EM

153

CHAPTER 6

HAZARD, VULNERABILITY, AND RISK ANALYSIS

This chapter describes how preimpact conditions act together with event-specific conditions to produce a disaster’s physical and social impacts. These disaster impacts can be reduced by emergency management interventions. In addition, this chapter discusses how emergency managers can assess the preimpact conditions that produce disaster vulnerability within their communities. The chapter concludes with a discussion of vulnerability dynamics and methods for disseminating hazard/vulnerability data.

Introduction

A disaster occurs when an extreme event exceeds a community’s ability to cope with that event.

Understanding the process by which natural disasters produce community impacts is important for four

reasons. First, information from this process is needed to identify the preimpact conditions that make

communities vulnerable to disaster impacts. Second, information about the disaster impact process can be

used to identify specific segments of each community that will be affected disproportionately (e.g., low

income households, ethnic minorities, or specific types of businesses). Third, information about the

disaster impact process can be used to identify the event-specific conditions that determine the level of

disaster impact. Fourth, an understanding of disaster impact process allows planners to identify suitable

emergency management interventions. The process by which disasters produce community impacts can

be explained in terms of models proposed by Cutter (1996) and Lindell and Prater (2003). Specifically,

Figure 6-1 indicates the effects of a disaster are determined by three preimpact conditions—hazard

exposure, physical vulnerability, and social vulnerability. There also are three event-specific conditions,

hazard event characteristics, improvised disaster responses, and improvised disaster recovery. Two of the

event-specific conditions, hazard event characteristics and improvised disaster responses, combine with

the preimpact conditions to produce a disaster’s physical impacts. The physical impacts, in turn, combine

with improvised disaster recovery to produce the disaster’s social impacts. Communities can engage in

three types of emergency management interventions to ameliorate disaster impacts. Physical impacts can

be reduced by hazard mitigation practices and emergency preparedness practices, whereas social impacts

can be reduced by recovery preparedness practices.

The following sections describe the components of the model in greater detail. Specifically, the

next section will describe the three preimpact conditions—hazard exposure, physical vulnerability, and

social vulnerability. This section will be followed by sections discussing hazard event characteristics and

improvised disaster responses. The fourth section will discuss disasters’ physical impacts, social impacts,

and improvised disaster recovery. The last section will discuss three types of strategic interventions,

hazard mitigation practices, emergency preparedness practices, and recovery preparedness practices.

Preimpact Conditions

Hazard Exposure

Hazard exposure arises from people’s occupancy of geographical areas where they could be

affected by specific types of events that threaten their lives or property. For natural hazards, this exposure

is caused by living in geographical areas as specific as floodplains that sometimes extend only a few feet

beyond the floodway or as broad as the Great Plains of the Midwest where tornadoes can strike anywhere

over an area of hundreds of thousands of square miles. For technological hazards, exposure can arise if

people move into areas where they could be exposed to explosions or hazardous materials releases. In

principle, hazard exposure can be measured by the probability of occurrence of a given event magnitude,

154

but these exceedance probabilities can be difficult to obtain for hazards about which the historical data are

insufficient to reliably estimate the probability of very unusual events. For example, many areas of the US

have meteorological and hydrological data that are limited to the past 100 years, so the estimation of

extreme floods requires extrapolation from a limited data series. Moreover, urbanization of the

watersheds causes the boundaries of the 100-year floodplains to change in ways that may be difficult for

local emergency managers to anticipate. Even more difficult to estimate are the probabilities of events,

such as chemical and nuclear reactor accidents, for which data are limited because each facility is

essentially unique. In such cases, techniques of probabilistic safety analysis are used to model these

systems, attach probabilities to the failure of system components, and synthesize probabilities of overall

system failure by mathematically combining the probabilities of individual component failure.

Figure 6-1. Disaster Impact Model.

Source: Adapted from Lindell and Prater (2003)

The greatest difficulties are encountered in attempting to estimate the probabilities of social

hazards such as terrorist attacks because the occurrence of these events is defined by social system

dynamics that cannot presently be modeled in the same way as physical systems. That is, the elements of

social systems are difficult to define and measure. Moreover, the interactions of the system elements have

multiple determinants and involve complex lag and feedback effects that are not well understood, let

alone precisely measured. Indeed, there are significant social and political constraints that limit the

collection of data on individuals and groups. These constraints further inhibit the ability of scientists to

make specific predictions of social system behavior.

