Rather than turning to broad-scale plate tectonics to investigate earthquakes, David Goldsby, associate professor of geological sciences, and Terry Tullis, professor emeritus of geological sciences, decided to take a closer look. Their study, published in this week's issue of the journal Science, examines earthquake processes on a microscopic scale.
But their findings are anything but small — they shed new light on the physics behind earthquake production.
About 10 years ago, the two began to examine the relationship between the rate at which rock faces slide past one another and the strength of the frictional forces between them. According to their study, heat is generated almost exclusively at microscopic contact points when rock faces slide past each other.
Under normal speeds, the generated heat has time to spread across the entire rock. But as plates approach earthquake speeds, heat does not have time to escape the contact points, producing an effect called flash heating, Goldsby said. Flash heating can result in temperatures of up to 1,800 degrees Celsius.
As the contact points get warmer, the heat decreases the frictional forces between the rocks and they become weaker.
In a series of experiments designed to mimic earthquake speeds, the researchers tried to determine the point at which frictional forces begin to significantly diminish. To isolate the effects of flash heating without raising the overall rock surface temperature, the two took special pains to slide the rocks' faces past one another at fast speeds and over short distances.
Once speeds reach 100 millimeters per second, rock strength plummets, Goldsby said. If their frictional-reducing model is correct, once rock faces reach that critical speed, a positive feedback loop is created. Weakened frictional forces allow the rocks to slip more quickly.
The concept of flash heating was first conceived in the 1930s, but Goldsby "was the really the first person who demonstrated (flash-heating) works for rocks," Tullis said.
The most exciting part of their study was "plotting up the data and showing that (friction strength) has this exact dependence on velocity," Goldsby said.
But it was not all smooth sailing — their efforts were complicated by the difficulty of proving that their results were not "flukey" and a result of the machines they used, said Tullis. "It was quite challenging to figure all that out."