it, since freshening from ice melt and rainfall makes surface water harder to sink. But confident claims about how much it has already weakened, or when it might collapse, run ahead of what two decades of data can settle.

• Why western boundary currents like the Gulf Stream are so much narrower and faster than the rest of their gyres.
• How the Southern Ocean's unbroken circumpolar path makes it the hinge of the global overturning.
• What sediments and ice cores suggest about abrupt past changes in the Atlantic overturning.

Every term the collection defines is gathered in the glossary.