This website is no longer actively maintained
Some material and features may be unavailable
PBS Online Savage Earth Logo Thirteen / WNET
Vertical Marker
The Restless Earth: Earthquakes
Back to Home
Hell's Crust: Our Everchanging Planet
The Restless Planet: Earthquakes
Out of the Inferno: Volcanoes
Waves of Destruction: Tsunami
Article Sidebar One Sidebar Two SidebarThree Animation
Quake Prediction

Quake Prediction

by Daniel Pendick

Can earthquakes be predicted? Many seismologists would probably answer, "Not yet, but eventually." But to date, nobody has been able to predict earthquakes reliably enough and over short enough time scales to allow the evacuation of threatened cities. Some scientists have entirely lost faith in earthquake prediction. They say that so many factors decide whether a fault will rupture that earthquakes could well be inherently unpredictable in a practical sense.

One basic idea behind quake prediction is that faults send out subtle but detectable warnings before they slip. Scientists have looked at a host of potential warning signals, or "precursors," including foreshocks, weird animal behavior, and changes in the water table, stream flow, well levels, and patterns of electrical currents in the ground. In 1975, the Chinese government made a successful prediction based on precursors like foreshocks, and as a result casualties from a magnitude-7.3 quake in the Haicheng-Yingkow region were relatively light. In the United States, scientists have hoped to predict quakes by noting changes in the speed of seismic waves passing through at-risk faults.

These methods have not really panned out, however. In 1976, Chinese earthquake prediction suffered a major blow when the Tangshan earthquake in northern China struck unpredicted and killed at least 240,000 people. (Earthquakes undoubtedly do have precursor signs, but they tend to be clearer in hindsight, and it is difficult to sort out signs that may be meaningful from ones that prove not to be.)

Man with equipment

A Chinese seismologist studies quake patterns.

Meanwhile, in 1983, scientists with the U.S. Geological Survey predicted that a moderate earthquake was due to strike near Parkfield, California. The prediction was based on the observation that earthquakes with magnitudes of about 6.0 had occurred there in 1857, 1881, 1901, 1922, 1934, and 1966. On average, this is every 22 years, so the prediction was made for 1988 plus or minus five years. A large research effort was mounted to monitor the region. When the quake did not hit by 1993, the prediction was cancelled.

The Parkfield incident contributed to an erosion of faith in specific earthquake prediction. But many seismologists are still hopeful about general earthquake forecasting. Instead of predicting specific events over short time scales (hours to days), the scientists hope to forecast the probability of earthquakes over longer periods. In 1988, the USGS said there was a high probability of a major earthquake in the Santa Cruz Mountains region sometime in the following 30 years. Sure enough, the Loma Prieta quake rumbled through the region in 1989.

These probability forecasts are based on the idea that a fault builds up strain until it reaches a critical point, when it is released as an earthquake. Then the whole process starts over. The cycle won't necessarily be perfectly regular, like the ticking of a clock. But it will repeat in a predictable way, say proponents of forecasting. According to this theory, faults that haven't had a quake in the longest time are most at risk.

This theory has been conventional wisdom for thirty years. However, it did come under attack recently when UCLA seismologist David Jackson and colleague Yan Kagan scrutinized a global set of forecasts made in 1979. They found that areas thought to be at low risk of earthquakes -- the ones that had recently had quakes -- actually experienced five times as many shocks as perceived high-risk areas. The seismology community is still debating the issue.

Still, the dream of quake prediction is very much alive and well in some quarters. A group of Greek scientists claim to be able to predict specific earthquakes based on changes in electrical activity in the ground and say that they have done so three times in the last decade, although their methods and evidence have been sharply criticized. And regardless of the ongoing controversy surrounding the Greek predictions, many seismologists believe the problem is not the inherent unpredictability of earthquakes, but our ignorance of the processes that trigger them. In any case, as long as earthquakes remain deadly and destructive, we will dream of being able to foresee the next "big one" coming before it strikes.

Article: All Stressed Out | Sidebar One: Learning from Earthquakes | Sidebar Two: Quake Prediction | Sidebar Three: Build Smart | ANIMATION
Hell's Crust: Our Everchanging Planet  |  The Restless Planet: Earthquakes
Out of the Inferno: Volcanoes  |  Waves of Destruction: Tsunamis


PBS Online   |   Thirteen Online