Earthquake Prediction Woes

A recent earthquake near Hollister, California may have done more damage to seismologists' hopes than to the surrounding countryside.

North of the city, the Calaveras and San Andreas fault zones intersect. A dense network of seismic and other instrumentation had been established there in hope of picking up any precursors to the next big earthquake. The earthquake came, one of magnitude 5.7, but no identifiable precursors came before it.

There were no foreshocks, no bulging or tilting of the earth, and no observed electric or magnetic field changes that could have been used to predict that a moderate earthquake was imminent. Even in retrospect, the only possibly significant events prior to the earthquake were a surge in flow rate from a spring some distance from the epicenter and reported unusual animal behavior. Still, it is uncertain that either of these was actually related to the earthquake's occurrence.

This failure to find definite precursors is counterbalanced by a success five years ago when a magnitude 5.2 earthquake near Hollister on Thanksgiving Day was successfully predicted on the basis of earth tilting and magnetic field changes. So far, this seems to be the story on earthquake predictions: some successes and some failures.

One result is much rethinking by seismologists on the nature of faults and how breakage occurs along them. Now failing is the simple concept of a big earthquake being caused by a clean break along a large fault surface uniformly stressed. Instead, seismologists are suspecting that fault zones are highly irregular with regions of high stress interspersed in regions of less stress. Twists and bends in the fault zones may be important too, since they may inhibit sudden breakages over large surfaces of a fault. All this could mean that rather elaborate, and perhaps impossibly difficult measurements might have to be made for 100% reliable predictions of earthquakes.