Those who believe that deterministic earthquake prediction, in a practical, useful form, is possible are forced to confront a long and dreary history of failure. Of the vanishingly small number of apparent successes, only one stands out, and it is the example most likely to be cited to refute the argument that useful predictions are, so far at least, impossible. The quake in question occurred over 30 years ago – the 1975 M 7.3 Haicheng, China earthquake. As the story goes, official government researchers combined their observations with those of a legion of citizen observers, and concluded that a major earthquake was imminent. Consequently, according to Wikipedia, “the China State Seismological Bureau ordered an evacuation of 1 million people the day before the earthquake.” It is usually accepted that this evacuation saved the lives of tens, or even hundreds, of thousands of people.

Much of the original information about this event relied on official, propagandized statements released by the Maoist government during China’s Cultural Revolution. China has, in recent years, become a more open society, and the authors of an article in the current (June 2006) Bulletin of the Seismological Society of America have obtained a great deal of declassified documentation and conducted numerous interviews with figures involved in the event. The article, “Predicting the 1975 Haicheng Earthquake,” written by Kelin Wang, Qi-Fu Chen, Shihong Sun, and Andong Wang, in my opinion, significantly alters, and reduces, the impact of the story on the whole issue of earthquake prediction. In particular, the authors’ findings are that there was “no official short-term prediction,” that “it was the foreshocks alone that triggered the final [evacuation],” that “local construction style and time of the earthquake also contributed to minimizing fatalities,” and “evacuation was extremely uneven across the disaster region and critical decisions were often made at very local levels.” In addition, they find that “to demonstrate the correctness of Chairman Mao’s ideology, the role of amateurs in monitoring precursory anomalies was exaggerated.” Readers of this forum are likely to be interested in the role played by such supposed precursory phenomena as animal behavior, groundwater levels, electrical signals (geomagnetic field and “telluric current”) leveling
measurements, “unusual light or sound,” and “changes in seismicity”. Prediction proponents who tend to fixate on these phenomena are likely to be disappointed that such evacuation as finally occurred was due to “the foreshocks alone” – particularly the foreshocks’ “large number . . . their increasing magnitudes, their consistent direction and distance . . . and . . . the damage they caused.” I submit that this foreshock sequence is highly unusual, in that it is not by any means a general component in the lead-up to major earthquakes, and that it is pedestrian in the sense that even prediction skeptics would not hesitate to take seriously such a pronounced and obvious sequence of events as occurred at Haicheng.

Of the other precursory phenomena, the authors single out only groundwater level and animal behavior as being of particular interest. However, even the observations of those are ambiguous and subject to interpretations which include the effects of such commonplace things as “vibrations caused by earthquake tremors that were not detected by the then very sparse seismic network.” One interesting paragraph in the BSSA article is the following:

Before the earthquake, there were a few sightings of groups of
disoriented mice unresponsive to people and cats around them. They
represent a tiny fraction of the total mouse population in that
area, and their behavior must have some site-speicific reason. One
possibility is that their dens were invaded by some toxic gas emitted
from the ground, perhaps due to fracturing before the large earth-
quake. Without knowing what happened to these mice, why most other
mice were not bothered in this way, it will be unfruitful to try to
use mouse behavior as an earthquake precursor.

The authors also point out that:

Earthquakes are sufficiently different from one another to cause
different anomalies. Earthquakes also modify geological structure
and affect certain rock properties (such as permeability and
porosity), such that the same anomaly may not occur, or occur
differently, for two consecutive earthquakes in the same region.
Besides, there are numerous anomalous phenomena that are not
followed by earthquakes.

That this is true is supported, I think, by the fact that, only 18 months after the Haicheng quake, a M 8 earthquake hit Tangshan, China without warning, killing over a quarter of a million people.

Mike Williams in Arroyo Grande, CA USA
"Prediction is very difficult, especially about the future" -Niels Bohr