A squad of Chinese paramilitary police walk by collapsed buildings in one of the worst earthquake-hit areas of Yingxiu town in Wenchuan county, in China's southwestern province of Sichuan
An accidental discovery has brought seismologists one step closer to being able to predict earthquakes. As part of an unrelated effort to measure underground changes caused by shifts in barometric pressure, a team of researchers found that increases in subterranean pressure preceded earthquakes along California's San Andreas Fault by as much as 10 hours. If follow-up tests advance the findings, seismologists may eventually be able to provide a few hours' notice for people to find safe haven prior to quakes. As the horrific images from China demonstrate, the effort is well worth the alternative. "Predicting earthquakes is the final goal for seismologists," says Fenglin Niu, the research team's lead author and a Rice University seismologist. "This is a start."
Reporting in the July 10 edition of the journal Nature, researchers used a high-tech equivalent of a stereo speaker lowered into a bore hole near Parkfield, Calif., a half-mile deep and five yards from a measuring device. For two months beginning in late 2005, researchers transmitted pulse signals three times per second, from the speaker to the measuring device, calculating travel time between the two stations. Surprised scientists learned the seismic waves slowed dramatically on only two occasions: two hours prior to a magnitude-1 temblor, and a startling 10 hours before a magnitude-3 quake.
The research team theorizes that the immense amount of pressure building along the fault causes small cracks within the rock during the final hours before an earthquake, increasing rock density and slowing the transmission signals. "The more cracks you have, the slower the seismic velocity," says study co-author Paul Silver, a geophysicist with the Carnegie Institution of Washington. Still unknown is whether there is any significance to the fact that the magnitude-3 quake had a much longer pre-seismic signal than the lower-magnitude quake, or whether it was simply because its magnitude was larger and its epicenter closer to the sensors.
If scientists can flesh out the new findings during future earthquakes — a two-year study at the same seismically active location begins this September — it could form the basis of a vastly improved early-warning system for quakes. Current earthquake-warning systems give just a few seconds' notice because they detect only P-waves, the fast-moving seismic waves that precede the more destructive waves released during a quake. Upgrading to a seismic stress meter, however, is still a long way off. "To use this for earthquake prediction, you need to know the precursor waves have a physical basis [that is, increased pressure and a pending quake] and that it's repeatable [with a larger sample size of quakes]," Niu says. He also hopes to test whether the stress signals would still be detectable on a larger scale, with the two sensors spaced more than a few yards apart.
Barring a major effort to drill multiple, half-mile-deep bore holes along fault lines, researchers would also need to develop a surface-based detection system capable of filtering out temperature swings, precipitation and other "noise" that could confuse their seismic readings. Says Silver, "We obviously have more work to do, but we're certainly encouraged because this is what people are looking for."