Research on Quakes Assures One Thing: Big One Gets Closer

The task of determining when The Big One will strike California remains as much a matter of educated guesswork as science, the results of three studies indicate.

A U.S. Geological Survey-sponsored panel released a report Thursday estimating the chances of major earthquakes along California’s best-known geologic faults to be 50% to 60% over the next 30 years. And while the scientists say these are the most accurate figures yet derived, uncertainties in the underlying data and new findings in Southern California are already clouding the results.

Another study by scientists at Caltech and Columbia University, published today in the journal Science, suggests a major quake could hit Southern California this year--or in a century. “In the words of Winnie the Pooh, things are getting complicated-er and complicated-er,” said Kerry Sieh, a Caltech seismologist who participated in that and a third, unpublished study of the San Andreas Fault.

Risks Increase Daily

Nevertheless, Californians should not let the ongoing scientific debate obscure the report’s basic conclusions that the earthquake risks are high and that they increase with every passing day, said Richard Andrews, deputy director of the state Office of Emergency Services.

“For the individual residents of California, the message is very clear,” Andrews said. “Until the big one happens, it (the risk) is going to get worse before it gets better.

“Unless we aggressively pursue earthquake hazard reduction measures, there is the potential for many people to die--perhaps more people than have ever died in a single day from a natural disaster in the history of the United States.”

If a large earthquake were centered in a metropolitan area, it could cause up to $60 billion in damage and kill 3,000 to 23,000 people, a congressional report estimated in 1987.

The eight-month statistical study released Thursday by the federal task force confirmed what has been the conventional wisdom among seismologists for years: that the two California regions most at risk of earthquakes into the next century are also its biggest population centers, the San Francisco Bay Area and Los Angeles.

The report says there is a 60% chance of a major earthquake, magnitude 7.5 or 8, along the San Andreas Fault in Southern California over the next 30 years. There is a 10% chance that could happen within the next five years.

But the hazards in fact could be much higher, the report notes, because the scientists did not attempt to consider the dozens of geologic faults within the metropolitan Los Angeles and Orange County areas. They concentrated on the San Andreas Fault and on three other relatively well-known faults in Northern and Southern California.

The report also concluded:

- There is a 50% chance of an earthquake of magnitude 7 in the peninsular San Francisco and East Bay areas along the San Andreas and Hayward faults.

The report placed special emphasis on how little is known about the Hayward Fault, which stretches along the heavily populated East Bay corridor from Berkeley to San Jose. Because 120 years have passed since the last large earthquake on the Hayward Fault, “the possibility of a large earthquake should be a serious societal concern,” the report says.

- The segment of the southern San Andreas with the highest likelihood, 40%, of a major rupture over the next 30 years is the Coachella Valley segment that runs through Indio.

- The chances of an earthquake of magnitude 6.5 to 7 along the San Jacinto Fault were put at 10% within the next five years and 50% over the next 30 years. The fault extends 150 miles southeastward from just north of San Bernardino to the Superstition Mountains near El Centro.

- Along the Imperial Fault farther east, the chance for a quake of magnitude 6.5 to 7 over 30 years was put at 50%.

These numbers should be considered as minimum estimates, cautioned the report by 12 government, academic and private research scientists, who were convened in April, 1987, by the National Earthquake Prediction Evaluation Council. The council advises the director of the U.S. Geological Survey.

Minimum Estimates

The numbers are minium estimates not only because of the limitations of the statistical methods they used but also because very little is known about the behavior of even the most-studied fault, the San Andreas.

“Right now these are our best estimates, not handed down from God and not to be taken as rigid numbers,” said Stuart P. Nishenko, a U.S. Geological Survey scientist from Denver. He is one of a handful of researchers nationwide who are trying to use statistics to predict earthquakes.

In the U.S. Geological Survey group, Nishenko provided key statistical expertise for an effort that took known and inferred data about the faults and incorporated them into a mathematical process for determining how the likelihood of an earthquake increases with time. The method is based on how fast the fault slips during “quiet” intervals and how long it has been since the last quake.

But so little is known about Earth’s faults that the data supporting only two of 17 of the estimates were deemed “reliable” by group members.

An illustration of how uncertain the data can be was evident last November. Using the mathematical process, members tentatively determined that the Superstition Hills segment of the San Jacinto Fault near El Centro had a 20% chance of a magnitude 6.5 earthquake in the next 30 years. Later, on Nov. 23, they threw out the figure, deciding that the data behind it was too poor.

The next day, a 6.6-magnitude quake occurred on the fault.

Significant Leap

Nishenko sees such prediction problems as a reflection of how much more study needs to be done before prediction models will really work. In the meantime, the probability figures issued Thursday are a significant, if tentative, leap for the national earthquake hazards assessment program overseen by the U.S. Geological Survey, he said.

However, more information does not necessarily make the nascent earthquake-prediction field easier.

For instance, along the Mojave section of the San Andreas, the area nearest Los Angeles, the most recent studies have produced results so complex that they do not easily fit the prediction model used by the working group.

In today’s report in Science, researchers at Columbia and Caltech cite tree-ring evidence that an 1812 earthquake--thought to have been caused by a coastal fault--actually occurred on the Mojave section of the San Andreas.

The narrowness of the tree rings from that year told them that the trees’ roots had been disturbed by the earthquake. The quake was so large that it also destroyed the San Juan Capistrano mission 50 miles south of the fault section, which runs from Three Points to Cajon Creek northwest of San Bernardino.

That finding and an unpublished study dating quake-disturbed beds of prehistoric peat in the area have cut the average repeat time for large earthquakes there to 131 years, compared to 145 previously, Sieh of Caltech said.

Error of More Than 100 Years

Furthermore, the intervals between quakes during the last 10 episodes there were as short as 45 years and as long as 330 years, Sieh said. This suggests that using an average interval to predict a quake could carry with it a margin of error of more than 100 years.

Although Nishenko found that the altered numbers do not significantly affect the probability result along that section, they do show that the area does not play by a basic rule in his model--that earthquakes occur at roughly regular intervals.

“The common notion is that strain builds up along the fault, then it’s relieved completely in a big earthquake,” Sieh said. “In fact, if we have regular loading of the fault (with stress) but irregular occurrence, that means the strain is only partially released.”

Indeed, the 131-year average repeat time for the Mojave section would mean that--if the average held true every time--the next big quake there would be due this year. The last major event there was in 1857.

But because the quakes there seem to occur in close clusters of two or three, with clusters separated by longer intervals, the 131 years since the last big quake probably means that the Mojave section is in a quiet phase that could last for another century, Sieh said.

After detailing these results, the authors of the earthquake prediction study lament, “Paradoxically then, the segment for which we have the most . . . data yields a long-term hazard assessment that has considerable uncertainty.”

The report gamely continues that, “not without some trepidation,” the group had calculated the probability of a major quake along the Mojave section as 20% over the next 30 years.

Authors of the tree-ring study in Science were Gordon C. Jacoby Jr. and Paul R. Sheppard of Lamont-Doherty Geophysical Observatory at Columbia and Sieh. The second report, which established the more precise dates for quakes along the Mojave section of the San Andreas, was authored by Sieh and Minze Stuiver of the University of Washington and David Brillinger of UC Berkeley. The study has been accepted for publication by the Journal of Geophysical Research.

Andrews said that, because the group’s figures represent the best guesses scientists can make, the state will use the report to continue its ongoing program of promoting earthquake preparedness.