By Curt Suplee
Washington Post Staff Writer
Monday, May 25, 1998; Page A03

Earth is the original retread. Although the planet is about 4.5 billion years old, you can hardly find an acre of ocean floor that's a day over 200 million. That's because the surface crust beneath the seas is continuously replenished as new rock wells up from the quasi-molten depths, called the mantle.

Scientists have obtained the first deep images of the way new planetary crust is formed at mid-ocean ridges, as hot rock from the upper mantle (shown above) rises at the junction between tectonic plates.   (Graphic by Patterson Clark/The Washington Post)

The total surface area remains constant because the old crust is pushed aside and recycled. It plunges downward into the mantle again in titanic trenches called subduction zones that lie at many of the boundaries between the 20 separate tectonic "plates" that cover the planet's surface like the polygons on a soccer ball.

The fastest sea floor creation on Earth is taking place along a 5,000-mile ridge called the East Pacific Rise (EPR), where two great plates are separating: the Pacific plate, speeding west at five inches per year; and the Nazca plate, headed east at half that speed. So that's where the National Science Foundation recently sent a team of scientists to see exactly what happens when the planet re-surfaces itself.

The process was not a complete mystery. "Of course, we knew that on average every year 20 cubic kilometers [five cubic miles] of new oceanic crust are formed worldwide, and that the material that forms that crust comes from the partial melting" of rock from the mantle, said geophysicist Sean C. Solomon of the Carnegie Institution of Washington. "But we've never actually seen that happen. It was entirely a product of theoretical models."

Many months and $7 million later, the NSF consortium of seven institutions had actually measured the extent of the magma, as the upwelling soft rock is called. And as the team reports in the May 22 issue of Science, their sea floor instruments brought back not only an unprecedented view of Earth's crustal makeover system, but several major surprises.

One is that the semi-molten mass beneath the EPR ridge is hundreds of miles wide, far broader than many theories assumed, and is not centered under the crack. Yet "99 percent of the melt that forms new crust comes right out at the ridge crest within one kilometer" of the crack, said geologist Donald W. Forsyth of Brown University. What forces combine to force it out of the single narrow crevice are unknown so far.

"Another big question was at what depth the melting begins," said Robert S. Detrick of the Woods Hole Oceanographic Institution. Traditionally, "people had thought that it started maybe at 100 kilometers [60 miles] or 125 at the outside." But the new data show that melting begins as deep as 100 miles and perhaps even 120.

Below that level, strange as it may seem, the rock is solid. It doesn't melt because it suddenly gets hotter near the surface. Just the opposite: It cools slightly as it travels upward, though it's still somewhere around 2,400 degrees Fahrenheit when the melting starts -- as a direct result of drastic pressure reduction. Rock melts more easily at lower pressures, just as water boils more easily at high altitudes where air pressure is lower. For example, at the top of Mount Everest, water boils at around 160 degrees F.

At 100 kilometers (60 miles) below the surface, the pressure on the rock is about 30,000 times atmospheric pressure at sea level, or around 450,000 pounds per square inch. At 10 kilometers (six miles), it's only one-tenth as great.

Only a few percent of the rising rock actually melts outright, but it's enough to make the whole volume less rigid and dense. And that's enough to affect the speed at which seismic waves can travel through the material.

The researchers dropped 51 seismometers over the side in two parallel arrays about 500 miles long stretching across the EPR ridge. Each instrument, roughly the size of a refrigerator, automatically leveled itself after it landed and then settled in to listen to the sounds of distant earthquakes from around the world as the vibrations traveled through various sections of the planet.

All the detectors heard the same vibrations from the same quakes. But some heard them between 0.7 seconds and 3 seconds later than others. The sizes of those time lags tell the scientists a great deal about the rock through which the vibrations passed on the way to each individual detector. Molten rock conveys vibrations a few percent more slowly than solid rock. So by analyzing the pattern of timing for the various seismometers, the scientists developed a detailed picture of the mantle below.

Collectively, the various readings produce a view of the subsurface "similar to CAT scans in medical tomography, revealing features of interest" said University of Oregon geologist Douglas R. Toomey.

The project (called MELT for Mantle Electromagnetic and Tomography Experiment), lasted from fall of 1995 into the following year, and would not have been possible even a couple of years earlier because the necessary detector technology was not readily available. The two key innovations, Toomey said, "were power efficiency and disk storage. We needed to record continuously for seven months" with battery-powered seismometers. Thus they had to be extremely stingy on electrical consumption while also recording the often faint signals from hundreds of quakes, amounting to about a gigabyte (1 billion 8-digit bits of information) per unit. At the end of the collection period, the instruments floated back to the surface for retrieval.

Many of the quakes the instruments recorded were caused by the very process the scientists were studying. Two-thirds of the Earth's crust is made beneath the oceans. Ocean crust -- chiefly a heavy form of igneous rock called basalt -- is heavier than the thicker stuff of which continents are made.

In general, continents are composed of granite-like rock rich in comparatively light silicon-oxygen compounds, aluminum, potassium and sodium. Typically these are about 10 percent lighter than the basalt of which oceanic crust, so continents are buoyant. When oceanic plates hit continental plates, they slide under the higher-riding continental plates. The presence of those subduction trenches makes earthquakes frequent around the western Pacific rim and in western South America.

But seismic activity is not necessary to study the rigidity, temperature and composition of rock in the upper mantle. Another collaboration of MELT researchers placed sensors that record electrical and magnetic fields -- which also vary with rock characteristics -- in the same area. They are scheduled to report their findings this week at the American Geophysical Union meeting in Boston.

"It's like the blind men and the elephant," Forsyth said. "Each of us is seeing the phenomenon with a slightly different tool. And we haven't put it all together yet."