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.
OUTLOOK FOR THE FUTURE We can now assess the prospects and the promise of earthquake prediction on the basis of real data and observations. The Panel unanimously be- lieves that reliable earthquake prediction is an achievable goal. We will probably predict an earthquake of at least magnitude 5 in Califor- nia within the next five years in a scientifically sound way and with a sufficiently small space and time uncertainty to allow public acceptance and effective response. A program for routine announcement of reliable predictions may be l0 or more years away, although there will be, of course, many announcements of predictions (as, indeed, there already have been) long before such a systematic program is set up. Research on prediction continues, with definite successes and prom- ising prospects. To achieve an effective prediction system, more fun- damental research and field testing are required. There are many gaps and unresolved problems in our understanding of earthquake phenomena. How does the earth's crust behave before, during, and after an earthquake? How large are the stresses responsible for earthquakes? How do the physical and chemical properties of the inhomo- geneous crustal rocks change under stress in the earth? What type of observable phenomena do these changes produce? Neither the current theoretical models nor the available laboratory and field data answer all these questions. A better understanding of the whole process from the accumulation of strain to the dynamics of earthquake faulting is necessary for a scientific approach to earthquake prediction. The principal uncertainties in our knowledge concern two questions: (l) If well-identified precursory phenomena occur, will they, in fact, be followed by earthquake? (2) For the various possible precursory phe- nomena, how large and of what character must deviation from base-level values be before they can be regarded as true signals of an impending earthquake? Earthquake-magnitude predictions will most likely be based essen- tially on the duration of an observed precursory episode, as suggested by the Soviet observations, but many observations will be needed to achieve a high confidence level for such predictions. Another method . for estimating magnitude might be to determine the areal extent of ob- served anomalous phenomena; presumably, more-widespread areal anomalies will be associated with larger earthquakes because of the larger extent 3l
32 of faulting. This attractive possibility requires observational and experimental tests. Another attractive model, quite at odds with this one, suggests that the magnitude of an impending earthquake will depend on the dynamical conditions governing the extension of a fracture. This proposal suggests that magnitude/areal-extent relations hold only for smaller events and that larger events "break out" of a confined focal zone into regions that do not display precursory anomalies. Though short-term precursors have been reported for a few small and moderate earthquakes, we are still very uncertain about the precision with which such phenomena can be used to estimate the time of occurrence of a future large earthquake. For example, we do not know whether we could recognize precursory phenomena extending over a period of 40 years, or whether we could forecast the termination time of the anomalous period with sufficient precision to be of real value. Even if some pre- cursors of large earthquakes last too long to permit satisfactory predic- tion, numerous historical observations and instrumental measurements of anomalous phenomena, such as sudden uplifts of the land, occurring a few hours or days before great earthquakes give us hope that other pre- cursors can be found that will permit prediction without excessive tem- poral uncertainty. Study of these processes and effects may ultimately give us a capability for reliable prediction of moderate and large earthquakes. Recent intensive efforts to identify velocity changes prior to moderate-size earthquakes have resulted in many disappointments as well as encouragements, and we doubt that any single phenomenon will alone constitute a basis for a successful monitoring program for earthquake prediction. Our best prospect for reducing false alarms to a minimum is through a system that would monitor a wide variety of physical param- eters rather than rely on a single kind of observation such as velocity changes. Obviously, a much better understanding of the physical pro- cesses that occur before and during earthquakes would help immeasurably in determining which parameters should be monitored. It has become apparent from the available observations that earth- quakes may be preceded by different physical changes in different geo- graphic and tectonic regions. Observable earthquake precursors that are prominent in one region, such as the San Andreas fault in California, may not be observed in other regions, or even on other faults in Cali- fornia. For example, the kinds of velocity changes that preceded the Blue Mountain Lake earthquakes in New York State have not been observed as strong precursors of San Andreas earthquakes. On the other hand, anomalous ground tilts may have preceded some San Andreas earthquakes. It may turn out that earthquake prediction will be based on different sets of criteria in different regions rather than on a single universal criterion that applies to all regions. These differences will require monitoring in many different earthquake-prone regions if earthquake pre- diction is to be achieved in all vulnerable parts of the nation (or the world). Thus far, all documented earthquake predictions have been made on the basis of data and observations from dense networks of instruments in epicentral areas. The outlook for predicting earthquakes with instru-
