The energy -- or magnitude -- of an earthquake is dependent on three factors: how hard the rock is (sheer rigidity), how fast the fault ruptures, and the size of the fault's rupture area (length times depth). Given this information from previous quakes in an area, scientists should be able to estimate the maximum size of future earthquakes.
This, scientists say, is more useful than predictions about when an earthquake will occur. If an engineer knows, for example, the maximum force a trembler will generate in a specific location, he can design structures to withstand it. At the same time, property owners can save money by not building stronger structures than needed.
Historical information about sheer rigidity and speed of rupture is generally available; however, fault area is hard to measure. Although the length of many faults is directly measurable at the surface, the depth of the seismically active part of a fault is not directly observable and is assumed to be the maximum depth of previous quakes. In areas where there are few records of earthquakes, the depth is often assumed to be constant.
University of Houston associate professor of geosciences Hua-Wei Zhou has found these assumptions about depth to be incorrect. Using a tomographic technique he developed to analyze seismic data from 37,000 earthquakes in southern California since 1981, Zhou has produced 3-D images of the subsurface which show that the maximum depth of the quakes varies widely both along and across the faults.
According to Zhou, the depth of an earthquake depends on the type of underlying rock structure. In areas with minerals that become plastic at lower temperatures, the maximum depth of quakes is shallower than in areas with harder basement rocks. Because the other factors which determine seismic depth (temperature, strain rate, and fluid pressure) are constant at a given depth, it is this lateral variation in rock formation which controls the depth, and thus the magnitude, of potential earthquakes. The composition of faults can therefore be used to estimate the potential magnitude of quakes.
Zhou hopes that this same technique will be able to predict maximum shaking in areas away from the fault itself. Then people will be able to judge risk more precisely and conduct better cost-benefit analyses prior to making construction decisions.
Harold Magistrale at San Diego State University assisted Zhou in analyzing the data for this study. The three-year project was 100 percent funded by the National Science Foundation, and the results were published in the August 2, 1996 issue of Science, the widely-circulated magazine of the American Association for the Advancement of Science.