UCSB researchers using three-dimensional computer models have found that earthquakes work much differently than previously assumed.

Researchers at the Institute for Crustal Studies found that when a crack moving along an earthquake fault reaches a stronger area where the fault requires more energy to break, the so-called “barrier” can focus the quake’s energy into intense bursts, causing extra-fast rupture speeds along the fault.

The group, consisting of physics graduate student Eric Dunham, physics professor Jean Carlson and postdoctoral researcher Pascal Favreau, began the study in spring 2002 and made their findings last September, but didn’t realize the implications until recently, Dunham said. The results were published in the March 7 edition of Science magazine.

Dunham said the findings contradict the previously held belief that a stronger region along an earthquake fault slows the energy down.

“Instead, the barrier concentrates the energy, and when it finally breaks, it releases extreme pulses to the sides,” he said.

Carlson said a barrier could be caused by the presence of stronger rock, or a bend or protrusion in the fault.

The study was noteworthy for its original goals as well as its findings, Dunham said.

“Previous studies have assumed that faults have uniform material strength,” he said. “We looked at a more abstracted model with varying strengths.”

In order to handle the massive computing demands of generating three-dimensional models of faults with varying strength, the group used a technique called parallel computing, in which researchers use several different computers working together doing different functions in the processing. Dunham said the technique, which involved the use of as many as eight computers at once to perform the simulations, would have been impossible until recently.

“Only in the past few years have computers been fast enough in a three-dimensional medium, which is how we know Earth,” he said.

The group received funding for the study from the Keck Foundation, which donated $1 million to the university three years ago to fund investigation of materials under stress during earthquakes.

Researchers are only now gaining the knowledge about fault “barriers” that could be used to apply the findings practically.

“We only really know the strength of regions that have already broken,” Dunham said. “But we can look at a quake that didn’t break one region and know that if it did, it could be very hazardous.”