4. Probabilities Based on Strain Accumulation

Scientists also use strain measurements to estimate the likelihood of future large earthquakes. When the strain in the rocks due to plate movement builds to a critical level, sudden slip results in an earthquake. Normally this slip takes place along one fault segment, but in very large earthquakes, about 7.5 or larger, more than one segment may move. The more slip that takes place during one earthquake, the more strain will be released. Thus, the longer, on average, it will be until the next large earthquake.

Using this method, the first step is to determine which fault segments have slipped in the past -- these are the segments most likely to slip in the future. Once a fault segment is identified, the potential magnitude of an earthquake on this segment can be estimated by determining the length of the segment. For example, when there is sudden slip on a fault segment 25 to 50 miles long in California, there is a magnitude 7 earthquake. A magnitude 8 earthquake typically results from slip on adjoining segments whose total length is 200 or more miles.

Anticipating when the next earthquake will strike along a given fault segment involves determining how much time has gone by since the last earthquake along the segment, how much strain was released in the last earthquake, and how fast the strain is building up along the segment. With this information, scientists have calculated the time required for the strain to grow to dangerous levels, typically 70 to 280 years along faults in the San Francisco Bay Area. Written history in California covers less than 250 years, but detailed geologic studies of fault zones have allowed the dating of a few prehistoric earthquakes.

The graph at right shows how the strain may have increased at an average rate of about 0.75 inches per year along the Santa Cruz Mountains segment of the San Andreas fault -- the segment that slipped on October 17, 1989, causing the Loma Prieta earthquake. Sudden slips reduced the level of strain along this segment during earthquakes in 1865 and 1906. The slip in 1906 was only about 5 feet, much less than the 15 feet of slip measured along parts of the San Andreas fault to the north. Based on this information, scientists suggested in 1981 that another damaging earthquake on the Santa Cruz Mountains segment was likely between then and 1996; the Loma Prieta earthquake proved their projections correct.

Although this strain model seems relatively simple, our information about strain is incomplete. Moreover, considerable judgment is required to determine the average time between large earthquakes on each segment, exactly where fault segments begin and end, the magnitude of anticipated earthquakes, the magnitude and amount of slip for some earthquakes in the 19th century that were not recorded by instruments, and the best statistical methods to use in calculating the probabilities.