Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
6 Damage and Losses to Special Facilities and Urban Systems \ The methods used to estimate losses for the general population of buildings must be modified for application to lifelines, facilities with essential emergency functions, or facilities with a potential for very large losses. [IEE[INES Lifelines (or utilities and infrastructure systems) include railroad airport, motor vehicle, water, telephone, electricity, natural gas or oil pipelines, sewage, port and airport, and communications services. The words systems and services are central to the distinction between the loss estimation process for lifelines as compared to buildings. Service outages are almost always a prominent concern addressed by lifeline loss studies. Property losses are also important, but casualties associated with damage to lifelines usually are small. A lifeline, such as a water or electrical utility's facilities and functions, must be analyzed as a system rather than as separate, unrelated structures. Securing the active cooperation and support of local lifeline owners, operators, and regulators is the key to producing a satisfactory loss estimate. An understanding of how the system operates is essential. 53
54 The first step in the analysis of lifelines is to estimate the prom ability that components of the system will fail. Examples of com- ponents would be bridges in highway routes, switchyards and trans- former stations in power systems, and pumping stations in water and sewage systems. The estimation of losses to the individual compo- nents of a system has a less extensive historical loss experience to support the development of construction class motion-damage rela- tionships than with buildings. The most ambitious attempt to date to develop classes that include nonbuilding structures is ATC-13, in which some 30 classes are related to lifelines. The panel believes that the DPMs in ATC-13 provide the best available guidance, especially for bridges, although adjustments for local conditions will generally be necessary. ATC-13 DPMs should be used with a knowledge of the specific definitions of the classes. For example, a DPM that was devised for the case of seismically anchored electrical equipment should not be applied to a case where the equipment is unanchored. Buried pipelines are more vulnerable to collateral hazards such as fault ruptures, landslides, and liquefaction than they are to ground shaking. The probability of failure of such a pipeline under these collateral hazards will depend on the detailed characteristics of the ground movement and the material, age, depth of burial, and wall thickness and diameter of the pipeline. There are examples of suc- cessfuT pipeline performance as well as failures for each of these collateral hazards. For any specific pipeline and detailed character- istic of ground movement, an evaluation of pipeline performance can be made. However, such detailed evaluations are beyond the feasible scope for a large geographic loss estimation study that includes many such pipelines. For these studies, the probability of failure of buried pipelines should be treated as being rather high (greater than 50 percent) wherever collateral hazards are postulated to occur. Similarly, with the exception of bridges, which are potentially vulnerable to both effects, highway and railway networks are also more vulnerable to collateral hazards than they are to ground shak- ing. The probability of failure of links in such networks due to col- lateral hazards should be treated similarly to that described above for buried pipelines. Once the probability of individual components failing has been estimated, the next step is then to evaluate the influence of the failure of components or segments on the performance of the system, as a whole. If analytical models exist for the system (utilities will often
55 have such models), the effect on overall performance of the loss of some components can be estimated readily. Lacking any available system-wide models, expert opinion based on available data must be user} to estimate the outages that might be expected. In either case, the result is a scenario describing the state of the lifeline; that is, its ability to provide service following the earthquake. ~ addition to degradation in system-wide performance, there may be localized outages. For example, pipes in local distribution systems may fad] as a result of ground shaking or ground failures, and streets may become clogged by debris. These local failures contribute to the overall problem of restoring service. The time needed to restore service is an important factor in planning for disaster response. This is, in part, a matter of the time required to bring components back to a serviceable condition (e.g., to fix breaks in pipelines or to inspect bridges). The AT~13 report contains t~rne-to-restore-service matrices for a number of life- I~ne components. Restoring service also depends on the emergency response capability of the lifeline operator or of other emergency response agencies. A utility with an earthquake-resistant radio sys- tem, personnel who undergo annual earthquake exercises to test their ability to carry out preassigned tasks, and back-up plans for using emergency bypasses should be much more able to contain the impact of earthquake damage than another utility that does not have these capabilities ant] experiences the same darnage. Considerations such as these must be handled on a case-by-case basis after evaluation of the utilities' emergency preparedness. Loss estimation studies have seldom incorporated lifelines into the study to the same extent as building losses. Lifeline loss esti- mation methods are not as mature as for building loss estimation, and the problems are very complex. There has been considerable re- search into methods for evaluating the performance of interconnected systems in probabilistic terms, but ~ yet these sophisticated meth- ods have not been used in conventional, multipurpose, regional-scale loss studies. The pane! encourages more systematic and sophisti- cated studies of losses to lifeline systems, partially for the purpose of aiding in the maturing of lifeline loss estimation. However, ad- ditional damage statistics will accumulate only slowly, because so many factors affect the behavior of components in lifeline systems.
