Seeking the Magic Bullet : SCIENCE FILE: An exploration of issues and trends affecting science, medicine and the environment : Once thinking they were on the brink of predicting quakes, scientists know they have a way to go.

With a full gray beard that spills down onto his chest, all Allan Lindh needs is a flowing white robe to make him look like a guru descended from the mountaintop.

It would be in keeping with the visual image if he suddenly began uttering incomprehensible incantations. But instead he is one of the handful of people who can speak lucidly about a subject that has vexed scientists for decades.

As chief seismologist at the U.S. Geological Survey’s western headquarters in Menlo Park, Calif., it is his job to steer the meager but concentrated efforts to learn how to predict earthquakes.

A couple of decades ago, scientists thought they were on the brink of success. But today, nearly all of them believe they vastly underestimated the challenge. Nevertheless, Lindh and others believe there are some promising areas now being explored.

Researchers scattered around the world are hunting for a telltale twitch, some signal, that the earth is about to move. They are studying possible foreshocks, ground movements, electrical signals--anything that might be a sign of an impending earthquake.

At stake is the solution to one of science’s most enduring mysteries, and perhaps the security of millions.

Scientists have been looking for an event--such as deformation of the ground adjacent to a fault--that would be a “magic bullet” telling them a quake is about to strike. What is needed first, Lindh says, is to understand earthquakes better and let the predictions evolve. Science is, he adds, getting closer.

“Not everybody believes this stuff, of course, but on several fronts we are making incremental progress in understanding earthquakes. I think that’s the only rational way to try to predict them.”

That point was underscored recently when seismologists Gregory Beroza of Stanford University and William Ellsworth of the Geological Survey reported the discovery of a very faint, irregular motion that immediately preceded all 30 quakes they had studied. The pattern of motion came within five seconds of each quake, indicating that the fault was going through a preparatory stage.

That discovery reinvigorated hopes that a series of events may precede earthquakes, thus providing some means of predicting them.

No country has tried harder than China to predict earthquakes, and no one else has had such astonishing results. In 1975, a major earthquake in the Haicheng area was predicted early enough to evacuate hundreds of buildings, saving thousands of lives. The scientists made their prediction after the seismically inactive area was struck with thousands of small earthquakes, which they interpreted as foreshocks.

Lucile Jones was working on her doctorate at the Massachusetts Institute of Technology when reports of the prediction reached the West.

Jones, then a 23-year-old whiz kid, spent five months in China and she became convinced that the Chinese had in fact predicted the Haicheng earthquake, saving as many as 30,000 lives.

But just one year later another quake struck in northeastern China at Tangshan. There was no warning. More than 300,000 people died. The Chinese clearly had not found the “magic bullet.” Yet Jones was convinced that the Haicheng triumph was more than luck. “The Haicheng earthquake was predicted in a large part because they guessed that foreshocks were foreshocks,” says Jones, who has returned to China three more times.

Jones believes that earthquakes beget earthquakes, and some of them are likely to be foreshocks of larger quakes. The great California quake of 1857 was preceded by a strong foreshock several hours earlier. But the 1906 quake that virtually destroyed San Francisco struck without a foreshock.

It would be convenient if foreshocks were clearly different from other quakes, and some experts believe that may turn out to be the case. Jones’ research suggests, for example, that many foreshocks are followed by tightly clustered smaller quakes before the main shock arrives, as in Haicheng.

Other scientists believe that the shock waves from foreshocks may have a different seismic signature than other temblors. They may travel more slowly, for example.

Some concede, however, that scientists may never be able to determine which quakes are foreshocks. Antony Fraser-Smith believes it may not be necessary.

In the fall of 1989, from his research site in the Santa Cruz Mountains between Stanford and the Pacific Ocean, the physicist and professor of electrical engineering at Stanford University began picking up a strange signal. “We thought it was some kind of interference” and the readings were ignored, he says.

The Loma Prieta quake struck a few days later, on Oct. 17, 1989, and the epicenter was only three miles from where Fraser-Smith had stashed his equipment. The “funny signals” that the equipment had recorded for 12 days continued up to the quake.

And the records revealed something startling. “The signals took off and got very, very big in the three hours just before the earthquake,” Fraser-Smith says. “And then they stopped.” There was no doubt in his mind that the powerful signals were caused by some kind of change in the ground along the fault before the earthquake. It was not all that surprising because scientists in other countries--particularly Japan and China--have reported detecting electromagnetic signals during and before earthquakes, but few in this country had taken the claims seriously.

Reports of Fraser-Smith’s claims hit the world of seismology like a thunderclap. If his results could be duplicated elsewhere, a valuable new tool had been found. Instruments like Fraser-Smith’s could be positioned along major faults, and they could give a strong warning days before major earthquakes.

But no one, including Fraser-Smith, has been able to come up with a satisfactory explanation for what could have caused the signals. Seismologists will not embrace Fraser-Smith’s findings until another instrument at another location comes up with the same results. That could come at the central California community of Parkfield.

For years, scientists have been expecting an earthquake of at least magnitude 6 to strike Parkfield. The historical record suggests that a temblor of that size has ruptured along the San Andreas Fault there about every 22 years. Nearly a decade ago, Lindh and William H. Bakun, a fellow seismologist at Menlo Park, recognized that Parkfield presented a unique opportunity. The apparent regularity of earthquakes at Parkfield meant scientists could place the right instruments there before the quake and measure every change that the fault goes through just before, during, and after the quake. So in 1985 Lindh and Bakun called for the establishment of a unique program at Parkfield.

It is known today as the Parkfield Experiment, and the area around the tiny community is probably the most heavily instrumented region in the world. But Parkfield has balked. The quake should have struck by the end of 1992.

There have been a lot of small temblors, but the quake that scientists have been waiting for has not come. That has become a sore point among many seismologists. Some argue that Lindh and Bakun erred in their assumptions, particularly regarding the regularity of the Parkfield quakes. Lindh and Bakun remain true believers, but even they are showing signs of frustration. “I sure as hell wish it would get its act together and get it over with,” says Bakun. “We’re ready.”

“If praying would do it,” Lindh says, “it would have happened.” Despite the frustration, there is no other place in the country where they are as likely to trap a moderate-size earthquake. Fraser-Smith has placed two of his instruments there, three miles apart, and for him the Parkfield Experiment is make or break.