New study upends a theory of how Earth's mantle flows

By developing a better picture of the underlying engine of plate tectonics, scientists hope to gain a better understanding of the mechanisms that drive plate movement and influence related process, including those involving earthquakes and volcanoes.

When we look out at the earth, we see its rigid crust, a relatively thin layer of rock that makes up the continents and the ocean floor. The crust sits on tectonic plates that move slowly over time in a layer called the lithosphere. At the bottom of the plates, some 80 to 100 kilometers below the surface, the asthenosphere begins. Earth's interior flows more easily in the asthenosphere, and convection here is believed to help drive plate tectonics, but how exactly that happens and what the boundary between the lithosphere and asthenosphere looks like isn't clear.

To take a closer look at these processes, a team led by scientists from Columbia University's Lamont-Doherty Earth Observatory installed an array of seismometers on the floor of the Pacific Ocean, near the center of the Pacific Plate. By recording seismic waves generated by earthquakes, they were able to look deep inside the earth and create images of the mantle's flow, similar to the way a doctor images a broken bone.

Seismic waves move faster through flowing rock because the pressure deforms the crystals of olivine, a mineral common in the mantle, and stretches them in the same direction. By looking for faster seismic wave movement, scientists can map where the mantle is flowing today and where it has flowed in the past.

Three basic forces are believed to drive oceanic plate movement: plates are "pushed" away from mid-ocean ridges as new sea floor forms; plates are "pulled" as the oldest parts of the plate dive back into the earth at subduction zones; and convection within the asthenosphere helps ferry the plates along. If the dominant flow in the asthenosphere resulted solely from "ridge push" or "plate pull," then the crystals just below the plate should align with the plate's movement. The study finds, however, that the direction of the crystals doesn't correlate with the apparent plate motion at any depth in the asthenosphere. Instead, the alignment of the crystals is strongest near the top of the lithosphere where new sea floor forms, weakest near the base of the plate, and then peaks in strength again about 250 kilometers below the surface, deep in the asthenosphere.

"If the main flow were the mantle being sheared by the plate above it, where the plate is just dragging everything with it, we would predict a fast direction that's different than what we see," said coauthor James Gaherty, a geophysicist at Lamont-Doherty. "Our data suggest that there are two other processes in the mantle that are stronger: one, the asthenosphere is clearly flowing on its own, but it's deeper and smaller scale; and, two, seafloor spreading at the ridge produces a very strong lithospheric fabric that cannot be ignored." Shearing probably does happen at the plate boundary, Gaherty said, but it is substantially weaker.

Donald Forsyth, a marine geophysicist at Brown University who was not involved in the new study, said, "These new results will force reconsideration of prevailing models of flow in the oceanic mantle."

Looking at the entire upper mantle, the scientists found that the most powerful process causing rocks to flow happens in the upper part of the lithosphere as new sea floor is created at a mid-ocean ridge. As molten rock rises, only a fraction of the flowing rock squeezes up to the ridge. On either side, the pressure bends the excess rock 90 degrees so it pushes into the lithosphere parallel to the bottom of the crust. The flow solidifies as it cools, creating a record of sea floor spreading over millions of years.

This "corner flow" process was known, but the study brings it into greater focus, showing that it deforms the rock crystals to a depth of at least 50 kilometers into the lithosphere.

In the asthenosphere, the patterns suggest two potential flow scenarios, both providing evidence of convection channels that bottom out about 250 to 300 kilometers below the earth's surface. In one scenario, differences in pressure drive the flow like squeezing toothpaste from a tube, causing rocks to flow east-to-west or west-to-east within the channel. The pressure difference could be caused by hot, partially molten rock piled up beneath mid-ocean ridges or beneath the cooling plates diving into the earth at subduction zones, the authors write. Another possible scenario is that small-scale convection is taking place within the channel as chunks of mantle cool and sink. High-resolution gravity measurements show changes over relatively small distances that could reflect small-scale convection.

"The fact that we observe smaller-scale processes that dominate upper-mantle deformation, that's a big step forward. But it still leaves uncertain what those flow processes are. We need a wider set of observations from other regions," Gaherty said.

The study is part of the NoMelt project, which was designed to explore the lithosphere-asthenosphere boundary at the center of an oceanic plate, far from the influence of melting at the ridge. The scientists believe the findings here are representative of the Pacific Basin and likely ocean basins around the world.

NoMelt is unique because of its location. Most studies use land-based seismometers at edge of the ocean that tend to highlight the motion of the plates over the asthenosphere because of its large scale and miss the smaller-scale processes. NoMelt's ocean bottom seismometer array, with the assistance of Lamont's seismic research ship the Marcus G. Langseth, recorded data from earthquakes and other seismic sources from the middle of the plate over the span of a year.