Two hundred kilometres off Peru, a ship lowers a sensor on a wire. For the first hundred metres the oxygen reading falls the way anyone would expect, steeply, as sunlit water gives way to dark. Then it stops falling, and stops doing anything at all. For several hundred metres of further descent the trace runs flat — a vertical stripe of ocean in which, as far as the instrument can say, nothing changes. Somewhere near a kilometre down the number climbs again, and the ordinary deep sea resumes.
That flat stretch is an oxygen minimum zone, and the tropical Pacific holds the two largest on Earth. The southern one, reaching out from South America, is roughly six hundred thousand cubic kilometres of water; the northern one, off Central America, is about three times larger. Together with a few others, they amount to less than one per cent of the ocean by volume. They have been in the same places, at the same depths, for as long as anyone has been able to check.
The obvious explanation is appetite. These zones sit beneath some of the most productive surface water in the world, where upwelling feeds enormous plankton blooms. What that surface makes, it eventually drops. Dead cells and faecal pellets sink into the dark, bacteria strip them for energy, and stripping them consumes oxygen. Spend the oxygen faster than it arrives and the water goes empty.
Half of that is right, and it is the wrong half to lead with. Sinking matter is respired in every ocean on Earth; consumption is ordinary. What is rare is having nothing come to fix it. Wind-driven gyres push surface water downward along particular density surfaces and carry it around the basin — but the geometry of that circulation leaves regions those streamlines never enter. Oceanographers have called them shadow zones since Luyten, Pedlosky and Stommel described the ventilated thermocline in 1983. The oxygen minimum is not a map of where the eating is hardest. It is a map of where the water is oldest.
Which makes it a different object from the lake in No. 65. A lake’s summer boundary is a lid: warm water floating on cold, sealing the depths from the air, and the lake dismantles it every autumn when the surface cools. Even Tanganyika, permanently anoxic below a couple of hundred metres, is a closed basin holding still water. An oxygen minimum zone has no lid and no walls. Water crosses it continuously, sideways. The layer is not sealed. It is unvisited.
Below a certain point the layer stops being about oxygen at all. Once concentrations fall to a few billionths of a mole per litre, microbes switch to the next-best thing to breathe and begin respiring nitrate. The switch is sharp — denitrification is half shut down by around two hundred nanomolar of oxygen, a quantity most instruments cannot even see. So this sliver of ocean, under one per cent of the volume, performs something near a third of the entire ocean’s loss of fixed nitrogen, and vents nitrous oxide while doing it. The world’s nitrogen budget is balanced in a place that runs out of air.
We could not measure the bottom of it for a long time, for an embarrassing reason. Water was fetched in plastic bottles lowered on a wire, and the plastic carried dissolved oxygen down with it and leaked it into the sample. Every reading of nearly nothing was inflated by the instrument’s own contamination. The 2021 atlas that finally mapped these zones in three dimensions — about fifteen million sensor measurements gathered over forty years — got around it by giving up on the values entirely and looking instead for depth ranges where the profile stopped changing. The signature of the layer is not a number. It is flatness.
Mapped properly, the shape is not smooth. Ribbons of oxygenated water intrude; there are bites taken out of the shallow edges where something is delivering air the models did not expect. Animals treat the boundary as an address, migrating up through it at dusk and settling back down to its ceiling at dawn, using a layer nothing can breathe in as a place nothing can follow them into.
And the boundary is not one surface. It is a nest of contours, and under warming they are moving in opposite directions. Model projections have the outer envelope — water below roughly a hundred and twenty micromoles per kilogram — expanding steadily, because the circulation that supplies it is slowing. The anoxic core, below twenty, contracts slightly over the same period, because production at the surface falls and there is less to respire. The dead zone is growing. The dead zone is shrinking. Both statements are true, and which one you get depends on which line you agreed to call the edge.
≈
One loop I’m watching
Next: a salt marsh sits in a band of tidal elevation only a few tens of centimetres wide — too low and it drowns, too high and it stops being a marsh. It holds its place against a rising sea by trapping sediment and growing peat at roughly the rate the sea comes up, and the loop is neatly perverse: flooding a little deeper delivers more sediment and grows more plant, so sinking is what pays for rising.