The Great Ocean Conveyor

If you've ever seen that famous ribbon looping around the globe — the ocean conveyor belt — here's a fun origin story. Wally Broecker remembers it starting life as a cartoon in Natural History Magazine in nineteen eighty-seven, sketched to explain the Younger Dryas, that sudden cold snap near the end of the last ice age. It wasn't meant to be a logo, just a simple picture for lay readers. Then it went viral, decades before "viral" meant what it does now. And in hindsight, that's part of its magic: it compresses a messy, three-dimensional tangle of currents and climate feedbacks into one memorable idea. The Atlantic takes in salt, gives off heat, and in the process helps keep Europe's winters weirdly mild for its latitude. Let's make that picture sharper without losing the simplicity. The conveyor has two limbs. In the north, surface water gets cold and salty enough to sink and become North Atlantic Deep Water, or NADW. That dense tongue slides south as a deep western boundary current. Near Antarctica, an even denser mass — Antarctic Bottom Water, or AABW — under-rides it. After a long tour, these deep waters mix and eventually return toward the surface. It's not a neat, closed pipe. As Broecker likes to stress, NADW doesn't travel as a labeled package; it blends and loses its identity well before it's halfway around Antarctica. But taken as a whole, that sinking in the north and return at the surface moves an enormous amount of water and heat. Where does the push come from? Oddly enough, from the sky. The Atlantic exports freshwater to the atmosphere — think evaporation carried away by winds — and that leaves the remaining seawater saltier and denser. Here's the lever in numbers. If the atmosphere removes about zero point thirty-five Sverdrups of freshwater from the Atlantic — a Sverdrup is a million cubic meters per second — the ocean must send saltier water out at depth and bring back slightly fresher water at the surface to keep the budget balanced. The deep outflow has a salinity near thirty-four point nine per thousand. The return flow needs to be around thirty-four point three per thousand to compensate. That zero point six per thousand gap between what leaves and what returns is roughly sixty times bigger than the tiny salinity differences you'd otherwise trade between basins. From that contrast, Broecker and colleagues argue that a seemingly modest vapor loss can drive a roughly twenty-Sverdrup conveyor. That salt-driven circulation is a heat engine for the North Atlantic. The math is simple and powerful. When surface water turns into deep water, it has to cool. Each cubic centimeter that makes the jump gives up about seven calories of heat to the air. Multiply that by a lower-limb flux around twenty Sverdrups and you get on the order of four times ten to the twenty-first calories released each year — about thirty-five percent of the sunlight that hits the Atlantic north of forty degrees north. That's why winters in London and Paris don't feel like Winnipeg. It's the conveyor on full blast, dumping heat into the storm tracks that feed Europe. How do we know how strong that deep flow really is? One elegant line of evidence comes from radiocarbon. Deep Atlantic waters are "older" in a radiocarbon sense because they've been away from the air longer; you can translate that age into a renewal rate. The tricky bit is that deep waters are mixtures. Northern-sourced waters start with a different radiocarbon signature than southern ones. Broecker breaks that down by using end members with initial apparent ages of about minus sixty-eight per mil in the north and minus one hundred fifty-eight per mil in the south, then tracking how much of each is in a given sample. To keep that mixing honest, he leans on a property oceanographers call PO-star. It's a bookkeeping trick: you take phosphate and dissolved oxygen and combine them with fixed weights — one hundred seventy-five and one point ninety-five in his preferred formulation — to create a quasi-conservative tracer that moves with the water mass. PO-star tells you, parcel by parcel, how much is northern versus southern source. Once you correct for those ingredients, you can compute fluxes. Do that math and you land in a tight neighborhood. The radiocarbon-based estimate for NADW pumping into the deep Atlantic is about twenty-three Sverdrups, with Antarctic Bottom Water contributing around four Sverdrups, for a total near twenty-seven. The deep Atlantic reservoir volume is roughly one point five five times ten to the seventeenth cubic meters, and the average residence time clocks in at about one hundred eighty years. That pairing — a big pool and a century-scale turnover — gives