September/October 1997


The Northridge earthquake of January 17, 1994 and the Kobe earthquake, which occurred one year to the day later on January 17, 1995, were both very damaging and disturbing events. The magnitude 6.7 Northridge earthquake occurred on a blind thrust fault that had not been identified as a potential seismic source, produced near-fault ground motions that were about 50% stronger than expected, and caused brittle failures in the moment frame connections of steel buildings located well away from the region of strongest shaking. Although the loss of life was moderate, in part due to its early morning occurrence, the earthquake caused direct losses estimated at about $30 billion. Although the magnitude 7 Kobe earthquake was not surprising geologically, it occurred on a relatively inactive fault in a region that conventional wisdom (and governmental policy) had grown to regard as being of low seismic risk. The scale of death and destruction inflicted by the Kobe earthquake was unprecedented in a modern urban environment, and it was profoundly shocking to Japan that such a disaster, which killed nearly 6,000 people and caused direct losses of $100 billion, could occur there.

The Northridge earthquake prompted a modest temporary increment in research funding to gather and analyze data in order to learn from the experience of the disaster, but research funding quickly returned to normal levels. In contrast, the Kobe earthquake caused a major transformation of earthquake hazard-mitigation research in Japan. Large capital expenditures were made in the immediate aftermath of the earthquake and are continuing today. For example, detailed seismic surveys were made to image the structure of the northwestern edge of Osaka basin, which is controlled by the fault system on which the Kobe earthquake occurred. This is important because the zone of intense destruction that ran through Kobe parallel to but offset from the fault is believed to have been caused by ground motions that were amplified by basin edge effects. The Science and Technology Agency installed a network of 1,000 digital strong motion instruments (the K Net) throughout Japan in a span of six months following the earthquake, making a total of five existing or planned national strong-motion networks operated by governmental or private organizations. The Science and Technology Agency is now planning an earthquake engineering testing facility that will cost about $400 million. The increase in research expenditures in Japan in response to the Kobe earthquake was about two orders of magnitude larger than that in the United States following the Northridge earthquake and has persisted much longer.

Why have these two earthquakes produced such different responses in the respective earthquake hazard-reduction research programs? One reason may be that in contrast to the unlucky "direct hit" that near-fault rupture directivity effects wreaked on the Kobe region, the 1994 Northridge earthquake was a lucky "near miss" in Los Angeles because strong rupture directivity effects were restricted to the mountains along the northern edge of the San Fernando Valley. However, there are many places in the United States that are just as vulnerable to near-fault rupture directivity effects as was Kobe, and the faults in many of these places have higher slip rates and hence generate earthquakes more frequently than Kobe. Thus, the different damage levels of the Kobe and Northridge earthquakes do not provide a rational justification for the discrepancy in level of funding for earthquake hazard mitigation research.

The discrepancy is more understandable from the perspective of national differences in seismic risk. In contrast with the United States, practically the whole of Japan is subject to severe seismic risk. What compounds the risk in Japan is that its capital, Tokyo, is particularly vulnerable (having last been destroyed in the great Kanto earthquake of 1923) and that too many functions vital to the livelihood of Japan are centralized in Tokyo. The United States has much more redundancy in its ability to withstand a single major earthquake disaster.

Still, these mitigating circumstances do not seem sufficient to warrant the discrepancy in level of funding of earthquake hazard-reduction research. Some thoughts on the adequacy of the National Earthquake Hazards Reduction Program (NEHRP) that I wrote about two years ago still seem to be relevant. These thoughts were prompted by questions that I was asked to respond to in preparing testimony for a hearing on the National Earthquake Hazards Reduction Program (NEHRP). The hearing was held by the Subcommittee on Basic Research of the Committee on Science of the United States House of Representatives on October 24, 1995. I went beyond my own area of practice in engineering seismology in answering some if the questions.