33 ments remote from the epicentral areas, or "teleprediction," is not cer- tain. It is reasonable to assume that in the near future predictions will be made primarily by the networks in epicentral areas. Remote ob- servations may give some indications, however, and may be useful in identifying areas that are good candidates for close monitoring. On the basis of present experience and understanding, it is reason- able to say that reliable prediction of smaller earthquakes will precede that of larger earthquakes. Small earthquakes occur frequently, their precursors occur over a short period of time, and their sources can be defined with regional networks. Routine prediction of earthquakes of magnitudes 6 or less may be possible in well-instrumented areas within the next ten years. Large earthquakes occur infrequently, and may re- quire monitoring over a much longer period of time to test the prediction capability. Experience with smaller earthquakes no doubt will be applied to larger ones. To accelerate this process and to test tech- niques for predicting larger earthquakes, we must instrument and monitor different active areas simultaneously. Only in this way can we obtain adequate data about larger earthquakes during the next l0 to 20 years. In addition to long-term data that will be obtained from future in- struments, a significant history of seismological and earth-deformation data is recorded in existing bulletins and instrumental records. Al- though much of this older information is of insufficient precision or relevance to be of use in studying possible long-term precursors to major earthquakes, careful analysis of these existing records is impor- tant. For example, two relevant searches are those for possible anoma- lies in long-term tide-gauge data, and possible variations in earthquake travel-times to long-established seismographic stations in regions of large earthquakes. Earthquake control is likely to be farther in the future than earth- quake prediction. Nevertheless, prediction may permit identification of regions in which studies could be made of the feasibility of earth- quake control by fluid injection or by other means that may be devel- oped. Earthquake-control experiments have been tried successfully on a small scale in an oil field at Rangely, Colorado. Before earthquake control can become a reality, however, much more must be known about the physical processes involved, including the magnitudes of the stresses, permeability and porosity of rocks, and variations of fluid pressure along fault zones. Studies such as those at Rangely, as well as a number of laboratory studies now under way, should improve our knowledge in this areaâand ultimately our capability to modify or control at least some types of earthquakes. Additional studies in more representative tectonic regions must now be undertaken to fully explore the feasibility of earthquake control. The outlook for the future of earthquake prediction and control de- pends on the extent of the national commitment and the size of the national program. With optimum support, the routine announcement of re- liable predictionsâmeeting criteria of time, space, size, and probabil- ity of occurrenceâmay be possible in ten years in well-instrumented areas, although very large earthquakes may present special problems. The size of the program and the level of support must be consistent with
34 the magnitude of the task and the available trained manpower. In the interests of making the most significant advances in the shortest period of time, the wide variety of expertise currently avail- able in the United States in universities, private industry, and various federal and state agencies should be applied to the earthquake-prediction problem. We believe that a truly effective program will require a com- mitment for ten years that includes a large increase to several times the current annual funding. Much of the effort would be concentrated on intense instrumentation of a few experimental areas of high seismicity. If this l0-year research program is successful, subsequent implementa- tion of the resulting earthquake-prediction capability for all seismic areas of the United States, on a continuing basis, will probably require a comparable national annual commitment. Research in predicting earthquakes has progressed to the point where it is advisable that a continuing overview of the program be presented to the Executive Branch of the United States Government at regular in- tervals, since its ramifications will affect large segments of our so- ciety. The setting up of a representative group of competent scientists to accomplish this advisory task is one of the recommendations of this report. Similarly, planning should begin now for organized societal response to earthquake prediction. It is essential that we develop a prediction capability and effective response system in concert if the nation is to benefit immediately from this new development when it ar- rives. Now is the logical time to make a national commitment to an effective earthquake-prediction-and-response program, and to allocate the neces- sary resources to the task.