you both the scale and the rhythm of the conveyor's lower limb. What about the water making its way back to the South Atlantic at the surface? There are three main suppliers. Antarctic surface water slips through the Drake Passage with a salinity around thirty-three point eight per thousand. Warm, salty Indian Ocean water sneaks in via Agulhas leakage at roughly thirty-five point one per thousand. And intermediate waters form along the northern edge of the Antarctic sector, sitting near thirty-four point three per thousand. You might think you could just use salinity to weigh the fractions. Not so. You can get the same bulk salinity by different recipes — one example Broecker plays with is mixing roughly one point six parts Drake Passage water with one part Agulhas water. The numbers work out the same, but the story under the hood is different. Salinity alone can't tell you who did the heavy lifting. Follow the deep limb south and you can watch the imprint of NADW fade and blend. At about three kilometers depth across the Atlantic, NADW still leaves a strong mark. But after the flow rounds the tip of Africa, it mixes with Antarctic-developed deep waters. PO-star settles into a homogenized value around one point thirty-seven, and by the time those mixtures fill the Indian and Pacific basins, a rough recipe holds: think one part NADW to about two parts Antarctic-origin water. That mixing is global housekeeping. It also ties back to the salt budget: surface waters in the Atlantic north of forty degrees north sit, on average, about one gram per liter saltier than the Pacific. If the conveyor didn't export that excess salt continually, the Atlantic would drift upward by roughly one point four grams per liter each millennium, and deep-water formation would choke. Models help translate that circulation into climate fingerprints. Using the Princeton ocean model in the late nineteen eighties, Syukuro Manabe and Kirk Bryan's group, in work highlighted by Manabe and Stauffer, found two distinct states for the Atlantic: conveyor on and conveyor off. In the on state, with a thermohaline component of about twelve Sverdrups in their setup — the real ocean is closer to twenty — sea surface temperatures in the northern Atlantic ran roughly five degrees Celsius warmer than in the off state. That's a huge regional imprint. It says the conveyor is not just moving mass; it's painting a thermal pattern on the atmosphere. What happens upstairs when you switch that pattern? An atmospheric general circulation model can tell you. David Rind and colleagues took the surface temperature differences between those two ocean states — reconstructed from glacial to interglacial changes — and asked the atmosphere to respond. In winter, the warming signal arced across Europe toward Siberia. That's the conveyor's heat output going airborne, visible not just over the ocean but bleeding into the continents downwind. If heat is the payoff, freshwater is the hand on the dimmer switch. North of forty degrees north, precipitation plus river runoff exceed evaporation by about zero point thirty Sverdrups. Add about one Sverdrup of relatively fresh water slipping in through the Bering Strait — which amounts to roughly zero point zero six Sverdrups in freshwater terms — and the region sees a total of around zero point thirty-six Sverdrups of freshening pressure. At today's conveyor strength, that water gets swept away efficiently. Now turn the knob and see how salinity responds. With a twenty-Sverdrup flush, the northward-flowing surface waters dilute by about zero point sixty-three per thousand as they pass into the Nordic Seas. Weaken the conveyor to fifteen Sverdrups and that dilution grows to roughly zero point ninety-four per thousand. Drop it to ten and you're near one point twenty-six per thousand. Somewhere along that slide, the surface gets too light to sink, deep-water formation stalls, and the conveyor can flip off. In model experiments, the off mode pushes surface salinity in the northern Atlantic down by about three per thousand. And in simulations by Ernst Maier-Reimer and Uwe Mikolajewicz, even modest freshwater additions to the NADW source region killed the thermohaline circulation with a response time on the order of decades — fast, in ocean terms — given that twenty-Sverdrup flushing rate. Paleoclimate adds the drama. If you want a proxy for conveyor state, Broecker points you to air temperatures around the northern Atlantic. Turn the conveyor on or off and Greenland and Europe shift by something like five to eight degrees Celsius. The Greenland ice cores — Camp Century, Dye three, and later, deeper records — show exactly those kinds of jumps during glacial times, rising and