Click for next page ( 86


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 85
Working Paper A Types and Examples of Loss Estimation Studies Potential users of loss estunates have different objectives, and a loss estimate study can only be called successful when it meets the purposes for which it is intended. Loss estunate studies can be categorized according to: type of losses estimated, kinds of facilities encompassed, certainty and detail, time span, and geographic scope. These considerations can be combined in a variety of ways in a particular study, and it would be impossible to discuss all of them. The categorization scheme depicted in Figure A-1 is only one way of structuring this subject matter. Other ways of categorizing and ana- lyzing earthquake loss estimation methods may be found in reviews of the field conducted from the 1930s to the present by Freeman (1932), McClure (1973), Boissonade and Shah (1982), Steinbrugge (1982, 1986), Reitherman (1985), Scawthorn (1986), and Whitman (1986~. Table A-1 divides earthquake loss estimation methods into five basic types, which can be characterized in terms of the combination of aspects presented in Figure A-1. Type I: General Type IT: Hazard Reduction Type IlI: Emergency Planning Type IV: Financial Risk ~ Type V: Economic Impact The methods presented in Figure A-1 all have a low degree of 85

OCR for page 85
86 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifeli nes Large potential for loss All buildings or structures CERTAINTY DETAI L high low high low TIME SPAN Hypothesized (scenario) Cumulative set of Predicted Actual earthquake earthquakes to earthquake earthquake occur in time span , ~ , G EGG RAPH IC SCALE Local Regional/state National FIGURE A-1 Aspects of earthquake loss estimation studies. certainty, which reflects the inherent uncertainty in the field of earth- quake loss estimation and is not necessarily indicative of method- ological errors or weaknesses in any particular method. In many engineering applications, the term accurate connotes a method that can reliably produce estimates that do not deviate much, say no more than perhaps 10 percent, from the actual results. Earthquake loss estimation methods that ace reliably of such accuracy (even in the case of facility-specific studies with high levels of effort) do not exist. Earthquake loss estimates that might prove to be in error by a factor of 3* are often considered accurate in this field. The word certainty is used here to describe the degree of confidence in a Toss estimate; an estimate with low certainty will have a large range of uncertainty *See footnote 2 in Chapter 4.

OCR for page 85
87 TABLE A-1 Purposes, Users, and Examples of Types of Lose Estimation Studies Type of Study Purpose Users Examples I: General Identify the general scope General public J. H. Wiggins of the earthquake problem as well as Company and to establish a basis for all other Engineering planning, prioritizing, users listed Geologists, and funding earthquake below Inc.1979; risk reduction efforts Algermissen et al., 1972 II: Hazard Guide hazard reduction Legislative, Alfors et al., reduction actions to reduce regulatory 1973; Ward, physical damage bodies; 1986; Office government of State officials Architect, and staffs; 1982; Los utilities and Angeles City corporations Planning Department, 1980 III: Emergency Facilitate more efficient Emergency Algerminsen planning emergency response response et al., 1972; agencies; D avis et al., utilities 1982a,b and corporations IV: Financial Rate earthquake risks of Insurance, Freeman, 1932; individual properties or mortgage California collective risk of lending, and Department of portfolios investment Insurance, industries 1985; Working Group Earth- quake Hazard Reduction, 1978 V: Economic Estimate economic losses National Applied impact (including indirect, security Technology long-term economic agencies and Council, 1985 impacts) national or regional planners about the best estimate, and conversely one with high certainty will have a small range of uncertainty. Specific examples of loss studies follow. TYPE I: GENERAL An example of this broadest type of loss study is the research

