Study of foreshocks lends support to concept of a silent start for quakes

STANFORD -- Most people think of foreshocks as the small earthquakes that trigger a big quake. Indeed, according to one scientific model of nucleation - the start of an earthquake - a tiny rupture triggers a small quake, which then triggers a larger one in a cascade of increasing size. But when Stanford graduate student Douglas Dodge analyzed the mechanics of foreshock sequences preceding six recent California earthquakes, he found that the small quakes were more likely to relieve the stress that led to the main quake rather than to increase it.

"Instead of taking the main shock closer to failure, the foreshocks took the main shock further away from failure," said Dodge's co-author, Gregory Beroza, associate professor of geophysics. Instead of triggering each other, both the foreshocks and the main shock appeared to be triggered by some other mechanism.

William Ellsworth of the U.S. Geological Survey in Menlo Park also was a co-author of the study. Dodge reported the results at the American Geophysical Union meeting in San Francisco on Wednesday, Dec. 13.

His findings lend support to an explanation of the observed weak beginning of earthquakes, the seismic nucleation phase, proposed by Ellsworth and Beroza in Science magazine on May 12, 1995. They showed that instead of cascading, earthquakes might start with a pre-slip stage, hours to days before the earthquake. This pre-slip nucelation phase would be silent, too slow or too subtle to pick up with seismic instruments. Ellsworth and Beroza presented data to show that the larger and more extensive the pre-slip, the more likely that the ensuing quake also would be large.

They said that if earthquakes begin with a pre-slip nucleation phase, it might be possible to detect that phase ahead of time and predict which quakes are going be large ones. If quakes begin in a cascade, there will be no way to predict which small quake is going to start a chain reaction leading to a large quake.

Earthquake precursors

Because few instruments are set up to detect subtle effects before an earthquake, it is difficult to study the nucleation phase and find out which model is more accurate. Dodge decided to look at one quake precursor that has a clear seismic signature - foreshocks.

About 40 percent of moderate to large earthquakes are preceded by nearby quakes that strike hours or days ahead of the main shock. Many are small and not clearly recorded. But Dodge devised a system to look at a set of foreshocks as an ensemble and refine the measurements of their properties. The result is an accurate view of the geometry of the foreshock sequence - pinpointing the epicenter of each earthquake and the orientation of its slip relative to the main shock.

Dodge analyzed foreshocks for three quakes in Southern California: the magnitude 7.3 quake at Landers in 1992; the 6.1 Joshua Tree quake in 1992; the 5.2 Upland quake in 1990. He also analyzed the 1986 magnitude 6.2 quake at Chalfant in the eastern Sierra, the 5.7 Mt. Lewis quake in 1986 near Livermore and the 4.7 Stone Canyon quake in 1986 near Hollister.

He found that the larger the main quake, the larger the geographical area covered by the eart