falling on millennial scales with a quickness that fits a circulation switch. Then look at the last nine to ten thousand years. Temperatures over Greenland level out. That postglacial stability looks like the conveyor locked in to the on position. How could the system jump so fast and so often in the ice ages? Broecker's salt-oscillator idea is a plausible script. Picture the conveyor humming: it exports salt, keeps the North Atlantic dense, and pulls heat poleward. Then big pulses of meltwater from ice sheets pour in and dilute the surface. During those on phases, he estimates meltwater fluxes around zero point twenty-five Sverdrups. That freshening lightens the surface and can push the system toward a stall. If it shuts off, salt export slackens, freshwater pools at the top, and salinity slowly creeps back up until the density is high enough to reignite deep convection. Switch back on, and you get an abrupt warming. The ice records even resolve events that sharp — jumps over about fifty years near twelve thousand seven hundred and again around ten thousand radiocarbon years before present — with the warming pattern spreading from the Canadian and Greenland coasts eastward toward Europe and Russia, just where a conveyor turn-on would vent heat. None of this says the conveyor is the whole climate story. Broecker and George Denton have both argued that abrupt warmings, especially those that end ice ages, need help from the atmosphere's big engines — the Hadley cell, cloud feedbacks, and shifts in water vapor. The ocean can shove, but the air can lean in. And even within the ocean, the restart problem is hard. No ocean general circulation model has yet produced a robust, physics-rich path for taking a shut-down conveyor and flipping it back on. Freshwater pooling at high latitudes presents a formidable barrier to re-establishing deep convection. There's a suspicion that brine processes — the salty plumes formed when sea ice grows and rejects salt — may be part of the boot-up sequence. But that's still an open engineering problem for the climate system. If you strip the logo down to its essential physics, a few numbers keep showing up. A lower-limb export near twenty to twenty-three Sverdrups. An atmospheric vapor loss of roughly zero point thirty-five Sverdrups that, through a zero point six per thousand salinity contrast, can drive that massive flow. A heat release over the North Atlantic worth about a third of the region's incoming sunlight. And a deep ocean that turns over on a timescale of a couple of centuries, mixing NADW and Antarctic waters until their fingerprints blur. Why should you care about the exact salinity of water sneaking through the Drake Passage, or a quirky tracer that combines phosphate and oxygen with constants you'll never memorize? Because those are the tools that turn a sketch into a diagnosis. They let Chen's radiocarbon lab or Broecker's tracer spreadsheets say, with numbers, how the system is wired today and how it behaved when the world was younger, icier, and jumpier. They tell us that the Atlantic really is saltier than the Pacific by about a gram per liter in the north, that this isn't a fluke, and that if we stopped exporting salt, the ocean would notice quickly in geologic time. There's a temptation, when you hear "conveyor off," to jump straight to catastrophe. Take a breath. The Holocene — the last nine or ten thousand years — looks stable in Greenland. Europe's winters are still milder than their latitude would suggest. Yet the paleoclimate record is waving a flag: the Atlantic thermohaline circulation can move abruptly when pushed by freshwater, and the pushes don't have to be huge. Fractions of a per mil in salinity matter when you're on the edge of convection. So here's where we end up. The logo is a simplification, but the physics it points to are real and quantifiable. Heat comes off the ocean when water sinks, and that heat keeps Europe warm. The sink is maintained by salt left behind as vapor blows away, and that salt budget can be tipped by meltwater, rain, and river flow. The deep limb is strong, globally connected, and thoroughly mixed with Antarctic waters on a timescale of centuries. And in both models and ice, we see two modes: on and off. The forecast? Keep the speculation short and the measurements long. Watch the freshwater budget in the subpolar North Atlantic and the Arctic gateways. Keep refining the tracers that pin down mixing and flux. And, as Broecker liked to remind anyone who'd listen, don't get seduced by a single lever. The conveyor is a big one, but it's part of a larger machine — ocean, atmosphere, ice — that we're still learning to read, one careful number and one good story at a time.