OCR for page 85
88 funded by the National Science Foundation (NSF) that produced property loss estimates for earthquakes as well as floods, expansive soil, landslides, hurricanes, and tornadoes Hi. H. Wiggins Company and Engineering Geology Consultants, Inc., 1979) for cumulative property losses during the years 1970 to 2000. Life loss was also estimated to some degree as well as the reduction in losses that could be expected by application of hazard reduction actions. The scope of such a study is very broad both geographically and in terms of considering more than one hazard. Although the results are more aggregated and less certain than those from studies focus- ing on an individual region and only one hazard, such comprehensive estimates are needed. Comparisons among hazards, between the con- tinuation of policies or the initiation of certain preventive actions and between the losses that would likely occur In the near term versus the Tong term, can be useful decision-making tools, especially at the national policymaking level. This type of study also enables compar- isons between the relative degree of risk faced by different states in relation to fixed analytical benchmarks, in contrast to comparisons of Tosses resulting from scenario events that vary in likelihood from one study to another. General studies are necessary if the intended application, such as selecting among policy options and evaluating the effectiveness of loss reduction programs, requires statements that say, for exam- ple: "Unless significant new steps are taken, the costs of replacing or repairing buildings destroyed and damaged by the nine natural hazards studied, during a typical year, are likely to increase more than 85 percent in the Midyear period between 1970 and 2000" Hi. H. Wiggins Company and Engineering Geologists, Inc., 1979~. Figure A-2 illustrates the characteristics combined in Type national-scale Toss estunation study using the above-mentioned study as the example. The types of losses estunated by such a study may vary, but two basic components are direct monetary cost of damage and casualties. The time variable is defined In terms of the cumulative losses estimated to occur in a given time span, in this case 197(}2000. The scope in terms of the kinds of facilities extends to all buildings, and both certainty and detail are relatively low. This study could also qualify as a Type T! (hazard reduction) study, which illustrates the overlap between categories. Regional Type ~ studies have fulfilled a variety of purposes, perhaps their most frequent use being as an emergency planning resource. The first of the studies (AIgermissen et al., 1972) sponsored

OCR for page 85
89 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifelines Large potential for loss All buildings or structures CERTAINTY DETAIL high low high low TIME SPAN Hypothesized (scenario) earthquake Cumulative set of Predicted Actual earthquakes to earthquake earthquake occur in time span l GEOGRAPHIC SCALE Local Regional/state National FIGURE A-2 Aspects of a Type I, national-scale loss estimation study, using the example of J. H. Wiggins and Engineering Geology Consultants, Inc. (1979~. Key: = aspects that pertain to this type of study. by the National Oceanic and Atmospheric Administration (NOAA) is a typical example of a Type ~ regional-scale study. This study of the San Francisco area projects a broad range of losses. Later NOAA and U.S. Geological Survey (USGS) studies are quite similar in their broad scope. (~n the mid-1970s, earthquake loss estimation projects and staff were shifted from NOAA to USGS with no significant change in the type of studies undertaken nor the methods used.) Casualties were estimated by time of day, by county, and according to hazard sources, that is, casualties that would occur within hospitals, schools, or dwellings are differentiated from other injuries and fatalities. Outages of utility services, transportation routes, and other types of functional losses suffered by lifelines were estimated. Property lodes involved only single-family dwellings. An updating loss study (Steinbrugge et al., 1981) also estimated property

OCR for page 85
go losses for commercial and most other building types. Figure A-3 shows the table of contents from the first NOAA study in order to indicate its scope, which is sunilar to later NOAA-USGS studies. Type ~ studies are often the first type of study to be conducted in a region. They are essential tools of seismic safety advocacy. As public policy, earthquake hazard reduction, or emergency planning activities develop, other studies with narrower foci may be conducted to support more specialized risk reduction efforts. Type ~ study elements are often adapted for use in other kinds of studies, and in some cases a Type ~ study can serve some of the more specific purposes requiring Type Il. Ill, IV, or V studies (Figure Add. The regional-scale study ~ much finer in detail than a national study, but at the cost of a smaller geographic scope. This trade-off between covering a larger area at a shallower level versus a smaller area in-depth is an inescapable constraint on all earthquake loss estimation studies. Figure A-4 categorizes a Type ~ regional study as including all but the overall econorn~c Trip act and national security types of losses; it may include the entire range of kinds of facilities. Its Toss statements have usually been predicated upon scenario events, and the certainty and especially its detail are greater than in the case of national-scale studies. In most cases, it is unportant that the study area boundaries or subarea boundaries match political boundaries demarcating cities or counties. TYPE II: HAZARD REDUCTION Type IT studies primarily support hazard reduction efforts, and the primary user is government agencies which adopt building codes regulating new construction or retroactive ordinances pertaining to existing hazardous facilities, land-use plans, and other laws and poli- cies. Type ~ studies are often used for this purpose, but Type IT studies emphasize this hazard reduction purpose with more specific reference to the codes, ordinances, voluntary standards, or other concrete policy options under consideration, and limit their scope to the specific physical hazards, resources, or jurisdictions of interest. On a state scale, cumulative losses over future time spans have been estimated in studies of California (Alfors et al., 1973) and Utah (Ward, 1986~. These two studies fit the pattern shown in Figure A-5, with the scope of the California study extending to all buildings and the Utah study focusing on particular types of facilities, such as schools and hospitals. By using a multidecade time

