Odds of Quake Put Too High, 3 Scientists Say

Earthquakes on the San Andreas Fault happen so irregularly that the state’s current long-term quake forecasts are not reliable and may grossly overestimate the chances of major quakes during the next several decades, according to three UC Santa Cruz researchers who challenged the views of California’s scientific Establishment on Tuesday.

“Our conclusion is that earthquakes are happening so aperiodically that they’re nearly random,” said Steven Ward of the school’s Institute of Tectonics, who was assisted by Saskia Goes and Sergio Barrientos. “It almost doesn’t make any difference how long ago the last earthquake was.”

Goes, addressing a work session of the annual meeting of the American Geophysical Union, said the three scientists believe that a 1988 scientific panel was off the mark when it put the chance of a major quake in Southern California in the next 30 years at 60%. The real odds may be about half that, she said.

Last week, in a dramatic statement of probability, a panel of scientists commissioned by the U.S. Geological Survey and the state Office of Emergency Services put the combined chances of a magnitude 7 quake on a Southern California fault as high as 47% over the next five years.

When Goes suggested Tuesday that the chances of such a quake on the San Andreas alone were no more than 33% in the next 30 years, a member of that panel, David Jackson of UCLA, rose at the meeting to point out that the projection the panel had made covered more faults than just the San Andreas.

However, it was the San Andreas that got by far most of the panel’s attention in its statements last week, and Goes replied that the Santa Cruz group felt the chances of a quake were less than the panel had said.

The Santa Cruz researchers also reported that a computer model they used to mimic 10,000 years of seismic activity on the San Andreas indicated that various segments of the fault do not build up stress and release it on any kind of regular basis.

Instead, Goes said, it appears that a segment interacts with its neighbors, exchanging stress in a complex way that makes the timing of any one earthquake on a particular segment uncertain.

The scientific panel last week contended that the recent Landers earthquake sequence has built up stress along two particular segments of the San Andreas, in the San Bernardino Mountains and in the Coachella Valley, moving up the projected date of a huge quake across both segments at once, something that has not occurred for 500 years.

The panel still could not put a precise date on the next quake, however.

The Santa Cruz researchers also criticized a 1985 assessment by another Geological Survey panel that there was a 95% possibility of a magnitude 6 quake on the Parkfield segment of the San Andreas Fault by Jan. 1, 1993.

With three weeks to go, the quake has not occurred and the researchers said Tuesday that they thought the panel had made a mistake in using the six known prior events on the Parkfield segment to come up with such a strong probability of the seventh occurring by a certain date.

“If you have less than about 10 events, things tend to look more periodic than they really are,” Ward said in a written statement. He said the real probability of a quake between 1985 and 1993 on this segment is only 32%.

Several scientists who have worked on the probability estimates criticized by the UC Santa Cruz researchers held a news conference here this week. When asked about the objections, they simply suggested that the Santa Cruz group join a large group now planning a new 30-year probability rating.