Earthquakes are known for their devastating effects above ground, but some of the most intense action occurs 10 kilometers below. Down there, the sudden lurching of Earth's tectonic plates produces enough friction to melt a slab of rock along the fault. Now, a team of geologists has shown that this melted rock seems to grease faults as they slide past each other, letting quakes really shake their thing.
It's no secret that earthquakes can melt rock: The resulting glassy layer of stone--called pseudotachylyte--rises to the surface when mountains form, allowing geologists a look. Some say that the melt accelerates the quake, just as melting ice under a skater's blade helps her race across a pond. But others believe the melted rock is so thick that it gums up the moving fault, possibly preventing more dangerous earthquakes.
To get to the bottom of things, an international research team, led by geologist Giulio Di Toro of the Università di Padova in Italy studied the process at two locales: in the Italian Alps and in a lab in Japan. The team estimated the amount of friction that occurred during the ancient earthquake by measuring the volume of pseudotachylytes at outcrops. If the liquefied rock had frustrated the quake, there would have been a lot of friction, and therefore heat. That heat would have melted even more rock, producing a large volume of pseudotachylyte. Instead, the researchers found a small volume, consistent with low friction and a slippery melt.
The team bolstered its findings in the lab by using a rotary machine to grind two chunks of rock from the Alpine site together. The low level of friction matched estimates from the field, the team reports today in Science. Di Toro says the findings could help seismologists better understand how the movement of a fault gets propagated from deep underground, because a slipping fault would translate into a more energetic quake than a braking fault.
Structural geologist Richard Sibson of the University of Otago in New Zealand says the combination of lab and field work won him over to the idea of a lubricated fault. "It makes a lot of sense," he says. But Sibson isn't convinced that the conditions Di Toro's team identified in the Alps occur along every fault. He speculates that lubrication might occur during some quakes, but another mechanism might play a role in others. "It raises the issue," he says, "Is there just one type of earthquake?"