Tiny earthquakes hint at larger shocks.

By recording thousands of puny earthquakes, a pair of seismologists has discovered a way to track the subtle mood swings of California's San Andreas fault. The technique may provide some warning when tensions along the infamous quake-maker reach the breaking point.

The study, reported in the July 30 SCIENCE, looked at 6,000 small earthquakes recorded near Parkfield, Calif., about halfway between San Francisco and Los Angeles. The segment of the San Andreas running through Parkfield has become one of the best-studied faults in the world because it appears to snap regularly every few decades, producing strong earthquakes.

In the late 1980s, Thomas V. McEvilly of the University of California, Berkeley and his colleagues installed seismometers at the bottoms of deep boreholes, 250 meters below the ground at Parkfield. This network of 10 instruments records quakes as small as magnitude -1, far too weak for people to feel.

Robert M. Nadeau and McEvilly discovered that many spots along the San Andreas produce characteristic microquakes months to years apart. By measuring the time between these carbon-copy quakes, the two researchers could detect surges of underground stress. "We're using these sequences of repeating earthquakes like instruments, essentially like strainmeters on the fault," says Nadeau.

Geophysicists have outfitted the San Andreas near Parkfield with several types of surface instruments that can monitor stress along the fault and the movement of rocks on either side of the fracture. The microquakes, however, provide a direct look at the kinds of changes going on deep underground. Taken together, the information from surface instruments and boreholes offers a more complete picture of the fault's behavior, say Nadeau and McEvilly.

Using the pattern of repeating quakes, the scientists deduced that a pulse of stress had migrated from the northwest to the southeast along the Parkfield stretch of the San Andreas fault starting in the late 1980s. The passage of this pulse preceded and overlapped a series of four moderate earthquakes between 1992 and 1994.

Nadeau and McEvilly surmise that the fault may display similar behavior before a major quake at Parkfield. The city has suffered five strong shakes since 1881, and seismologists in 1985 predicted that another one would happen by 1993 (SN: 7/5/97, p. 8). The expected jolt did not strike, but researchers continue to collect data in hopes of capturing any warning signals when it does happen.

The microquake technique could have uses elsewhere. A network of 20 borehole seismometers has already enabled Nadeau and McEvilly to detect repeating sequences of quakes along the Hayward fault, which runs along the east side of San Francisco Bay. "I could envision putting these [borehole seismometers] along any fault that has a seismic hazard potential," says Nadeau.

He and others caution, however, that nobody yet knows what kinds of deep signals will precede a large quake.

William L. Ellsworth of the U.S. Geological Survey in Menlo Park, Calif., has doubts about applying this technique elsewhere. Parkfield, he says, is a unique place, where the San Andreas produces many similar tremors. "It's hard to turn the results from Parkfield into a general method for looking at faults," he says.