July/August 1998


In a previous issue of Seismological Research Letters Paul Silver raised the question: "Why is earthquake prediction so difficult?" What makes the earthquake hazard so different from other natural hazards which are more susceptible to prediction efforts? While Paul raises valid arguments, I remain unconvinced that yet another multimillion dollar program is the solution to the problem. The fault may not lie in our lack of data, our inability to conceptualize the problem, nor in the lack of funds for whatever program may be preferred. I believe we need to refocus on the common goal of making earthquake prediction and forecasting a practical means to promote public safety.

Is this an obtainable goal? It has been in other fields, such as forecasting of weather, effects from solar storms, and to some extent volcanic and tsunami hazards. The history of how it came to be so in each of these fields shows that the process is, however, not natural. For example, during the first few decades of this century weather forecasting was done by private individuals for profit. Only with war did the importance of weather forecasting lead to the establishment of the National Weather Service. Likewise, the Pacific Tsunami Warning System was established only as a response to a highly destructive tsunami in 1946. Is an earthquake-related disaster far greater than Northridge or Loma Prieta required to bring us together in a cooperative effort at forecasting earthquakes? I hope we have the foresight to prevent this from happening.

There is an extraordinary interest in earthquake prediction by both professionals and nonprofessionals. The existence of large numbers of amateur predictions, some of which get much popular press coverage despite the efforts of the professional community to downplay their importance, is evidence of this interest. While we may often view the efforts of amateur seismologists as little more than irritants, amateurs are constructive in fields such as astronomy. They could have a similar role in seismology. The Public Seismic Network (PSN) is a possible mechanism to put the amateur seismologist to work. In China the public is encouraged to report anomalous behavior relating to possible precursors to earthquakes. From these observations, the Chinese have amassed archives of thousands of phenomena precursory to earthquakes and other natural disasters. The PSN could act as a vehicle for reporting of anomalous behavior as well as felt effects from earthquakes, a type of data the USGS has nearly abandoned in recent years. Furthermore, in regions of poor seismic coverage, the PSN could contribute seismograms and help determine epicenters.

I think we have all at one time or other during the course of our student years or professional careers dabbled, been interested in, pursued, or become obsessed with the problem of earthquake prediction. Most of us have even tried our hand at it. The result of these efforts was nearly universal failure, as the current state of earthquake prediction and the existence of this debate attests. It seems our own failures have predisposed most of us to take a dim view of efforts by others at earthquake prediction. Science is about "facts." But we weight the importance of "facts" through the filter of our experience. The experiential baggage we all bring to earthquake prediction causes us to view some facts, interpretations, or theories as important, others as useless or unimportant. We act according to these beliefs. For example, have you ever tried to find an unbiased evaluator for an earthquake prediction? We each color our evaluation with our previous experience.

If there is one thing we as earth scientists ought to have learned from experiences such as the plate tectonics revolution or the triggering of earthquakes several thousand kilometers from the 1992 Landers earthquake, it is that our present paradigms are not necessarily our future paradigms. Anyone who states that earthquakes are unpredictable will be found in the history of science alongside the physicists who at the turn of the century stated that all the important laws of physics had been discovered. Earthquake prediction is possible, albeit difficult, and perhaps not yet.

My optimism regarding the predictability of earthquakes is not unfounded. Previous work has shown statistical relations between earthquakes, weather-related phenomena, tectonic strains, magmatic movements, geomagnetic field changes, and other geophysical events. This suggests not only that there is a plethora of possible triggering agents but that the surface of the earth is sufficiently strained at all times that these small effects can and do trigger earthquakes. Recent work has suggested that earthquakes occur as part of a chaotic system. When a chaotic system like a loose pile of sand is near a major change, small effects can trigger the major event, and moderate-sized events occur with greater frequency as the major event is approached. I believe that short-term earthquake prediction will ultimately be achieved once we understand the triggering agents and their means of initiating seismic events, for it now seems likely that most earthquakes are triggered events.

The immediate causative agent may change the timing of the event by only a few days, however, so additional work is also needed to understand the intermediate earthquake forecasting process. The possibility of intermediate-range forecasting may also be within reach. A major problem has been the lack of earthquake catalogs which cover many seismic cycles. Recent work has produced usable long-term seismicity catalogs covering areas such as Turkey, New Zealand, North and South America, Europe, Japan, the Former Soviet Union, the Middle East, and China. These, combined with recent efforts in paleoseismology, make investigation of multiple seismic cycles possible in many areas of the world. Researchers will find these catalogs useful complements to the more commonly used catalogs from the United States.

The mathematics of chaotic systems is only beginning to reach the point where it can be applied to earthquakes. Fast computers, neural nets, and new computer algorithms allow us to examine concepts and data in ways unthinkable a decade ago. The history of science shows that great breakthroughs occur when new tools are combined with available data and innovative scientific thinking. Few problems in science cannot be overcome by enthusiasm, dedication, patience, cooperation, and intelligence.

The global nature of the earthquake hazard requires that there be international cooperation. This is best done through established organs of national and international governments. In the early years of geologic mapping, each state geological survey produced its state map without consultation with adjoining states. When maps were placed together, discontinuities existed at many state boundaries, hence state-line faults. Precursors to earthquakes have often been like state-line faults, never crossing international boundaries. Probably the best reason for international cooperation in the goal of useful earthquake prediction is hybrid vigor. We need new ideas to test our own. Earthquake prediction research is common in China and Russia, for example, and many innovative approaches and precursory observations are reported in the Journal of Earthquake Prediction Research, cosponsored by the State Seismological Bureau, China and the Academy of Sciences, Russia. I do not expect that most of my readers are aware of this English-language journal nor of many of the concepts on which articles in this publication are based.

As part of the International Decade for Natural Disasters Reduction (IDNDR), the United Nations with the assistance of the Chinese Academy of Science and the State Seismological Bureau has recently sponsored a workshop in Beijing, China with the intent of beginning an international process of natural disaster forecasting. The purpose of the workshop was to train participants from more than twenty countries in standardized statistical programs to evaluate forecasts and to watch for specific types of geophysical anomalies in their own region. Participants have now returned to their own regions and report anomalies through a central e-mail list server. They are urged to use the programs to evaluate and predict natural disasters and to report through the e-mail list. Several useful precursory anomalies have already been identified through this process. A second workshop on statistical earthquake prediction will be held in Hangzhou in June. These workshops provide the first step in a process which can lead to earthquake and other natural disaster forecasting on an international basis. I believe we should support efforts such as this. They cost us nothing (since we haven't yet paid our back dues to the United Nations), but they have significant potential for useful international scientific cooperation in the forecasting of natural disasters.

We are privileged to be among those few scientists who are working on a major and fundamental problem which will be solved within our lifetime. A thousand years from now people will remember that scientists of our era solved the earthquake prediction problem. We can choose to be a part of that process or a hindrance to it, but in the end a solution will be found. I wish I could enthuse all of you with the optimism I feel that this is a problem about to be solved.

Lowell S. Whiteside
Cooperative Institute for Research in Environmental Sciences
and National Geophysical Data Center
Boulder, Colorado

To send a letter to the editor regarding this opinion or to write your own opinion, contact Editor John Ebel by email or telephone him at (617) 552-8300.

Posted: 26 June 1998