Seismic theory could help reduce threat from earthquakes

California earthquakes may rock and roll to a definite beat. That, at least, is the basic premise of a concept called the seismic cycle, a notion that may ultimately aid in efforts to reduce earthquake hazards.

So far this suggestion, that the level of earthquake activity along specific sections of major faults varies according to a predictable cycle, is what scientists consider a working hypothesis: an interesting idea which needs more support before it can be classed as a full-fledged theory.

Nevertheless, a groundswell of supportive evidence has built up in the two decades since it was proposed. Most recently, the Morgan Hill earthquake, which rocked the San Francisco Bay area late last month - the largest to hit the region since 1911 - fits into the pattern of increasing earthquake activity predicted for the area by this hypothesis.

The basic concept was first formalized by the well-known Soviet earthquake researcher S.A. Fedotov in the late 1960s. He suggested that on a given stretch of a major type of earthquake zone called a strike-slip fault great earthquakes - Magnitude (M) 8 or above on the Richter scale - come at the end of a gradual crescendo of smaller-magnitude temblors and are followed by a period of relative seismic calm.

(Strike-slip faults occur at the boundaries of the huge, rigid, slow-moving plates that make up the planet's crust. Although major earthquakes occasionally occur in the middle of such plates, the lion's share of all earthquake activity takes place at the edges where they grind against one another. Thus, most of California's earthquakes originate at the boundary between the leading edge of the North American plate and the side of the Pacific plate.)

Three years ago, William Ellsworth, chief seismologist of the US Geological Survey (USGS) in Menlo Park, Calif., and several colleagues looked at the earthquake history of San Francisco to see how it fit the seismic-cycle hypothesis. What records there are in the Bay Area appear to fit extremely well, Dr. Ellsworth reports. While the shortness of the earthquake record adds considerable uncertainty, the recent M 6.2 Morgan Hill earthquake has added confidence to this interpretation, he says.

Looking back, a dearth of earthquake reports in Spanish mission records for the area from 1776-1807 suggests a lull in tremblor activity during that time. This wasn't the case in the 50-year period before the great 1906 earthquake. During this time the Bay Area was racked by a number of large shocks ranging from M 6.6 to 7.0. Following the great quake, however, was another period of comparative calm. It wasn't until the 1950s, in fact, that the number of temblors in the M 5-6 range increased noticeably. And since then both tempo and the size of major quakes has been on the rise.

''It's not just the record. There are also convincing physical reasons why such a cycle might arise,'' Ellsworth elaborates.

The two plates which meet in the Bay Area are grinding past each other at a rate of 3.5 centimeters per year. Ten kilometers or so beneath the surface, the rock is hot enough and soft enough to slip easily. But the material nearer the surface is cooler and more brittle. As a result, the fault tends to lock up. When this happens, tremendous pressures begin to build. Finally, the strain becomes too much and the fault slips, releasing energy stored over decades in a matter of seconds.

A great earthquake is one that releases virtually all the strain on the fault. Although it begins building again almost immediately, considerable time must pass before enough energy has been stored to trigger more large quakes. This explains the seismic lull. Gradually, however, the level of subterranean forces grows. After a time, this becomes manifest in a series of increasingly violent quakes culminating in the next great event.

When the amount of energy stored in the earth is large enough, sooner or later a great earthquake must release it. Temblors of M 8 shake the ground 10 times more violently than M 7 events, but they release 32 times the energy. Once enough energy has accumulated for a great earthquake, experts consider it unlikely that it can be dissipated by a series of smaller events. Because the rate at which the two plates move is constant over a long period of time, and this is the basic source of the strain, there is reason to believe that the time between great earthquakes might be fairly regular.

The implications of the seismic cycle for the Bay Area are both good and bad, Ellsworth explains. On the plus side, it suggests that the time for the next 1906-scale earthquake is not until somewhere between the years 2000 and 2100. On the minus side, the area appears likely to experience a number of potentially destructive earthquakes in the M 6-7 range in the interim. And these are most likely to strike near Silicon Valley or in the heavily built-up East Bay where damage could run into the billions of dollars.

The seismic cycle seems to fit well in the Bay Area, agrees Kerry Sieh, a geophysicist at the California Institute of Technology who studies prehistoric earthquakes on the San Andreas Fault. It is one of a handful of spots around the world that clearly exhibit such a pattern, he explains. The other areas include a fault zone in western Japan that has cycled four times since the year 600, and an area in southern Chile which has been rocked every 130 years or so since 1575 by some of the largest temblors in the world.

''In these cases, the timing of the great earthquakes seem to be uniform, with an uncertainty of 30 percent,'' Dr. Sieh says. So, if the average spacing between great quakes is 150 years, then a succeeding earthquake might strike anywhere from 105 to 195 years later.

Sieh's work suggests that each section of a major fault may shake to the beat of its own drummer, depending on the specific nature of the overlying rock. For instance, there is a portion of the central San Andreas which is creeping fast enough so that it is not building up a significant amount of energy - the rock there seems too weak to accumulate much strain.

Then there is another section of the fault which passes through Parkfield, near Fresno, which has exhibited a shorter cycle: It builds up enough energy for a M 5-6 earthquake and breaks. ''The reoccurrences seem very similar, right down to details like the timing of foreshocks and the way it breaks,'' Sieh explains. TABLE OF RICHTER SCALE Magnitude Description 3.0-4.9 Frequent, unnoticeable, no damage 5.0-5.9 Noticeable, low to moderate damage 6.0-6.9 Large, moderate to severe damage 7.0-7.9 Very large, severe damage 8.0-9.6 Great earthquake Note: An increase of 1 order of magnitude on the scale means an increase of 10 times in the violence of ground shaking and 32 times in the amount of energy released.