Physical Vulnerability

Human vulnerability. Humans are vulnerable to environmental extremes of temperature, pressure,

and chemical exposures that can cause death, injury, and illness. For any hazard agent—water, wind,

ionizing radiation, toxic chemicals, infectious agents—there often is variability in the physiological

Emergency management interventions

Emergency preparedness

practices

Hazard mitigation practices

Recovery preparedness

practices

Pre-impact

conditions

Hazard exposure

Social vulnerability

Physical vulnerability

Improvised disaster recovery

Hazard event

characteristics

Improvised disaster

response

Event-specific conditions

Social impacts

Physical impacts

155

response of the affected population. That is, given the same level of exposure, some people will die,

others will be severely injured, still others slightly injured, and the rest will survive unscathed. Typically,

the most susceptible to any environmental stressor will be the very young, the very old, and those with

weakened immune systems.

Agricultural vulnerability. Like humans, agricultural plants and animals are also vulnerable to

environmental extremes of temperature, pressure, chemicals, radiation, and infectious agents. Like

humans, there are differences among individuals within each plant and animal population. However,

agricultural vulnerability is more complex than human vulnerability because there is a greater number of

species to be assessed, each of which has its own characteristic response to each environmental stressor.

Structural vulnerability. Structural vulnerability arises when buildings are constructed using

designs and materials that are incapable of resisting extreme stresses (e.g., high wind, hydraulic pressures

of water, seismic shaking) or that allow hazardous materials to infiltrate into the building. The

construction of most buildings is governed by building codes intended to protect the life safety of

building occupants from structural collapse—primarily from the dead load of the building material

themselves and the live load of the occupants and furnishings— but do not necessarily provide protection

from extreme wind, seismic, or hydraulic loads. Nor do they provide an impermeable barrier to the

infiltration of toxic air pollutants.

Social Vulnerability

The social vulnerability perspective (e.g., Cannon, Twigg & Rowell, 2003; Cutter, Boruff &

Shirley, 2003) represents an important extension of previous theories of hazard vulnerability (Burton, et

al., 1978). As a concept, social vulnerability has been defined in terms of people’s “capacity to anticipate,

cope with, resist and recover from the impacts of a natural hazard” (Wisner, Blakie, Canon & Davis,

2004, p. 11). Whereas people’s physical vulnerability refers to their susceptibility to biological changes

(i.e., impacts on anatomical structures and physiological functioning), their social vulnerability refers to

their susceptibility to behavioral changes. As will be discussed in greater detail below, these consist of

psychological, demographic, economic, and political impacts.

The central point of the social vulnerability perspective is that, just as people’s occupancy of

hazard prone areas and the physical vulnerability of the structures in which they live and work are not

randomly distributed, neither is social vulnerability randomly distributed—either geographically or

demographically. Thus, just as variations in structural vulnerability can increase or decrease the effect of

hazard exposure on physical impacts (property damage and casualties), so too can variations in social

vulnerability. Social vulnerability varies across communities and also across households within

communities. It is the variability in vulnerability that is likely to be of greatest concern to local emergency

managers because it requires that they identify the areas within their communities having population

segments with the highest levels of social vulnerability.

Event-Specific Conditions

Hazard Event Characteristics

Hazard impacts often are difficult to characterize because a given hazard agent may initiate a

number of different threats. For example, tropical cyclones (also known as hurricanes or typhoons) can

cause casualties and damage through wind, rain, storm surge, and inland flooding (Bryant, 1991).

Volcanoes can impact human settlements through ash fall, explosive eruptions, lava flows, mudflows and

floods, and forest fires (Perry & Lindell, 1990; Saarinen & Sell, 1985; Warrick, et al., 1981). However,

156

once these distinct threats have been distinguished from each other, each can be characterized in terms of

six significant characteristics. These are the speed of onset, availability of perceptual cues (such as wind,

rain, or ground movement), the intensity, scope, and duration of impact, and the probability of

occurrence. The speed of onset and availability of perceptual cues affect the amount of forewarning that

affected populations will have to complete emergency response actions (Gruntfest, et al., 1978; Lindell,

1994c). In turn, these attributes determine the extent of casualties among the population and the degree of

damage to structures in the affected area.

