Click for next page ( 21


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



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 20
3 Ground-Shaking Hazard This chapter examines the selection of scenario earthquakes. Use of scenario earthquakes is not the only way to address loss estunation, but it is the most common method. There are two general approaches to evaluating scenario earthquakes that are commonly referred to as deterministic and probabilistic methods, although elements of judgment and uncertainty are present in both. DETERMINISTIC METHODS In this method, one or more earthquakes are postulated with- out explicit consideration of the probability that those events will occur. The most common form of this method is use of the largest earthquake known to have occurred in a region, and this event is termed the historical maximum earthquake. This approach is based on a premise that is geologically sound as well as intuitively convinc- ing: if an earthquake has occurred once, it can occur again. Usually this approach is acceptable to both the governmental users of loss estimates and the general public. Once a decision to adopt this basic approach has been made, various questions must be answered in order to establish a scenario earthquake. For example, will it be assumed that the same earth- quake reoccurs with the same extent, location, and type of fault- ing? The distribution of ground-shaking intensities outward from 20

OCR for page 20
21 the earthquake may have been recorded, and can then be used di- rectly. If this distribution was not thoroughly recorded, it will be necessary to use attenuation relationships (derived from analysis of data from many different earthquakes) to calculate some or all of this distribution. Alternatively, it may be decided that a different location should be considered, perhaps closer to the region being studied. In this case, use of attenuation relationships to calculate intensities is essential. If there are multiple faults near the region being studied, it will generally be desirable to consider separately the historical maximum earthquake for each fault. This is because each of these several earthquakes may produce the largest losses in some portion of the region. In some studies, two leveb of earthquakes have been used: the historical maximum earthquake and a smaller earthquake chosen by judgment. The smaller earthquake has often been taken to have a magnitude one unit less than the historical maximum earthquake. This practice has been adopted when planning for a response to several levels of disaster is deemed desirable, or when a repetition in the near future of a large historical maximum earthquake lacks credibility. There are also instances where earth scientists present convinc- ing evidence that an earthquake larger or closer than the historical maximum event should be considered. This may happen when there is geological evidence of earthquakes more severe than those that have occurred in historic time. The proper characteristics of the scenario earthquake for use in planning how to respond to a validated earthquake prediction would be the predicted earthquake's magnitude, location, or other avail- able seismological information accompanying the prediction. Ex- cept for the greater potential for controversy concerning predicted earthquakes, the other aspects of loss estimation are the same for nonpredicted scenario earthquakes. It is clear that this deterministic approach involves some judg- ment and uncertainty. Even in the most seisrn~c regions of the coun- try, no one knows when the next major earthquake will occur or just what it will be like; almost certainly there will be surprises. There is ~ ~ ~ .. . ~ . ~ ~ . .' ~ . ~ ~ ~ no clear dennltlon of the largest possible eartnqua~c some expert can always envision a larger eventand even if there were a well- defined maximum earthquake, it is not obvious that this immense earthquake is the proper basis for hazard reduction planning. As one

OCR for page 20
22 moves away from use of the actual historical magnum earthquake, and as use of attenuation relationships comes into play, uncertainty increases. As stated earlier, it is desirable that at least a rough in- dication of the probability of occurrence be attached to all scenario earthquakes, if only to convey to users and the public some indication of the likelihood of such an event. PROBABILISTIC METHODS As just noted, there are two situations where attempts to use the historical maximum earthquake run into difficulties. At one extreme is the situation where a very large earthquake has occurred within- recorded history, but it is thought unlikely that it will reoccur soon and in the same locale. The other extreme is the situation where it is thought relatively likely that there can be an earthquake larger than the historical maximum earthquake. (UHistorical" merely refers to a brief sample of the geologic timespan, up to about 400 years in the eastern United States and 200 years in the West, and some earthquakes that occur only once every several centuries are unlikely to be included.) For such situations, it would be useful to have a systematic method for selecting the scenario earthquakes that meet the criteria of being plausible but damaging. Probabilistic hazard analysis offers this possibility, and is dis- cussed in a report of the National Research Council (1987~. The elements of this method are sketched in Figure 3-~. Information is required concerning: the location of potential sources (such as known faults) of earthquakes, the probability that different magnitudes will occur within or along each source, and the attenuation of intensity away from the source, including uncertainty in the attenuation re- lation. This information is then formally combined to produce a ground-shaking versus hazard curve (Figure 3-1D), giving the proba- bility that any ground-motion level will be exceeded. An exceedance probability is selected and the associated ground-motion level (target level) is found from the hazard curve. Finally, the scenario earthquake is defined as the most likely event among those that produce ground motions more intense than the target level. The technology for this type of analysis is well ad- vanced, although there are often problems of statistically inadequate data for evaluating parameters required by the theory. One difficulty in the use of probabilistic ground-shaking hazard analysis is in selection of the probability of exceedance to be used

OCR for page 20
23 A. Seismic Source i (Earthquake locations in space lead to a distribution of epicentral distances fR Erg m) ~ Rupture f fore m) ' Site a Ll Distance r C. Ground motion estimation: A ~ m, r > o .g - a, - o - Fault i GA~m,rta ) Distance (log scale) B. Magnitude distribution and rate of occurrence for Source i: f M (m), hi f M (m) mO mmax Magnitude m D. Probability analysis: P[ A > a. in time t] /t Vj JJGAIm,r (a*) fM (m) f R (r ~ m) dmdr - _ ~ c ce Cal 0 _ \ \ \\ Ground Motion Level a. (log scale) FIGURE 3-1 Graphs indicating probabilistic seismic hazard analysis steps.

OCR for page 20
24 for defining a scenario earthquake. There are no generally accepted rules for this purpose. Some of the historical maximum earthquakes used for earlier loss studies have annual probabilities of about .002, which is equivalent to a mean recurrence interval of 500 years. The pane! rejects the notion of a single standard probability at this time, but accepts that, in the absence of a suitable historical maximum earthquake, a scenario earthquake with an annual proba- bility in the range from .001 to .005 is reasonable for disaster response and mitigation planning. Despite the lack of definite criteria, use of probabilistic seisrn~c hazard analysis offers the only rational means for selecting scenario earthquakes for many parts of the country. DESCRIBING INTENSITY OF GROUND MOTION As noted earlier, there is no generally accepted, objective, quan- titative scale for measuring the damaging effects of strong ground motion. This is because different buildings, structures, or other facil- ities respond in different degrees to various aspects (e.g., predominant frequency, duration, and so on) of ground motion. Most U.S. loss estimates have used MMI as a scale for the in- tensity of ground shaking. This scale involves subjective evaluation of the elects of ground shaking, and its use is subject to abuse and misinterpretation. However, In most parts of the country the histor- ical seismic record is known only in terms of MMI. Instrumentally recorded strong-motion data are much more sparse. While urging continued research to develop a satisfactory quan- titative measure of ground-motion severity, the pane} accepts the continued use of MMI as a basis for the usual loss estimate study. One aspect of MMI that does require careful attention is the meaning and use of intensities XI and XIl. The scale's criteria for these levels emphasize observations of ground failure, some of which may occur when other indicators of shaking severity imply a MMI as low as VI. The use of high MM! values in a loss estimate requires explicit explanation to avoid misunderstanding. Some on the panel interpret the MMI scale as implying that intensity X represents maximum possible ground shaking. Others feel that ground shaking stronger than that associated with MMI X is possible, and there have been some instances in which loss estimators have used MMI XT and XII to represent increasingly strong ground shaking apart from ground failures.

OCR for page 20
25 The pane} recommends that MM! XT and XIT not be used to indicate increased intensities of ground shaking. If this is nonetheless done, it is essential that the interpretation of these intensity levels be set forth very clearly, and an explicit statement of how the MMI scale was interpreted should be included in any study where it is used. EFFECTS OF LOCAL SITE CONDITIONS Local site conditions can have a great effect on earthquake losses. Greater losses often occur because of ground failures, increased in- tensity of shaking for some soil and topographic conditions, and selective amplification of ground motion at the frequencies critical to structural response. It is import ant to take site effects into account in a loss estimate. While geotechnical data collected at individual construction sites can be very valuable in this effort, more general- ized geologic mapping of districts and zones in a city or region is also useful and can lead to refinements in seismic hazard analyses. The essential requirement is to make clear whether the inten- sity in a scenario earthquake applies to the ground as it is locally found (i.e., no further correction for local soil conditions required) or whether it applies to some standard ground condition and must be further modified for actual local conditions. This is a matter requiring good communication among seismologists, geologists, and engineers.