OCR for page 85
91 Table of Contents PART A: ISOSEISMAL STUDIES. PART B: CASUALTIES AND DA\lAG E Section 1: Introduction Section 2: Bases for Analysi s Section 3: Effects on Local Medical Resources Major Hospitals Health Manpower. Medical Supplies . Bloodbanks Hospital Reserve Disaster Inventory (HRDI) Modules . Packaged Disaster Hospitals . Clinical Laboratories . . . . . . . . Ambulance Services . Nursing Homes Section 4: Demands on Medical Resources . Deaths and Inj uric s, Ex eluding Dam s Dams . . ~ Section 5: Effects on Immediate and Vital Public Needs Public Structures . Communications Transportation. Public Utilities. . Schools . . . . . . . . . c ~ . Mercantile, Industrial ~ and Warehousing . Homeless . ~ . . Fire Following Earthquake Selected Bibliography . . . 1 8 34 34 55 61 68 76 80 87 93 102 108 108 126 ~ 133 133 146 153 172 i88 194 200 208 215 FIGURE A-3 Table of contents of the first of the joint National Oceanic and Atmospheric Administration and U.S. Geological Survey studies. Source: Algermissen et al. (1972~.

OCR for page 85
92 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security r '~ KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifeli nes Large potential for loss All buildings or structures CERTAINTY DETAIL ~ high low high low TIME SPAN Hypothesized (scenario) Cumulative set of Predicted Actual earthquake earthquakes to earthquake earthquake occur in time span . . GEOGRAPHIC SCALE Local Regional/state National FIGURE A-4 Aspects of a Type I, regional-scale loss estimation study, using the example of Algermmsen et al. (1972~. Key: = aspects that pertain to this type of study. span and estimating cumulative losses, these studies provide a way for policymakers to develop long-term risk reduction strategies. The study of 229 hospitals having 1,077 buildings in six southern California counties (Office of State Architect, 1982) is a Type IT study of a large urban region within one state, with the scope limited to one kind of occupancy. Vuinerabilities were rated without regard to scenario or cumulative losses. On a broader geographic scale, limited also to one kind of facility, a survey of 800 major buildings on University of California campuses was conducted (McClure, 1984~. Losses in this case were estimated in terms of the relative risks faced by building occupants, assuming the buildings were subjected to the same strong level of shaking. These last two examples of studies indicate that for some hazard reduction purposes, relative

OCR for page 85
93 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifeli nes Large potential for loss All buildings or structures _ CERTAINTY DETAI L ~ high low high low TIME SPAN Hypothesized (scenario) earthquake Cumulative set of Predicted Actual earthquakes to earthquake earthquake occur in time span GEOGRAPHIC SCALE Local Regional/state National E`IGURE A-5 Aspects of a Type II, state-scale loss estimation study, using the examples of Alfors et al. (1973) and Ward (1986~. Key: = aspects that pertain to this type of study. risk ratings rather than estimated numbers of casualties In a given scenario earthquake may be the appropriate goal of the analysis. An example of a local-scale hazard reduction study is the en- vironmental impact report accompanying an ordinance that went into effect in Los Angeles in 1981 requiring the hazards of about 8,000 unreinforced masonry buildings to be reduced (Los Angeles City Planning Department, 19803. In this study, only one kind of structure was studied, only life losses were of concern, the time span was in terms of a future scenario earthquake, and the certainty and detail were higher than with typical Type ~ studies (Figure A-6. This Toss estimate study was calibrated with the earlier NOAA study of losses estimated for a broader area and without an explicit breakdown of casualties related to classes of construction (AIgermis- sen et al., 1973~. The 1980 L08 Angeles study provided the conclusion