A hazard’s impact intensity can generally be defined in terms of the physical materials involved

and the energy these materials impart. The physical materials involved in disasters differ in terms of their

physical state—gas (or vapor), liquid, or solid (or particulate). In most cases, the hazard from a gas arises

from its temperature or pressure. Examples include hurricane or tornado wind (recall that the atmosphere

is a mixture of gases), which is hazardous because of overpressures that can inflict traumatic injuries

directly on people. High wind also is hazardous because it can destroy structures and accelerate debris

that can itself cause traumatic injuries. Alternatively, the hazard from a gas might arise from its toxicity,

as is the case in some volcanic eruptions. Liquids also can be hazardous because of their toxicity, but the

most common liquid hazard is water. It is hazardous to structures because of the pressure it can exert and

is hazardous to living things when it fills the lungs and prevents respiration. Lava is solid rock that has

been liquefied by extreme heat and therefore is hazardous to people and structures because of its thermal

energy. Solids also can be hazardous if they take the form of particulates such as airborne volcanic ash or

floodborne mud. These are particularly significant because they can leave deposits that have impacts of

long duration.

The scope of impact defines the number of affected social units (e.g., individuals, households,

and businesses). The probability of occurrence (per unit of time) is another important characteristic that

affects disaster impacts indirectly because more probable hazards are likely to mobilize communities to

engage in hazard mitigation and emergency preparedness measures to reduce their vulnerability (Prater &

Lindell, 2000).

Improvised Disaster Response

Disaster myths commonly portray disaster victims as dazed, panicked, or disorganized but, as will

be discussed at greater length in Chapter 8, people actually respond in a generally adaptive manner when

disasters strike. Adaptive response is often delayed because normalcy bias delays people’s realization that

an improbable event is, in fact, occurring to them. Further delays occur because people have limited

information about the situation and, therefore, seek confirmation of any initial indications of an

emergency before initiating protective action. In addition, the vast majority of people respond in terms of

their customary social units—especially their households and neighborhoods—which usually consumes

time in developing social organizations that can cope with the disaster’s demands. Contrary to stereotypes

of individual selfishness, disaster victims often devote considerable effort to protecting others’ persons

and property. Accordingly, there is considerable convergence on the disaster impact area, as those in areas

nearby move in to offer assistance. When existing organizations seem incapable of meeting the needs of

the emergency response, they expand to take on new members, extend to take on new tasks, or new

organizations emerge (Dynes, 1970).

Improvised Disaster Recovery

Once the situation has stabilized to the point that the imminent threat to life and property has

abated, disaster-stricken communities must begin the long process of disaster recovery. Immediate tasks

157

in this process include damage assessment, debris clearance, reconstruction of infrastructure (electric

power, fuel, water, wastewater, telecommunications, and transportation networks), and reconstruction of

buildings in the residential, commercial, and industrial sectors. Improvised disaster assistance is derived

primarily from resources provided by individuals and organizations within the community. The victims

themselves might have financial (e.g., savings and insurance) as well as tangible assets (e.g., property)

that are undamaged by hazard impact. As one might expect, low-income victims tend to have lower levels

of savings, but they also are more likely to be victims of insurance redlining and, thus, have been forced

into contracts with insurance companies that go bankrupt after the disaster. Thus, even those who plan

ahead for disaster recovery can find themselves without the financial resources they need (Peacock &

Girard, 1997). Alternatively, victims can promote their recovery by bringing in additional funds through

overtime employment or by freeing up the needed funds by reducing their consumption below preimpact

levels. Friends, relatives, neighbors, and coworkers can assist recovery through financial and in-kind

contributions, as can community based organizations (CBOs) and local government. In addition, the latter

also can provide assistance by means of tax deductions or deferrals.

Disaster Impacts

As noted earlier, disaster impacts comprise physical and social impact. The physical impacts of

disasters include casualties (deaths and injuries) and property damage, and both vary substantially across

hazard agents. The physical impacts of a disaster are usually the most obvious, easily measured, and first

reported by the news media. Social impacts, which include psychosocial, demographic, economic, and

political impacts, can develop over a long period of time and can be difficult to assess when they occur.