56 FACILITIES WITH ESSENTIAL EMERGENCY FUNCTIONS In general-purpose loss estimates, the main focus of this study, special attention must be given to those facilities most essential to emergency response, such as fire stations, emergency operations centers, and hospitals. The key question is: How well will these facilities be able to perform after an earthquake? Obviously, if such facilities suffer severe structural damage, their usefulness wiB be negated or greatly impaired. However, even if there is little or no structural damage, the facility may be unable to func- tion effectively if nonstructural damage causes critical equipment to be dislodged or overturned, or if essential or dangerous chemi- cals have been thrown down from shelves, or if lifelines services are interrupted. Nonstructural damage is significant because it is generally more widespread than structural damage. Even a moderate level of ground shaking (such as Vl or VIT on the MM! scale) can cause nonstruc- tural damage, such as overturned gas cylinders or water heaters and the release of flammable or toxic gas. The inventory task of field surveying nonstructural characteristics for the building population at large has yet to be attempted in a large-scale study, but this effort should be undertaken for the smaller number of essential, emergency function facilities that are within the scope of a large-scale study. During a loss study, it generally is necessary to walk through each essential facility allowing sufficient time to assess the likelihood of severe structural damage, but it also is essential to ascertain whether critical equipment and supplies have been adequately secured, en cl whether back-up resources have been arranged to deal with util- ity outages. This is labor intensive work and requires earthquake engineering expertise, but these are unavoidable costs. Undertaking a detailed structural analysis of such facilities is generally beyond the scope of a loss study. However, it may be desirable to examine structural drawings and to utilize a rapid as- sessment method. Critical evaluation of these methods is beyond the scope of this report, but several such methods are described briefly in Working Paper E. Even though each emergency facility is inventoried, the problem of potential liability to those involved with this work may make it desirable for a number of such facilities to be grouped when stating expected Tosses. That is to say, the result is a scenario describing the functionality of the emergency response system as a whole and not the functionality of individual facilities.
57 ~ _ :: : i: : : _: : :: ::~: The flexible, first-story of this hospital building was overstressed during the 1971 San Fernando earthquake (M 6.6~. Although designed in accordance with the 1970 building code, this reinforced-concrete structure was so severely damaged that it was torn down. The intensity of ground shaking was much greater than the hypothetical intensity upon which the code requirements were based. Photo courtesy of P. C. Jcnr~inge. Aerial view of collapsed hospital buildings in Sylmar, California. The older, weaker buildings collapsed and the newer, stronger buildings survived with only minor damage during the 1971 San Fernando earthquake (M 6.6~. Photo courtesy of G. Howncr.
58 The ATC-13 report solicited from experts the DPM for mean percentage property Toss (termed damage factor by ATC-13 and gen- erally called damage ratio by others) versus the MMI for six classes of equipment (i.e., residential, office, electrical, mechanical, high technology and laboratory, and vehicles). The validity of these rela- tively gross groupings is unknown and untested at present, although considerable variability is known within these types of equipment. For example, in the mechanical category, many pumps are routinely bolted to concrete slabs and are relatively earthquake resistant, even where earthquakes are not specifically considered in design. Also within this overall category of mechanical equipment is air-handling equipment mounted on springs, and these items are usually quite vulnerable to earthquakes except where special seismic measures are taken. The six classes of equipment analyzed by ATC-13 are also not all-inclusive. The equipment DPMs were not directly used in the ATCi13 functional loss estunation process. Instead, experts were asked to assign recovery times to the damage states of Table ~2 (e.g., Toss damage state 1 = no damage) for 60 socioeconomic classes of build- ing use. For the class defined as health care services, for example, each expert had to decide how a given damage state (that now in- cluded structural plus equipment damage in its definition) affected the facility's functionality in terms of time to restore service to 30 percent, 60 percent, and 100 percent. Loss of function resulting from lifeline service outages were included as a separate step. To be valid, these relationships (of MMI to equipment damage, and of combined structural and equipment damage state to functional loss) must be defined more specifically than to say they represent "typical" Cali- fornia practice. It should be clearly recognized that there is less certainty and less maturity in such techniques for estimating functional loss than for estimating property loss. For essential facilities, individual field visits rather than reliance on general relationships are recommended. Considerable potential exists for improving est~rnates of the per- formance of essential facilities through research into the earthquake performance of nonstructural items and identification of typical non- structural conditions in different parts of the country. While the state of the art of quantitative nonstructural loss estimation is not well developed, guidance for identifying and reducing nonstructural vuInerabilities is available in works such as those by McGavin (1981, 1986), Reitherman (1985, 1986), Stratta (1987), and the Veterans
so Administration (1976, 1981~. For emergency planning purposes, references documenting functional lomes (e.g., Arnold and Durkin [1983] concerning hospitale) or other reports in the postearthquake reconnaissance and research literature are useful for pouting the way toward improving emergency response capabilities, even though quantitative response dysfunction can only be very approximately predicted. FACIIITI1:S WITH A POTENTIAL FOR LARGE LOSS In this category are large and densely occupied buildings and other facilities such as tank farmers, refineries, damp, liquefied natural gas (LNG) plants or storage areas, chemical plants, nuclear plants, and pipelines containing hazardous materials. The characteristic feature of these facilities ~ that failure could cause an enormous number of casualties as wed as very large property losses. Except possibly for large and densely occupied buildings, there may be only a few of these facilities in a given study area. Thus, the lo" estimator cannot take advantage of averaging out uncertainty in performance over many facilities, as can be done for the ordinary building stock. Unless the loss attributable to a facility can be stated with great confidence, including it may completely bias the projected overall lomes for the region. Since the low from an individual building or other facility cannot be estimated reliably except through very detailed and expensive analysis, it follows that possible losses from such facilities should not be quantitatively included in the overall estimated loss. Obviously, however, the existence of such potentially hazardous facilities cannot be ignored. They should be highlighted In the inven- tory, and the cognizant regulatory bodies should be urged to require that detailed studies be made. A large-scale multipurpose study can educate local officials and staff personnel about the potential threat and the need to map the location of such facilities.
60 Damage to a liquid storage tank during the 1971 San Fernando earthquake (M 6.6~. The sloshing of the fluid contents overstreseed the wall of the tank. Sometimes such overstressed tanks collapse "d combustible contents ignite. Photo comets of P. C. Jcnrunge.