Is NEHRP research user-driven? Some is, a notable example being the National Earthquake Ground Motion Mapping Project sponsored by the USGS, which provides input into the development of seismic provisions for national building codes. Another is the FEMA-sponsored SAC Joint Venture Project described below. However, much NEHRP research is driven more by the interests of researchers than by the needs of users, and consequently much of it does not have and may never have the potential for practical application. Also, the NEHRP agencies that fund research are not currently well set up to manage applied research. One way of enhancing the relevance of research and enhancing its implementation is to involve practicing professionals together with researchers in applied research projects. A notable example is provided by the SAC Joint Venture project. Brittle fractures in steel connections during the Northridge earthquake have effectively invalidated current design methods for steel-frame buildings. The SAC Joint Venture, a joint venture of the Structural Engineers' Association of California (SEAOC), Applied Technology Council (ATC), and California Universities for Research in Earthquake Engineering (CUREe), is conducting a coordinated three-year effort to develop cost-effective and reliable guidelines and standards of practice for design of new steel-frame construction and the inspection, evaluation, and rehabilitation of existing structures. This project is a model of how applied research projects involving the collaboration of researchers and practicing professionals should be conducted.

Should the NEHRP program emphasize short-term, applied research such as microzonation rather than improving basic Earth science knowledge? I think it needs to both improve basic Earth science knowledge and support applied research. There is a danger that basic research will be neglected if too much emphasis is placed on short-term applied research. I expect that what we learn from basic research in the next few decades will be much more useful for seismic hazard mitigation than what we know now. At the same time, I think that microzonation is a potentially very effective short-term tool for risk mitigation. Microzonation is the mapping of seismic hazards, expressed in relative or absolute terms, on an urban block-by-block scale, based on local conditions (such as soil types) that affect ground shaking levels or vulnerability to soil liquefaction. It is motivated by the observation, common to all earthquake disasters, that severe damage tends to be concentrated in discrete zones which may be separated by relatively unscathed regions. By identifing the localities within a region that are most subject to seismic hazards, it provides an effective means of prioritizing mitigation actions which may otherwise be financially or administratively unmanageable in scope.

There has been criticism that many of the technologies and practices developed by the NEHRP program have not been implemented and that research is far ahead of implementation. It may be that much research lacks practical relevance, but I do not think that research is far ahead of implementation. On the contrary, in many areas the technical questions are beyond our present capacity to give useful answers, and research is only beginning to produce findings that can be implemented. Some of these basic research problems are: What is the physics of earthquakes, and how can we predict earthquake behavior and earthquake effects? How can we realistically model the behavior of structures during earthquakes? How can we economically reinforce existing structures? How can sociological knowledge be used to enhance the effectiveness of earthquake preparedness, earthquake response, and the implementation of mitigation measures?

Should the federal government put more teeth into earthquake hazard reduction? I think it should. The experience of the Kobe earthquake of January 17, 1995 suggests that the United States may incur direct economic losses of $100 billion dollars or more from moderate magnitude earthquakes occurring within urban communities, in addition to loss of life and indirect economic losses. I do not think that the resources that are committed to earthquake risk reduction in the United States are commensurate with this very high level of risk to life and economic health. The only comparable external threat to our society, short of war, is AIDS. To date, the direct cost of dealing with AIDS in the United States has been $75 billion.

What are the best ways to achieve results? I think some of them are:

  • Introduce legislation that mandates or provides financial incentives for the adoption of seismic codes and the implementation of mitigation measures.
  • Provide better coordination of applied research, and involve researchers and practicing professionals jointly in focused applied research projects like the SAC Joint Venture.
  • Fund activities such as seismic microzonation that identify zones of special vulnerability to seismic hazards and thereby help to prioritize hazard mitigation work.
  • Calibrate methods for earthquake loss estimation, use them to estimate losses that may occur in moderate and large earthquakes in the United States, and use these loss estimates to prioritize earthquake hazard mitigation programs.
  • Provide better mechanisms for the development of building codes, which are revised every three years. The appropriate development of these codes should be less dependent on the voluntary contribution of free time by busy practicing professionals, and more dependent on a rigorously reviewed and adequately funded process involving the collaboration of practicing professionals and researchers. Future codes need to be performance-based, and address nonstructural as well as structural damage.
  • Provide adequate funding for pure research, applied research and development, and implementation. The NEHRP program consists of a very capable and committed community of professionals, but their productivity is being severely limited by a shortage of funds to support research and implementation.

Paul G. Somerville
Pasadena, CA 91101 USA

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: 18 January 1999