Plates on the Mantle

Sunday, December 01, 1996
With the theory of plate tectonics, geologists can explain a host of phenomena, from earthquakes to the origin of mountain ranges. Oddly enough, there is one feature of Earth’s surface that the theory fails to explain: the size of the crustal plates themselves. The theory predicts that a large number of relatively small plates should cover Earth’s surface rather than the few large plates that are observed. Hans-Peter Bunge, a geodynamicist at Los Alamos National Laboratory, may have found a way to make theory jibe with observation.

The theoretical problem is fairly straightforward: the continental and oceanic plates ride on a 1,800-mile-thick layer of hot, churning mantle rock. The flowing mantle is driven by heat from Earth’s molten core. Like boiling water in a pot, the mantle rock should flow roughly in a circular pattern: as hot rock rises and cool rock sinks, the horizontal extent of the flow should be about the same length as the 1,800- mile depth of the flow. But the horizontal flow--as marked by the boundaries of the tectonic plates at Earth’s surface--is not 1,800 miles wide but closer to 6,000 miles, on average.

Most geologists have assumed that the steely tensile strength of crustal plates might be squashing the mantle flow, making it spread out near Earth’s surface. But Bunge has used a supercomputer to create a highly realistic three-dimensional model of mantle flow, and he’s shown that it is independent of the plates. The shape of the flow seems to depend entirely on how friction in the mantle increases with depth.

Bunge generated two plateless models of the mantle--one of which assumed uniform mantle friction, and another in which mantle friction increased markedly with depth, the result of pressure compressing the rock. Only the latter model duplicated real-world mantle flow--even in the absence of surface plates. Because the mantle meets less resistance to its flow near the surface, it spreads out, distorting what would otherwise be the neat, nearly circular flow pattern predicted by theory.

To test whether the mantle’s squashed pattern might also be caused by the plates themselves, Bunge imposed 6,000-mile-wide extra-strong plates on both models. Again, only the model with increasing friction showed realistic mantle flow. We are the first to move the problem away from the plates and back to the mantle, says Bunge. By making this one assumption, we turn a fairly unrealistic model into a very realistic one.
Comment on this article