Despite the difficulty in measuring these social impacts, it is nonetheless important to monitor them, and

even to predict them if possible, because they can cause significant problems for the long-term

functioning of specific types of households and businesses in an affected community. A better

understanding of disasters’ social impacts can provide a basis for preimpact prediction and the

development of contingency plans to prevent adverse consequences from occurring.

Physical Impacts

Casualties. According to Noji (1997b), hurricanes produced 16 of the 65 greatest disasters of the

20 th

Century (in terms of deaths) and the greatest number of deaths from 1947-1980 (499,000).

Earthquakes produced 28 of the greatest disasters and 450,000 deaths, whereas floods produced four of

the greatest disasters and 194,000 deaths. Other significant natural disasters include volcanic eruptions

with nine of the greatest disasters and 9,000 deaths, landslides with four of the greatest disasters and

5,000 deaths, and tsunamis with three of the greatest disasters and 5,000 deaths. There is significant

variation by country, with developing countries in Asia, Africa, and South America accounting for the top

20 positions in terms of number of deaths from 1966-1990. Low-income countries suffer approximately

3,000 deaths per disaster, whereas the corresponding figure for high-income countries is approximately

500 deaths per disaster. Moreover, these disparities appear to be increasing because the average annual

death toll in developed countries declined by at least 75% between 1960 and 1990, but the same time

period saw increases of over 400% in developing countries (Berke, 1995).

There often are difficulties in determining how many of the deaths and injuries are “caused by” a

disaster. In some cases it is impossible to determine how many persons are missing and, if so, whether

this is due to death or unrecorded relocation. The size of the error in estimates of disaster death tolls can

be seen in the fact that for many of the most catastrophic events the number of deaths is rounded to the

158

nearest thousand and some even are rounded to the nearest ten thousand (Noji, 1997b). Estimates of

injuries are similarly problematic (see Langness, 1994; Peek-Asa, et al., 1998; Shoaf, et al., 1998,

regarding conflicting estimates of deaths and injuries attributable to the Northridge earthquake). Even

when bodies can be counted, there are problems because disaster impact may be only a contributing factor

to casualties with pre-existing health conditions. Moreover, some casualties are indirect consequences of

the hazard agent as, for example, with casualties caused by structural fires following earthquakes (e.g.,

burns) and destruction of infrastructure (e.g., illnesses from contaminated water supplies).

Damage. Losses of structures, animals, and crops also are important measures of physical

impacts, and these are rising exponentially in the United States (Mileti, 1999). However, the rate of

increase is even greater in developing countries such as India and Kenya (Berke, 1995). Such losses

usually result from physical damage or destruction of property, but they also can be caused by losses of

land use to chemical or radiological contamination or loss of the land itself to subsidence or erosion.

Damage to the built environment can be classified broadly as affecting residential, commercial, industrial,

infrastructure, or community services sectors. Moreover, damage within each of these sectors can be

divided into damage to structures and damage to contents. It usually is the case that damage to contents

results from collapsing structures (e.g., hurricane winds failing the building envelope and allowing rain to

destroy the furniture inside the building). Because collapsing buildings are a major cause of casualties as

well, this suggests that strengthening the structure will protect the contents and occupants. However,

some hazard agents can damage building contents without affecting the structure itself (e.g., earthquakes

striking seismically-resistant buildings whose contents are not securely fastened). Thus, risk area residents

may need to adopt additional hazard adjustments to protect contents and occupants even if they already

have structural protection.

Perhaps the most significant structural impact of a disaster on a stricken community is the

destruction of households’ dwellings. Such an event initiates what can be a very long process of disaster

recovery for some population segments. According to Quarantelli (1982a), people typically pass through

four stages of housing recovery following a disaster. The first stage is emergency shelter, which consists

of unplanned and spontaneously sought locations that are intended only to provide protection from the

elements, typically open yards and cars after earthquakes (Bolin & Stanford, 1991, 1998). The next step is

temporary shelter, which includes food preparation and sleeping facilities that usually are sought from

friends and relatives or are found in commercial lodging, although “mass care” facilities in school

gymnasiums or church auditoriums are acceptable as a last resort. The third step is temporary housing,

which allows victims to re-establish household routines in nonpreferred locations or structures. The last

step is permanent housing, which re-establishes household routines in preferred locations and structures.

