Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
SOCIAL IMPLICATIONS Successful earthquake predictions can lead to great reduction in loss of life, to smaller but still important reductions in quake-caused property damage, to enhancement of the margin of safety of critical facilities such as dams and nuclear reactors, and to more-effective and rapid restoration of normal living after the quake. The long- range goal of social policy with respect to earthquake predictions should be realization of these benefits. Little social benefit can result from earthquake predictions, no matter how accurate and precise they are, unless careful planning for response to those predictions has been undertaken and unless appropri- ate response agencies are prepared to implement those plans. Such planning must take into account uncertainties about the time and mag- nitude of a predicted quake. And the design of prediction systems intended for operational use rather than for research should be re- sponsive to the needs of social-response agencies as well as to the state of prediction technology. Much can be learned from past and current successful prediction-and-response programs for floods, hurri- canes , and tornadoes. Several recent studies have focused on prediction of and public re- sponse to natural disaster (White and Haas, l975), and the recent report on Earthquake Prediction and Public Policy, prepared by the NRC Panel on the Public Policy Implications of Earthquake Prediction (l975), focused specific attention on the social, economic, political, and legal implications of earthquake prediction. Our brief discussion, consistent with the conclusions of the latter report, will address primarily the interface between systems for predicting earthquakes and systems for societal response. Prediction capabilities will depend, for a long time to come, on relatively dense instrumentation of highly seismic areas. At present, amply instrumented regions in the United States have been chosen more on the basis of seismic activity than of social importance. Conse- quently, early successful predictions are likely to be for areas of relatively low population density. Eventually, a decision will have to be made about when and where to install instrumentation intended primarily to provide socially useful warnings rather than research data. The time for that decision may be as much as ten years away, but in any , 27
28 case, such a decision would be inappropriate until the whole warning- and-response system has been carefully thought through and planned. Public Release of Predictions Scientists and administrators concerned with the possibility of drastic public reaction to unsound or premature predictions have debated whether earthquake predictions should in fact be released to the public until a formal mechanism has been established for doing so. Earthquake predic- tions cannot be kept secret; the word will get out one way or another. Moreover, issuing a prediction publicly puts a burden of proof on the competent scientistâputs his scientific reputation on the line, so to speakâand may therefore encourage a strong sense of responsibility about the technical basis and timing of a prediction. Unsound and premature predictions will certainly be made, and public response to a prediction is likely to be very expensive. Decision-makers therefore need guidance about whether to respond to an earthquake predic- tion, and how to respond, if necessary. In the Panel's opinion, the best compromise between the scientist's freedom to make his view public and society's need to be protected from costly responses to false alarms is to form an official body to scrutinize and evaluate such predictions as soon as possible after they are made. If such a reviewing body con- cludes that a prediction is not well grounded in evidence, that conclu- sion, reached in time, is likely to obviate the costs of a needless social response. Such a conclusion about one prediction would almost certainly lead to social discounting of future predictions from the same source. The possibility that this might happen should tend to en- courage a high level of responsibility among predictors. If such a re- viewing body should in effect endorse a prediction, on the other hand, undertaking an appropriate social response to that prediction would then become an urgent task. With or without such a reviewing body, predictions will undoubtedly be made over the next decade by competent scientists and agencies, and by others. The Panel strongly believes that earthquake predictions should be accompanied by estimates of confidence level and by suffi- cient backup data so that their merits can be evaluated. One advantage of acquiring experience in social response to earth- quake predictions even before operational prediction systems are de- veloped is that it can offer guidance concerning response to false alarms as well. In any prediction system, including an operational one, false alarms are inevitable. Their numbers can be decreased only by increasing the weight of evidence required before a prediction is issued, thus increasing the possibility that a real earthquake may not be pre- dicted or that its prediction may be later than it might have been other- wise. But too high a false alarm rate is likely to lead to public complacency and thus make more difficult the task of public agencies trying to respond to warnings. The appropriate trade-off between false alarms and unpredicted earthquakes is a difficult question of social policy. Any evidence about response to false alarms that can be gathered
29 will be helpful in the design of an operational prediction system. In any case, that operational system will have to be protected, both in its procedures and in its public representation, from the consequences both of false alarms and of unpredicted earthquakes. If a useful but fallible technology exists, its failure in one or more instances should not be allowed to prevent its later use. The experience of the National Weather Service with false alarms and missed predictions, as well as with successful predictions, in its attempts to predict tornadoes indi- cates that such problems are not insoluble. Moving Toward an Operational Earthquake Warning System Thinking about the design of operational earthquake-warning-and-response systems should begin now, in time to permit thorough cost-benefit analysis of alternative designs. Such design thinking and cost-benefit analysis should be done by collaborating groups of experts, including seismologists, earthquake engineers, experts on social warning systems, other social scientists, and representatives of the governmental agen- cies, especially state and local, with which such operational systems must routinely interact. It seems likely that an operational prediction-and-warning system will issue seismological data routinely and special warnings when appro- priate. Warning categories will probably be small in number but very explicit, as are those now used for hurricanes and tornadoes, and re- sponse agencies should have pre-planned lists of things to do in re- sponse to each category of warning. These definitions and boundaries will help seismologists and others in converting earthquake prediction as a scientific achievement to its use as a social tool. Response Agencies The report by the NRC Panel recommends that the primary responsibility for planning and responding to earthquake predictions should be assigned to federal, state, local, and private agencies with broad concern for community and economic planning and for disaster preparedness and re- sponse, rather than to newly formed agencies established especially to deal with earthquake prediction and warning or to agencies primarily concerned with emergency response. Any operational prediction agency should be organized around seismo- logical rather than political boundaries. Response agencies, on the other hand, are mostly under local control, organized in a manner that respects political boundaries. Thus, a potential conflict exists be- tween the geographic basis of a prediction and the jurisdictional bound- aries of response agencies. While the resulting political problems could presumably be worked out for predictions with a reasonably remote date or a relatively large time uncertainty, they would be more diffi- cult to resolve under the time pressures that would be generated by a prediction that an earthquake is only hours away. Consequently, social
30 responses to earthquake predictions can probably be considered under two headings: planning for future earthquakes, and responding to imminent earthquakes. The problem of command and control for the immediate pre-earthquake and post-earthquake periods will probably require special organization. Needs will exist for backup communications, on-the-scene law enforcement, fire control, utilities management, and the ability to quickly muster recovery teams with heavy equipment. One of the principal concerns of individuals in or evacuated from a disaster area is protection of life and property. Compliance with emergency measures is much more likely if such protection can be relied on. Perhaps the most serious potential undesirable consequence of earthquake-prediction capability is complacency about other kinds of preparation for earthquakes. Prediction cannot prevent earthquakes, so good earthquake engineering design in siting and building, and appro- priate land-use planning, will continue to be as indispensable when a prediction capability exists as they are now. Since we have relatively little experience in planning for the con- sequences of earthquakes, and still less in planning for exploitation of a prediction capability, major efforts to study these problems in advance are needed now. Investigative groups should be ready to study social response to earthquakes as they occur. As social-response mechanisms evolve, simulations will help to study their effects and to train those who must operate them. Earthquake prediction, like other technological capabilities, can be used well or poorly. Used well, as we have said, it holds great poten- tial for saving lives, reducing property damage, and smoothing the return to normal post-earthquake living.