OCR for page 85
94 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security Kl N DS OF FACI LITI ES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifeli nes Large potential for loss All buildings or structures CERTAINTY DETAIL ~ high low high low TIME SPAN Hypothesized (scenario) Cumulative set of Predicted Actual earthquake earthquakes to earthquake earthquake occur in time span _ . GEOGRAPHIC SCALE I Local Regional/state National FIGURE A-6 Aspects of a Type II, local-scale lose estimation study, using the example of Los Angeles City Planning Department (1980~. Key: ~ = aspects that pertain to this type of study. that in a great earthquake (the same scenario earthquake used in the 1973 NOAA study), the number of fatalities within the city would de- cTine from 8,500 to 1,500 if the retroactive standards for unreinforced brick buildings were implemented. TYPE m EMERGENCY PLANNING The NOAA-USGS study (Type I) also fits this category, but another example is the work done by the California Division of Mines and Geology to identify functional losses to lifelines in the urban areas of Los Angeles (California Division of Mines and Geology, in progress; Davis et al., 1982a) San Francisco (Davis et al., 1982b; Steinbrugge et al., in progress), and San Diego (ReichIe et al., in progress). The characteristics of this type of study, at the regional scale, are shown

OCR for page 85
95 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifelines Large potential for loss All buildings or structures CERTAINTY DETAIL high low high low Tl M E SPAN Hypothesized (scenario) Cumulative set of Predicted Actual earthquake earthquakes to earthquake earthquake occur in time span rim ~ GEOGRAPHIC SCALE Local Regional/state National FIGURE A-7 Aspects of a Type III, regional-scale loss estimation study, using the examples of Davis et al. (1982 a,b). Key: ~ = aspects that pertain to this type of study. in Figure A-7. When such a study is devoted to a smaller geographic area, the detail of the results increases, as shown by the finer scale of the maps used to portray the results. The fire department of Orange County, California has extended the detail of one type of emergency planning study concerning transportation routes to the level of the neighborhood surrounding each fire station, looking at each roadway route that leads from the station to the outside area and considering potential route blockages such as collapsing bridges or building debris (C. Nicola, Orange County, California, Fire Department, personal communication, 1986~. This might be called the "street map" scale of Type ITI studies.

OCR for page 85
96 TYPE IV: FINANCE RISK This type of study is distinguished by its focus on direct prop- erty loss and the fact that its prunary user for many decades has been the insurance industry, and to a lesser extent the mortgage lending and investment industries. Several examples are provided in an earlier, widely published work in the field of earthquake loss es- timation, Earthquake Damage and Earthquake Insurance (Freeman, 1932~. Freeman produced regional-scale property loss estimates for all areas of the United States; the nature of this study is diagrammed in Figure A-8. Another study that fits this pattern of a state- or regional-scale financial study is the annually updated report issued by the Department of Insurance in California, which estimates aggre- gate losses in each of the various regional-scale zones of the state for properties covered by earthquake insurance (California Department of Insurance, 1985~. In these two cases, the rating of the risk of experiencing prop- erty damage or insurance losses extends essentially to Al buildings, and the intended user is broadly defined as the property insurance industry or government regulators having industry-wide insurance concerns. Some Type IV studies conducted for a given company, however, limit themmeIves to a smaller scope those facilities that are contained in that particular company's portfolio of insured, fi- nanced, or owned properties. An appendix to a comprehensive report by the Office of Science and Technology Policy (OSTP) discussed one particular aspect of the subject matter that falls under the heading of Type IV studies: the risk faced by various sectors of the financial industry during certain possible earthquake prediction situations (Working Group on Earthquake Hazards Reduction, 1978~. This analysis pointed out the need to divide a financial sector, such as mortgage lending, into smaller categories when analyzing earthquake risk, because of the different characteristics in terms of assets, liabilities, income, and expenses of institutions such as commercial banks, savings and loans establishments, and life insurance companies. This type of study is an exception to the rule that financial risk studies generally focus on the monetary cost of damage as the type of loss of concern. TYPE V: ECONOMIC IMPACT Type V studies deal with the decrease in the economy's pro- duction of goods and services that might result months after the