Households vary in the progression and duration of each type of housing and the transition from

one stage to another can be delayed unpredictably, as when it took nine days for shelter occupancy to

peak after the Whittier Narrows earthquake (Bolin, 1993). Particularly significant are the problems faced

by lower income households, which tend to be headed disproportionately by females and racial/ethnic

minorities. Such households are more likely to experience destruction of their homes because of

preimpact locational vulnerability. This is especially true in developing countries such as Guatemala

(Peacock, Killian & Bates, 1987), but also has been reported in the US (Peacock & Girard, 1997). The

homes of these households also are more likely to be destroyed because the structures were built

according to older, less stringent building codes, used lower quality construction materials and methods,

and were less well maintained (Bolin & Bolton, 1986). Because lower income households have fewer

159

resources on which to draw for recovery, they also take longer to transition through the stages of housing,

sometimes remaining for extended periods of time in severely damaged homes (Girard & Peacock, 1997).

In other cases, they are forced to accept as permanent what originally was intended as temporary housing

(Peacock, et al., 1987). Consequently, there may still be low-income households in temporary sheltering

and temporary housing even after high-income households all have relocated to permanent housing

(Berke, et al., 1993; Rubin, Sapperstein & Barbee, 1985).

As is the case with estimates of casualties, estimates of losses to the built environment are prone

to error. Damage estimates are most accurate when trained damage assessors enter each building to assess

the percent of damage to each of the major structural systems (e.g., roof, walls, floors) and the percentage

reduction in market valuation due to the damage. Early approximate estimates are obtained by conducting

“windshield surveys” in which trained damage assessors drive through the impact area and estimate the

extent of damage that is visible from the street, or by conducting computer analyses using HAZUS

(National Institute of Building Sciences, 1998). These early approximate estimates are especially

important in major disasters because detailed assessments are not needed in the early stages of disaster

recovery and the time required to conduct them on a large number of damaged structures using a limited

number of qualified inspectors would unnecessarily delay the community recovery process.

Other important physical impacts include damage or contamination to cropland, rangeland, and

woodlands. Such impacts may be well understood for some hazard agents but not others. For example,

ashfall from the 1980 Mt. St. Helens eruption was initially expected to devastate crops and livestock in

downwind areas, but no significant losses materialized (Warrick, et al., 1981). There also is concern about

damage or contamination to the natural environment (wild lands) because these areas serve valuable

functions such as damping the extremes of river discharge and providing habitat for wildlife. In part,

concern arises from the potential for indirect consequences such as increased runoff and silting of

downstream river beds, but many people also are concerned about the natural environment simply

because they value it for its own sake.

Social Impacts

For many years, research on the social impacts of disasters consisted of an accumulation of case

studies, but two research teams conducted comprehensive statistical analyses of extensive databases to

assess the long-term effects of disasters on stricken communities (Friesma, et al., 1979; Wright, et al.,

1979). The more comprehensive Wright, et al. (1979) study used census data from the 1960 (preimpact)

and 1970 (post-impact) censuses to assess the effects of all recorded disasters in the United States. The

authors concurred with earlier findings by Friesma, et al. (1979) in concluding no long-term social impact

of disasters could be detected at the community level. In discussing their findings, the authors

acknowledged their results were dominated by the types of disasters occurring most frequently in the

United States—tornadoes, floods, and hurricanes. Moreover, most of the disasters they studied had a

relatively small scope of impact and thus caused only minimal disruption to their communities even in the

short term. Finally, they noted their findings did not preclude the possibility of significant long-term

impacts upon lower levels such as the neighborhood, business, and household.

Nonetheless, their findings called attention to the importance of the impact ratio—the amount of

damage divided by the amount of community resources—in understanding disaster impacts. They

hypothesized long-term social impacts tend to be minimal in the US because most hazard agents have a

relatively small scope of impact and tend to strike undeveloped areas more frequently than intensely

developed areas simply because there are more of the former than the latter. Thus, the numerator of the

160

impact ratio tends to be low and local resources are sufficient to prevent long-term effects from occurring.

Even when a disaster has a large scope of impact and strikes a large developed area (causing a large

impact ratio in the short term), state and federal agencies and NGOs (e.g., American Red Cross) direct

recovery resources to the affected area, thus preventing long-term impacts from occurring. For example,

Hurricane Andrew inflicted $26.5 billion in losses to the Miami area, but this was only 0.4% of the US

GDP (Charvériat, 2000). Recovery problems described in the studies reported in Peacock, Morrow and

Gladwin (1997) were determined more by organizational impediments than by the lack of resources.

