The Ocean Within

By Carl Zimmer|Saturday, October 01, 1994
Old oceans never die: they are preserved, in scattered fossil form, in the rock of Earth’s mantle, 250 miles down.

Jules Verne had an impressive track record of predicting technological progress, having foretold in his books the invention of rocket flight, the submarine, and television. At first glance he would seem to have had less luck with geology; in Journey to the Center of the Earth Professor Otto Lidenbrock and his nephew Axel travel underground and discover that our planet is hollow. But the Lidenbrocks also sail across a vast underground ocean, and geologists now think Verne was onto something there. Where researchers once envisioned nothing but dry rock, they are now discovering signs of huge reservoirs of water deep underground--250 miles deep--that may play a key role in the evolution of Earth.

When the theory of plate tectonics was being formulated in the 1960s and 1970s, most researchers would have told you that water could go underground--but not very far. The plates that make up Earth’s crust form from upwelling hot magma at midocean ridges and move like a conveyor belt until they encounter another plate; at that point one of the two colliding plates may dive into Earth’s mantle, in a process called subduction. Seawater jumps onto this conveyor belt at several stages. It works its way into the crystal structure of the hot rock at the ridge, and it squeezes itself into the pores that open as the rock cools. In the ensuing millions of years, as the plate moves away from the ridge, soggy organic matter rains onto it incessantly, coating it with a waterlogged veneer of sedimentary rock. When the plate finally dives into Earth’s interior, most of this water goes down with it.

Until recently it was thought that the water then quickly came back up. By the time the slab dived 60 miles, geologists figured, the increasing pressure would squeeze the rock like a sponge, sealing its pores and forcing it into a new and denser phase (an arrangement of the rock’s crystal lattice) that had no space for water. The liberated water would rise into the rock above. Like salt spread on ice, the water would dramatically lower the rock’s melting point. Heat from the rubbing of the plates would then melt the rock, and the magma would rise to the surface to form volcanoes, which would spew water vapor into the atmosphere. Later the water would rain to the ground, flow to the ocean, and eventually be sucked back into the mantle.

This neat circuit of plumbing sprang a leak in the mid-1980s. As geologists improved their bookkeeping, they discovered that much more water was being subducted into Earth than was coming out of subduction-zone volcanoes. At the same time, research with diamond-anvil cells--devices that squeeze slivers of rock between diamond tips to mantle pressures while heating them with lasers to thousands of degrees--was revealing new phases of rock that could hold on to water as a plate plunged into the mantle. Only when the descending slab reached a depth of hundreds of miles would its water be forced into the surrounding rock. It became conceivable that there were reservoirs of water down there.

Now it is becoming more than conceivable: there is evidence that the reservoirs really exist. Some of the evidence comes from seismic tomography, a technique that essentially allows researchers to CT-scan Earth. Earthquake waves travel fastest through rock that is cold and dense, slower through rock that is hot. By analyzing seismic waves from many different earthquakes received at many different seismometers, a computer can reconstruct a three-dimensional picture of Earth’s interior. Princeton seismologist Guust Nolet and one of his graduate students, Alet Zielhuis, recently made tomograms of the mantle beneath eastern Europe. There they discovered what they think is a buried ocean--not one you could sail on, to be sure, since the water is in the form of scattered molecules and droplets trapped in the rock, but still a huge volume of H2O.

Between 500 and 400 million years ago there was an ocean at the surface of this part of the world. The ocean was rapidly disappearing, though, as two continental blocks were slamming together to create Europe. In the process an oceanic slab was thrust into the mantle below what is now Poland and Ukraine. Near this slab, at a depth of 250 miles, Nolet and Zielhuis found two regions, each hundreds of miles wide, where seismic waves traveled 4 percent more slowly than average.

That is a much greater slowdown than normal variations in rock temperature or composition could produce. It is the sort of slowdown seismic waves undergo when they travel through rock that is actually molten. Rocks are like particles connected by springs, Nolet explains, and when you pull one, a wave travels through the springs. You can think of melting rock as breaking some of them, which slows down the seismic wave.

There is only one problem with this explanation: at a depth of 250 miles, mantle rock doesn’t normally melt. The pressure is too great. Nolet realized that something abnormal was going on--or rather, two things. The first is water. As the laboratory research had suggested, Nolet thinks the oceanic slab that was buried under eastern Europe carried water down 250 miles and then injected it into the surrounding rock, lowering its melting point.

Under normal conditions, the water would have gone unnoticed. But eastern Europe is abnormal in a second way. Over the past hundreds of millions of years, as Europe has shifted back and forth, it has twice crossed over a hot spot: a narrow plume of hot rock rising from deep in the mantle. Hot spots sometimes poke through the crust and make volcanoes; the one that eastern Europe passed over is now under Mount Etna in Sicily. Under eastern Europe, though, all the plume did was heat the mantle--and melt those regions that were rich in water and thus easily meltable. The two seismically slow regions under eastern Europe, says Nolet, are places where the hot-spot plume hit a trapped ocean of water released millions of years earlier by the ancient subduction zone. The plume acted like a flashlight, exposing parts of the hidden ocean to the scrutiny of seismologists.

There may be other such oceans. The White Mountains of New Hampshire are geologically similar to eastern Europe: they were the site 300 million years ago of a subduction zone, and 100 million years ago they passed over a hot spot. No one has done a seismic tomogram of the White Mountains, but geochemist Tom Torgerson of the University of Connecticut has found what he says is another sign of an underground water reservoir: high levels of the light form of helium, helium 3.

Once helium gets into the atmosphere, it quickly floats out to space, so any helium you find near the surface is gas that has recently seeped up from the mantle. In White Mountains groundwater, Torgerson measured helium 3 levels 1,000 times higher than are typically found elsewhere on Earth--which is evidence, he thinks, that a mantle plume and an underground water reservoir have been conspiring to bring lots of helium up from the depths. When a plume heats up water-rich rock, Torgerson explains, the rock melts in bubbles. Those bubbles suck up lots of gases like helium, just as when you boil water the bubbles suck in oxygen and carry it away, he says. The helium is the smoking gun for Nolet’s hypothesis.

Nolet envisions hidden oceans all over the world, wherever plates once plunged into the mantle. Rough calculations suggest that a volume of water equal to that in the oceans on Earth’s surface may be trapped in these reservoirs. This water might be crucial to many geologic processes. If the mantle plume below New Hampshire hadn’t hit water, for example, creating bubbles of buoyant magma, the White Mountains might be much smaller today. And when water squirts out of the subducted slabs, it may make the dehydrating rock shudder, creating the enigmatic earthquakes that occasionally rumble deep inside the planet--like the one under Bolivia last June that was felt as far north as Toronto.

The buried oceans may also be important to the surface ones we’re familiar with. The old water cycle that researchers once envisioned is being replaced by a new one: some subducted water quickly goes back to the surface through volcanoes, but some flows down to reservoirs 250 miles underground. Mantle plumes that pass through them force the water out and gradually back up to the surface. But Nolet points out that in coming eons, as the interior of Earth cools, mantle plumes may occur less frequently. Then more water will go down than comes up. The surface oceans will gradually drain into Earth’s interior, making our planet a dry world like Venus or Mars. In the end Jules Verne may have the last laugh: of all Earth’s oceans, only his may remain.
Comment on this article

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar

Log in to your account

Email address:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »