predictions, such messages should provide maps of earthquake damage potential, including such near-instantaneous changes as landsliding and liquefaction. Within this framework, short-term predictions can provide crucial information, provided the users are properly prepared to receive it; preparation must be made through public education programs. California has begun to issue short-term earthquake advisories based on probabilistic models of foreshock activity. These low-probability advisories—a 2 to 5 percent chance of a larger event in a 3-day window—have proved their worth by spurring individuals and institutions toward earthquake mitigation actions they had previously ignored. More precise short-term predictions may, some day, again revolutionize the technical ability to give earthquake warnings. Even if and when this goal is achieved, the translation of scientific knowledge into planning, preparation, and action will remain the most important task for scientists and public officials alike.

Modern computer-networking capabilities promise a new form of hazard mitigation involving real-time seismology that can immediately identify the most severely damaged areas and assign emergency relief priorities. It is now technically possible to determine earthquake source parameters such as size, depth, and direction of rupture propagation immediately after a large earthquake. Large-scale deployment of the new generation of broadband seismic instruments with satellite or other telemetry capabilities is making this a reality. The intensity distribution of earthquake ground shaking often exhibits a very irregular spatial pattern because of variations in the crustal structure near the epicenter, in site response, and in the force mechanisms—such as strike-slip, along the San Andreas Fault, or dip-slip, in western Mexico. When seismic networks are supplemented by portable instruments, generic path effects and site responses in an earthquake-prone region can be defined. It then becomes possible to quickly determine the spatial distribution of ground shaking and the damage exposure in an entire epicentral region.

Instantaneous communication capabilities offer further opportunities to mitigate earthquake damage. In cases where the earthquake is centered some distance away, it is possible to warn a region of imminent strong shaking as much as several tens of seconds before the onset of damaging shaking because of the relatively slow speed of seismic waves. Endangered regions could respond to the warnings in time to shut down delicate computer systems, isolate electric power grids and avoid widespread blackouts, protect hazardous chemical systems and offshore oil facilities, and safeguard nuclear power plants and national defense facilities.

A simple warning system using this concept has already been incorporated into the Japanese railroad system. A similar strategy is being implemented in the form of a tsunami warning system that will alert areas around the Pacific when any large submarine earthquake occurs in the Pacific basin. Real-time seismic and geodetic systems are also critical components of volcano monitoring systems that can provide fairly short-term warning of impending explosive eruptions.

Tsunami Hazards

A tsunami along several hundred kilometers of the coast of Nicaragua on September 2, 1992, resulted in over 100 deaths as a 25-foot wave inundated coastal areas (Figure 5.16) and in places had run-ups of up to 1,000 m. Tsunamis are large ocean waves most commonly generated by the uplift or depression of sizable areas of the ocean floor during large subduction-zone earthquakes; significant tsunamis have also resulted from volcanic eruptions and large landslides or submarine slides. Like earthquakes, great sea waves are of little consequence in remote regions. In the open ocean they are hardly noticed, but as they approach the shore the waves increase in amplitude as they move into shallower water, depending on the nature of the local submarine topography. The resulting tsunami hazard in many coastal areas is far greater than is often appreciated. For example, during the great Alaskan earthquake of 1964, the loss of life from the tsunami generated in the offshore area was more than 15 times as great as the loss of life directly attributable to earthquake shaking; much of it occurred far from Alaska. Indeed, during the past 50 years, significantly more people have been killed in the United States by tsunamis than by other effects of earthquakes—although these statistics could change radically overnight with a major earthquake in a metropolitan area.

The areas of the United States most affected by tsunamis are Alaska, Hawaii, and the Pacific Northwest. Hilo, Hawaii, has been hit repeatedly by tsunamis originating as far away as southern Chile, and the same Chilean earthquake that produced tsunami devastation in Hilo in 1960 caused 200 deaths in far more distant Japan. Very recent geological field studies suggest that large prehistorical tsunamis occurred along the Oregon-Washington-British Columbia coast, probably generated by



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement