Click for next page ( 7


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 6
Introduction An earthquake loss estimate is a description or forecast of the effects of future or hypothetical earthquakes. Loss generally encom- passes deaths and casualties; direct repair costs; damage or functional loss to communication, transportation, and other lifeline systems; emergency response and emergency care facilities; the number of homeless people; and the impact on the economic well-being of the region. Earthquake Tosses may be estimated to: Identify especially hazardous geographical areas; . Identify especially hazardous groups of buildings or other structures; Aicl in the development of emergency response plans; Evaluate overall economic impact; ~ Formulate general strategies for earthquake hazard reduction, such as land-use plans or building codes, or evaluate the effectiveness of earthquake programs; . Support advocacy efforts aimed at establishing priorities and budgets for earthquake programs; ~ Aid in obtaining quick estimates, made during the first hours following an actual earthquake, of the approximate impact of the earthquake; and quake. Estimate the expected consequences of a predicted earth- 6

OCR for page 6
7 The estimation of property losses to assess property insurers' risks has been one of the more common uses of earthquake loss estimates, but is only lightly addressed in this report because the emphasis here is on the broader range of public agency uses. This report focuses on Toss estimates of the type being funded by the Federal Emergency Management Agency (FEMA). They are intended for local and state government use, primarily for disaster response planning and to aid in the formulation of near- and Tong- term strategies for earthquake hazard reduction. This type of large- scale loss estimate study encompasses a city, region, state, or even the nation, and it looks at more than one type of loss, typically including life loss or casualties, property Toss, and functional loss or outages of essential services. A number of such studies have been completed or are under way. Figure 1-1 illustrates the geographic scope of past or in-progress large-scare Toss studies, while Table 1-1 lists these major studies. During the 1970s, the National Oceanic and Atmospheric Ad- min~tration (NOAA) and the U.S. Geological Survey (USGS) as- sembled teams of experts, predominantly engineering consultants and federal government geoscientists, who produced large-scale Toss studies that set the basic pattern for the scope and methods of others to follow. The first four were devoted to the metropolitan areas of San Francisco (AIgermissen et al., 1972), Los Angeles (AIgermissen et al., 1973), Puget Sound (Hopper et al., 1975), and Salt Lake City (Rogers et al., 1976~. These are sometimes collectively referred to as the NOAA-USGS studies. Some of the more recent studies have been sponsored by FEMA and carried out by consulting firms. In response to a National Security Council request for an eval- uation of potential impacts on the defense industry impacts, FEMA also initiated a recent large-scare effort aimed at modeling the re- gional economic effects of a major earthquake. This effort involved a study by the Applied Technology Council (ATC) of methods for preparing an inventory of facilities and for estimating damage and functional loss. The result was a report, Earthquake Damage Eval- uation Data for California, known as ATC-13 (Applied Technology Council, 1985~. FEMA also began in-house efforts and supported work by consultants to apply these new methods to selected eco- nomic sectors and regions. Differences exist among the techniques employed in these studies, arising from different levels of earthquake risk in various parts of the country, different objectives and budgets, and different authoring

OCR for page 6
\; at ~ ~: Air L7' 8 ~13 so C) o , - au - o o o 0 4a ~ o J ret {~ ~ C . a, . <, ,; . ~ ., so ~ `e Jo o 4, 3 d ~ ~ V ~ ~ 3 ~ _ U] . - ~ d 0 .~ lam =) d 0 =8111, ~ O I, ~ ._ _' o O ~ Cal

OCR for page 6
9 TABLE 1-1 Areas of the United States Where Large-Scale Lose Studies Have Been Completed or Are In Progress Areas Study 1. San Francisco, California 2. Los Angeles, California 3. Puget Sound, Washington 4. Salt Lake City, Utah 5. Honolulu, Hawaii 6. Central United States 7. Anchorage, Alaska 8. Boston, Massachusetts 9. Charleston, South Carolina 10. Puerto Rico and Virgin Islands 11. Clinton County, New York 12. San Diego, California Algermissen et al., 1972; Debris et al., 1982b; FEMA, 1980; Steinbrugge et al., 1981; Steinbrugge et al., in progress Algermissen et al., 1973; Blume et al., 1978; FEMA, 1980; Steinbrugge et al., 1981; Davis et al., 1982a; Scawthorn and Gates, 1983; Degenkolb, 1984; California Division of Mines and Geology, in progress Hopper et al., 1975 Rogers et al., 1976; U.S. Geological Survey, in progress Furomoto et al., 1980; Steinbrugge and Lagorio, 1982 Mann et al., 1974; Liu, 1981; Allen and Hoshall et al., 1985 Alaska Division of Emergency Services, 1980; URS/Blume, in progress Whitman et al., 1980; URS/Blume, in progress Lindbergh et al., in progress Geoscience Associates, 1984 and 1985; Molinelli and Oxman, in progress Geoscience Associates, in progress Reichle et al., in progress = . aNumbers correspond with studies noted in Figure 1-1. Organizations. Hence, inconsistencies can be found among the results of the various studies, and no clear guidance exists for conducting such studies. FEMA anticipates the need for future loss estimation efforts. Seeking to encourage studies that are done in a technically sound, ef- ficient, consistent manner that will satisfy the needs of users, FEMA asked the National Research Council to provide "evaluations and rec- ommendations with regard to methodologies which should be used for earthquake loss estimation by FEMA and state and local govern- ments in earthquake preparedness and mitigation planning." This work statement for the council's Pane] on Earthquake Loss Esti- mation Methodology, within the Committee on Earthquake Engi- neering, required that the applicability of recommended methods be

OCR for page 6
10 nationwide in scope, or that advice be provided for modifying recom- mended methods to fit regional variations. In addition to reviewing present methods, FEMA requested recommendations for testing and further development of methods to produce more accurate and com- prehensive loss estimates. The next section of this chapter presents an overview of the basic method used to carry out a loss estimate. This is followed by a discussion in Chapter 2 of the purposes and nature of Toss estimates as viewed by potential users, and then by more comprehensive reviews of the techniques and methods available for completing the several parts of a loss estimate. Recommendations for research and development leading to better loss estimates are given in Chapter 9. Seven working papers in Part IT treat many subjects in more detail. Several important points of a general nature must be emphasized: . The methods examined in this report rely on averaging dam- age and losses over a large group of facilities, and hence apply to groups of facilities and not to individual buildings. There are tech- niques for examining in detail the seismic resistance of individual structures, and brief reference will be made to such techniques. How- ever, any such detailed analysis can be expensive and time consuming and therefore generally is not feasible as part of a large-scale study. When methods intended for large numbers of buildings are used to estimate Tosses for individual buildings, the results may be m~slead- ing. . This report emphasizes large-scare loss estimates, the basic method and some of the detailed techniques of which are applicable to other types of studies. . No loss estimate prepared today, or in the foreseeable future, can be completely accurate. There are major gaps in our knowledge, both as to the time of occurrence, magnitude, and location of future earthquakes and as to the manner in which the ground and structures will respond to earthquakes. Any loss estimation inherently involves significant uncertainties. . Despite their limitations, loss studies that are properly con- ducted and used with an understanding of the methods' limitations can be of great value. These studies have played an important role in developing earthquake programs throughout the country, and are an important too! for initiating effective programs in areas where earthquakes are a significant threat but have received little atten- tion, or where few practical hazard reduction or emergency planning countermeasures exist.

OCR for page 6
11 . Loss studies in and of themselves do nothing to reduce seis- mic risk unless they lead to implementation of hazard reduction or emergency planning measures, or facilitate the development of pub- lic policy. Earthquake loss estimation is an important preliminary step toward taking appropriate actions for earthquake loss reduction. This is the most basic purpose underlying earthquake loss estimation. We study earthquake losses so they can be reduced. BASIC METHOD As previously noted, earthquake loss estimates may be made for many different purposes. Thus, studies may differ as to the types of Tosses considered, the extent of the geographical area involved, and the kinds of facilities included. Facilities is a term of broad scope that includes buildings as well as other structures such as bridges and utility stations and lifeline systems such as water distribution networks and airports. The detail in which the analysis is carried out and the manner in which the losses are aggregated and presented also may vary. (Working Paper A presents a scheme for categorizing loss estimates and briefly discusses examples of each category.) Although the techniques used to carry out various types of studies may differ, a basic underlying method is common to airnost all loss estimation studies. The Tao Main Components of an :Earthquake Doss Estimation Study Figure 1-2 illustrates two components comprising the basic struc- ture of a loss estimation study. One component, the seismic hazard analysis, involves the identification and quantitative description of the earthquake (or earthquakes) to be used as a basis for evaluating losses. This part of the study falls primarily within the disciplines of geology and seismology, and this geoscience effort must be coor- dinated with input from the broad field of civil engineering. The phrase seismic hazard might seem to refer to all hazards to life and property posed by earthquakes, but the term has a technical meaning restricted to the behavior of the ground, apart from any effects on the built environment. The second component, the vulnerability analysis, entails analy- sis of the vulnerability of buildings and other man-made facilities to earthquake damage and the losses that may result from this damage.

OCR for page 6
12 I r SEISMIC I I VULNERABILITY \ / LOSS ESTI MATE FIGURE 1-2 Basic structure of an earthquake loss estimation study. This effort pr~rnariTy involves engineers, architects, and experts in local real estate patterns or socioeconom~cs, although other disci- plines (e.g., utility system operators, urban planners, and disaster preparedness and response specialists) may contribute to identifying steps that can alter the losses caused by damage. The information assembled from these two components is com- bined to produce the loss estimate. Close communication among the . . . . . . . .. . . . .. .. . . . technical people un~ertaxmg the two parts, and Wltn the intended users, is vital to ensure proper coordination. In most loss estimates, the primary emphasis is on damage and Tosses caused directly by the shaking of the ground. The bulk of this report deals with the evaluation of the ground-shaking hazard and with the ejects of ground shaking on buildings and other facilities. However, other aspects of the seismic hazard, referred to as collateral hazards, often are important. They include fault ruptures, landslides, liquefaction, tsunamis, and seiches. Landslides may occur in the absence of shaking, but earthquakes often trigger the sliding of susceptible slop es. Liquefaction ~ the state whereby a normally solid soil (saturated with ground water and usually sands of low density or compaction) turns to a mud-like or fluid consistency when shaken. Tsunamis are seismic sea waves (sometimes popularly called tidal waves). Seiches are sloshing or oscillating waves in bodies of water, generated by earthquakes in reservoirs, lakes, and enclosed harbors. In some earthquakes col- lateral hazards may be more destructive than the ground-shaking hazard, but the technology for evaluating these hazards and their

OCR for page 6
13 ejects is not as well developed as that relating to the ground-shaking hazard. In a similar vein, most loss estimates focus on the more or less direct effects of the damage caused by an earthquake: fatalities and injuries, Toss of function, and the cost of repairing clamage. Various other negative effects are called indirect losses. Other types of in- direct but potentially import ant consequence of damage include fire ancI flooding from dam failure. Another type of indirect consequence is the econorn~c impact of Toss of function on the owners of commercial property, on the region immediately affected by the earthquake, and on a larger region economically linked to the affected area. Again, these losses may be as important as the more direct losses, but the techniques for evaluating them is much more complex and not as well advanced. The Ground-Shai~ng Hazard The basic building block for a description of the ground-shaking hazard is a map displaying the intensities of ground shaking over the study region for an individual earthquake. In general, the in- tensity will vary over the region, depending on the size and source characteristics of the event, its location, and local geologic materi- als and topographical conditions. Such a description ~ a scenario earthquake. Most loss estunate studies use one or several scenario earthquakes to define the shaking hazard. Loss estunates based on specific earth- quakes are relatively easy to understand and explain. In addition, use of specific earthquakes makes it possible to include diverse types of losses, some of which are best described partially by words rather than merely by numbers. The use of several such events allows a range of assumptions and hypotheses to be analyzed and then syn- thesized in terms of their effects on facilities, without reliance on a single, perhaps unlikely occurrence. A more comprehensive but difficult to interpret display of the hazard consists of calculating the seismic shaking by considering many possible different earthquakes. Such events can cover a wide range of magnitudes and locations, and each can be assigned a prob- ability of occurrence. This approach leads finally to probabilities of occurrence for earthquake losses (as described in Working Paper C). These results can be presented as loss-frequency curves, which give the annual frequency with which different levels of loss are expected

OCR for page 6
14 LL He a: C] 10 1 LL X O 10 2 - m a: m o a: CL 10-3 10-4 10-5 o FIGURE 1-3 Loss-frequency curve. Existing Hazard Without Mitigation - Hazard With Mitigation - - - - LOSS to occur (Figure 1-3~. Summing these frequencies for levels above a specific value gives, for the study region, the annual probability of exceedance of losses. Representing the hazard as a loss-frequency curve is ideally suited for study of the relative merits of various mitigative actions. That is, loss-frequency curves corresponding to different possible ac- tions (including no action) may be compared. The method works best when all the consequences of an earthquake can be expressed by a single number, such as dollar loss. When multiple losses of different types are involved, the use of multiple scenario earthquakes finds wider favor. Regardless of the number of earthquakes used to represent the seismic hazard, there is no single, uncontroversial measure of the damageability of ground motions. For one of the most commonly utilized measures of intensity Modified MercaDi Intensity (MMI)- there are even basic disagreements as to the interpretation and defi- nition of the scale. A strong need exists for communication at the beginning of a Toss study among those who will evaluate the ground-shaking hazard

OCR for page 6
15 and ground failures, those who will determine the losses resulting from that seismic hazard, and those who will utilize the results of the study. Vulnerability There are two steps in a vulnerability analysis: (1) developing an inventory of the buildings and other facilities to be considered in the study, and (2) establishing for each inventory category the relationships among intensity of ground shaking (and, in some cases, ground failures), resulting damage, and associated losses. A key step is to develop a list of the categories of facilities to be considered, that is, to select a classification system. Selection of this system requires a compromise between different objectives. On the one hand, a very detailed classification system, with many categories, allows fine distinction to be made among buildings with different seismic resistance. On the other hand, a coarse classification system with only a few categories simplifies the inventory effort and makes it more economical. It is also inappropriate for a classification scheme to divide facilities into many different categories if the un- derlying state of the art is unable to distinguish among the predicted performance of the categories. Reaching an optimum compromise requires close communication among the parties conducting the loss study. For purposes of evaluating damage, facilities are usually inven- toried by placing them in different groups. . Buildings that provide working space or residences for people; . Lifelines, such as transportation, communications, water, sewage, and electricity systems, that are vital to the functioning of an area; Essential facilities, such as hospitals, and fire and police sta- tions, that are vital to postdisaster response; and . Facilities with a potential for large lo - , such as large and densely occupied buildings, dams, and chemical plants. Lifelines must be treated differently than buildings because they form interconnected systems that extend over large areas. Essential facilities, if they are to be included, must receive more careful atten- tion and individual surveys and analyses. Facilities with a potential for large loss pose a very special problem. Clearly their presence and potential for large loss must be noted, but losses cannot actually be estimated without analyses of the likelihood that potential damage

OCR for page 6
16 will actually occur in a given scenario earthquake, and this requires very detailed study wed beyond the scope of a typical loss estimate. It is easier, for example, to map the area that would be flooded if a certain dam were to fail than it is to determine whether the dam actually would fail in various earthquakes. CONSIDERATIONS OF UNCERTAINTY The foregoing discussion has presumed that loss est~rnates take the form of scenarios or a loss frequency curve. For the former, one or more particular earthquakes are postulated to occur, and the losses expected from each are described. For the latter representation, the probabilities of various levels of loss are indicated. Whichever method is used, the uncertainty in the loss estimates should be indicated, such as by giving a range of possible losses. The uncertainties in loss estimates derive Tom several sources. First is uncertainty in the ground-motion intensity and ground fail- ures for a given event. Second is uncertainty in estimating damage given the intensity and ground failures. Third is uncertainty in es- timating the losses given damage to the facility. Finally there is uncertainty in the process of inventorying the number of facilities in each building classification and geographic area. Each of these elements could be made more precise with additional effort and re- sources, but uncertainties are inevitable in any practical study of earthquake losses and should be expressed and quantified.