OCR for page 85
97 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifelines Large potential for loss All buildings or structures _ , CERTAINTY DETAI L high low high low T TIME SPAN Hypothesized (scenario) ~ Cumulative set of Predicted Actual earthquake earthquakes to earthquake earthquake occur in time span GEOGRAPHIC SCALE Local Regional/state National FIGURE A-8 Aspects of a Type IV, regional-scale loss estimation study, using the example of Freeman tl932~. Key: ~ = aspects that pertain to this type of study. earthquake, rather than just the unmediate damage. The most re- cent effort of this type has been initiated by the Federal Emergency Management Agency (FEMA). The ATC-13 study is the engineering component of the overall FEMA method that will use the ATC-13 estimates of initial damage and decrease in functionality to forecast the effects on local, regional, and national economies and national security. Therefore, the ATC-13 method may also be thought of as a Type I, general-purpose loss estunation technique. The primary motivation for the ATC-13 study was concern over the ability of de- fense industries to supply militarily essential products after a major C)aTifornia earthquake, and it was requested by the National Security Council (NSC). Figure A-9 illustrates the basic characteristics of this proposed type of study (which has to date been implemented only In pilot projects). The ATC-13 or earthquake engineering portion of

OCR for page 85
98 TYPE OF LOSS Monetary cost of damage Casualties Homeless Functionality of essential facilities Safety problems of potentially high hazard facilities Economic impact National security KINDS OF FACILITIES Selected facilities (e.g., occupancy, ownership, construction) Essential facilities Lifeli nes Large potential for loss All buildings or structures CERTAINTY DETAI L high low high low TIME SPAN Hypothesized (scenario) Cumulative set of Predicted Actual earthquake earthquakes to earthquake earthquake occur in time span GEOGRAPHIC SCALE l Local Regional/state National FIGURE A-9 Aspects of a Type V, regional-scale loss estimation study, using the example of FEMA Fedloss Method and its ATC-13 engineering component (Applied Technology Council, 1985~. Key: = aspects that pertain to this type of study. this overall economic loss modeling method could also be applied to other purposes and diagrammed differently, but is outlined here in the context of its original purpose. ATC-13 as of this date has not been used to produce a complete, published loss study. Much detail exists with the FEMA/ATC-13 type of study be- cause its aim was to clevelop an inventory of almost every facility according to about 500 econorn~c sectors for most of the state of California. It makes precise statements about the amount of damage and the functional loss that could be suffered in each facility. It includes industrial tanks, tunnels, and other nonbuilding structures in greater detail than other methods. The AT~13 method devel- ops losses on the scale of each individual factory, for exan~ple, first estimating damage and then determining the number of days after

OCR for page 85
99 the earthquake before 30 percent, 60 percent, and 100 percent of pre-earthquake functioning is restored. Because hard data or relatively accurate field-acquired informa- tion describing the construction characteristics of all buildings in a region do not exist, and because of few data on the connection between building damage and loss of function, the ATC-13 method provides inferences for constructing an inventory from readily avail- able socioeconomic data bases. Expert opinion was used to develop damage and loss relationships for a large number of types of facilities. The ATC-13 example is a reminder that when great detail is sought on a large scale requiring that the loss estimation method answer a number of difficult questions in a detailed, quantitative waycertainty must be sacrificed. Widely accepted, easily applied, and objective ways of rating the certainty of loss estimate methods do not exist, and the issue of what constitutes acceptable certainty or acceptable detail can be decided only by reference to the fitness of the study for its intended purpose. This again brings up the important subject of the users and uses of loss estimation studies, a theme throughout this report.