Earth scientists studying the background of western North American earthquakes are revising their perception of the timing of the ``big ones.'' They also wonder if their understanding of quake-causing strain along California's San Andreas fault may be literally 90 degrees in error. They have learned just enough to show the need to pursue such research faster and more extensively in order to estimate the earthquake hazard realistically.
Geologists Kerry Sieh of the California Institute of Technology and Minze Stuiver of the University of Washington, working with statistician David Brillinger of the University of California at Berkeley, have refined their timetable for the last 10 big quakes along the San Andreas fault near Wrightwood, Calif. These quakes are reflected in sediment layers at the Pallett Creek ski resort.
Earlier estimates put the average time between big quakes at 145 years. Sieh and his colleagues have cut that to 131 years. More important, they find a suggestive clumping in the earthquake dates, which range from about AD 677 to the last big one on Jan. 9, 1857. They cluster in bursts of two or three shocks at relatively short intervals with periods of 200 to 300 years between bursts.
This clustering is further suggested by a study of tree-ring data Sieh made together with Gordon C. Jacoby and Paul E. Sheppard of Columbia University's Lamont-Doherty Geological Observatory. The 1813 growth rings of a set of ancient conifers growing along the San Andreas fault near Wrightwood reflect severe growth disruption. The scientists reported recently in Science that it probably is due to the large quake of Dec. 8, 1812, which had been wrongly attributed to a different fault system. This means two big quakes occurred on this section of the San Andreas within 45 years.
The last quake occurred on this fault section 131 years ago - the overall average interval. It is important to extend this research to try to determine whether the fault is in one of the periods where quakes occur relatively frequently or in one of the quieter periods with several centuries between shocks.
Using these new studies and other data, the US Geological Survey (USGS) last month refined its estimates of earthquake probability for California. Experts had considered the Los Angeles-San Bernardino area more likely than San Francisco as the site for the next big San Andreas quake, because so much strain was released in the great San Francisco shock of 1906. But the USGS now thinks San Francisco is almost as likely as Los Angeles for a major fault rupture.
Such estimates are still rough. And the USGS emphasizes the need for more research to better define the western North America earthquake threat.
Among other things, geologists need to know whether their general concept that strain builds up along the fault is correct. For a couple of decades, there have been indications that the main earthquake-causing strain is building across the fault. If this is true, it would change the way geologists estimate the likelihood of fault rupture. A consortium of 42 universities led by Stanford University geophysicist Mark Zoback has been drilling into the crust near the fault at Cajon Pass to measure the strain. But last April it had to stop short of the depth needed to settle the question when the National Science Foundation ran out of money for the project because of budget cuts.
This was false economy. A major California quake would be immensely costly. The better geologists can anticipate its nature and timing, the better authorities can prepare to minimize the catastrophe. There is a big economic and social premium on giving earth scientists the support they need to reduce the uncertainties in their earthquake knowledge.
A Tuesday column. Robert C. Cowen is the Monitor's natural science editor.