that a magnitude 6 or greater earthquake would strike the region in 1993. In the mid-
1980s, seismologists with the United States Geological Survey set up an enormous number
of instruments along the Parkfield section of the San Andreas to monitor the expected
earthquake. While they were right that an earthquake was due in that segment of the fault,
they were quite far off in predicting the earthquake’s timing. A magnitude 6.0 quake did
not strike Parkfield until 2004, 11 years late.
Scientists have recognized some indicators that allow them to recognize that a large earth-
quake is likely. Large earthquakes are often preceded by small tremors, called foreshocks,
that occur between a few seconds and a few weeks before a major quake. However, many
earthquakes are not preceded by foreshocks and clusters of small earthquakes are not neces-
sarily followed by a large earthquake.
Large earthquakes are also often preceded by the tilting of the ground surface, which is
caused by the buildup of stress in the rocks. Seismologists measure the ground tilt and use
the changes to predict an impending earthquake. While this technique has been somewhat
successful, it has also been a part of predictions of earthquakes that never came and has
failed to predict some that did. Water levels in wells fluctuate as water moves into or out
of fractures before an earthquake. This information can also be used as a possible, but
uncertain, predictor of large earthquakes.
The most successful earthquake prediction was on February 4, 1975. At the recommendation
ofChineseseismologists, officialsevacuatedmanyoftheresidentsoftheManchurianprovince
of Liaoning. Although the region was not prone to earthquakes, the seismologists made their
prediction because the area experienced about 400 small foreshocks over a few days. The
night of the evacuation an earthquake of magnitude 7.3 struck the town and only a few
hundred people died. An estimated 150,000 people may have been saved. However, a little
more than a year later, Chinese seismologists failed to predict the Tangshan earthquake,
which killed more than 250,000 people. One month after that, Chinese officials evacuated
residents of the Guandong Province for an earthquake that never came.
There is value in predicting the arrival of seismic waves from an earthquake that is already
taking place. Seismometers can detect P-waves a few seconds before more damaging S-waves
and surface waves arrive. Although a few seconds is not much, a coordinated computerized
system can use that time to shut down gas mains and high tension electrical transmission
lines, and initiate protective measures in chemical plants, nuclear power plants, mass transit
systems, airports and on roadways.
Folklore tells of animals behaving erratically just before an earthquake. Mostly these anec-
dotes are told after the earthquake, when people remember back to the time before the
shaking began. Memories are notoriously faulty. However, Chinese scientists actively study
the behavior of animals before earthquakes to see if there is something to the anecdotes.
One interesting tale involves the number of animals killed in the 2004 Boxing Day Tsunami,
which appeared to be surprisingly low. Reports abound suggesting that the animals had
a “sixth sense” that warned them of the danger. In Sri Lanka’s Yala National Park, for