Quake Warning System to Be Re-Evaluated for Credibility : Seismology: Experts will discuss setting broader prediction criteria after alert expires with no strong temblor occurring in Central California.

As the latest high-level alert for a strong earthquake expired Wednesday in Central California without such a quake occurring, the director of the state Office of Emergency Services said the alert system will be formally re-evaluated to preserve its credibility.

Director Richard Andrews said the California Earthquake Prediction Evaluation Council, which consists of about a dozen quake experts from throughout the state, would be summoned into session by early next month to discuss establishing a broader set of criteria that would have to be met before an alert was declared.

Since 1982, under various criteria, there have been 11 notices, advisories or alerts of possibly impending earthquake or volcanic activity issued by the Office of Emergency Services and the U.S. Geological Survey in California. None of the events has occurred in the time frame cited.

No change in the state’s procedures for alerts would be made without the concurrence of the federal Geological Survey, which has contributed most of the money to fund the $20-million quake prediction experiment in Parkfield, a Monterey County hamlet, Andrews said in an interview from Sacramento.

Scientists issued a 72-hour A-level alert Sunday for a seven-county area of Central California after a magnitude 4.8 temblor on Middle Mountain, five miles northwest of Parkfield. The moderate shock, which caused no damage or injuries, was part of a swarm of about 25 earthquakes of lesser magnitude that began Saturday afternoon. Middle Mountain lies in what is one of the most frequently active segments of the San Andreas Fault. Similar moderate quakes there preceded the last two magnitude 6 earthquakes at Parkfield in 1934 and 1966.

Andrews said it is possible that more factors than moderate shocks in sensitive areas--such as Sunday’s quake, or a 4.7 shock on Middle Mountain in October--should be considered before an alert is automatically issued.

He said other precursors that could be evaluated include changes in so-called aseismic creep of the fault--slow movement without any quakes occurring--electromagnetic readings, tilt meters and strain evaluations.

The scientist in charge of the experiment for the Geological Survey, geophysicist John Langbein, said in Menlo Park that he has no problems with a re-evaluation such as the one suggested by Andrews. But he said he does not want to use other precursory phenomena because it would reduce the chances that an alert could be issued before the actual quake.

On Sunday, even with automatic procedures and using only the 4.8 temblor as the criterion, it took the state and federal authorities about 30 minutes to issue the alert. In 1966, when the last magnitude 6 quake at Parkfield occurred, the only moderate foreshock came just 17 minutes before the larger quake.

Langbein said the best course would be to simply assign a lower probability to the occurrence of a larger quake after a moderate shock.

“Maybe next time, we’ll decide that the chances of the larger quake should be stated as 1 in 4 rather than 1 in 3,” he said.

When Sunday’s alert was issued, scientists said the chances of a strong quake were 1 in 3. By Tuesday, they put the odds at 1 in 20.

Langbein and another Geological Survey scientist, Bill Ellsworth, suggested Wednesday that the fact that shocks strong enough to mandate an A alert occurred only 13 months apart in the Parkfield area--when earlier probabilities had suggested such alerts would average 54 months apart--may indicate that events are building in the area for a strong quake in the near future.

The Parkfield segment of the San Andreas Fault, about 180 miles northwest of Los Angeles, has had quakes between magnitude 5.5 and 6.3 in 1857, 1881, 1901, 1922, 1934 and 1966. In 1985, the Geological Survey, in the first official earthquake prediction in U.S. history, said there was a 95% chance that such a temblor would occur before the start of 1993, but it has not occurred yet.

“We’re further along the earthquake cycle now,” Langbein said Wednesday. “Another magnitude 5 foreshock may be that kick that’s needed.”

The scientists believe the alert system is important because eventually it may be used to warn Californians of the Big One, a catastrophic magnitude 7 or 8 temblor.

Ellsworth said that evidence suggests that the last estimated magnitude 8 earthquake in Southern California, recorded in 1857 and centered at Ft. Tejon at the top of the Grapevine near Interstate 5, was triggered by two estimated magnitude 6 quakes in the Parkfield area.

So, he suggested, new strong, magnitude 6 earthquake activity in Parkfield, particularly if repeated at short intervals, could lead state and federal authorities to issue a more serious alert, warning Southern California of the possibility of a 7 or 8 quake.

“That’s what the alert procedure is really all about,” Ellsworth said.

In addition to using the Parkfield experiment to issue short-term warnings, scientists say the many instruments the Geological Survey has implanted around the town may help them record more precisely than ever before what happens before, during and after a sizable quake.

A few yards from the ranch house where Ellsworth and a dozen other Geological Survey scientists were headquartered this week is a camera poised over the San Andreas Fault, constantly taking a picture of various posts every six minutes. When quake activity such as last Sunday’s quake takes place, the camera operates constantly.

Even when there is no alert, one USGS scientist, Rich Liechti, is on duty in Parkfield, checking and repairing the various instruments.