Psychosocial impacts. Research reviews conducted over a period of 25 years have concluded that

disasters can cause a wide range of negative psychological responses (Bolin 1985; Gerrity & Flynn, 1997;

Houts, Cleary & Hu, 1988; Perry & Lindell, 1978). These include psychophysiological effects such as

fatigue, gastrointestinal upset, and tics, as well as cognitive signs such as confusion, impaired

concentration, and attention deficits. Psychological impacts include emotional signs such as anxiety,

depression, and grief. They also include behavioral effects such as sleep and appetite changes, ritualistic

behavior, and substance abuse. In most cases, the observed effects are mild and transitory—the result of

“normal people, responding normally, to a very abnormal situation” (Gerrity & Flynn 1997, p. 108). Few

disaster victims require psychiatric diagnosis and most benefit more from a crisis counseling orientation

than from a mental health treatment orientation, especially if their normal social support networks of

friends, relatives, neighbors, and coworkers remain largely intact. However, there are population

segments requiring special attention and active outreach. These include children, frail elderly, people with

pre-existing mental illness, racial and ethnic minorities, and families of those who have died in the

disaster. Emergency workers also need attention because they often work long hours without rest, have

witnessed horrific sights, and are members of organizations in which discussion of emotional issues may

be regarded as a sign of weakness (Rubin, 1991). However, as Chapter 11 will indicate, there is little

evidence of emergency workers needing directive therapies either.

The negative psychological impacts described above, which Lazarus and Folkman (1984) call

emotion focused coping, generally disrupt the social functioning of only a very small portion of the victim

population. Instead, the majority of disaster victims engage in adaptive problem focused coping activities

to save their own lives and those of their closest associates. Further, there is an increased incidence in

prosocial behaviors such as donating material aid and a decreased incidence of antisocial behaviors such

as crime (Drabek, 1986; Mileti, et al., 1975; Siegel, et al., 1999). In some cases, people even engage in

altruistic behaviors that risk their own lives to save the lives of others (Tierney, et al., 2001).

There also are psychological impacts with long-term adaptive consequences, such as changes in

risk perception (beliefs in the likelihood of the occurrence a disaster and its personal consequences for the

individual) and increased hazard intrusiveness (frequency of thought and discussion about a hazard). In

turn, these beliefs can affect risk area residents’ adoption of household hazard adjustments that reduce

their vulnerability to future disasters. However, these cognitive impacts of disaster experience do not

appear to be large in aggregate, resulting in modest effects on household hazard adjustment (see Lindell

& Perry, 2000 for a review of the literature on seismic hazard adjustment, and Lindell & Prater 2000;

Lindell & Whitney, 2000; and Whitney, Lindell & Nguyen, 2004 for more recent empirical research).

Demographic impacts. The demographic impact of a disaster can be assessed by adapting the

demographic balancing equation, Pa – Pb = B – D + IM – OM, where Pa is the population size after the

disaster, Pb is the population size before the disaster, B is the number of births, D is the number of deaths,

IM is the number of immigrants, and OM is the number of emigrants (Smith, Tayman & Swanson, 2001).

161

The magnitude of the disaster impact, Pa – Pb, is computed for the population of a specific geographical

area and two specific points in time. Ideally, the geographical area would correspond to the disaster

impact area, Pb would be immediately before disaster impact, and Pa would be immediately after disaster

impact. In practice, population data are available for census divisions (census block, block group, tract, or

larger area), so a Geographical Information System (GIS) must be used to estimate the impact on the

impact area. Moreover, population data are likely to be most readily available from the decennial

censuses, so the overall population change and its individual demographic components—births, deaths,

immigration, and emigration—are likely to be estimated from that source (e.g., Wright, et al., 1979). On

rare occasions, special surveys have been conducted in the aftermath of disaster (e.g., Peacock, Morrow &

Gladwin, 1997). The limited research available on demographic impacts (Friesma, et al., 1979; Wright, et

al., 1979) suggests disasters have negligible demographic impacts on American communities, but the

highly aggregated level of analysis in these studies does not preclude the possibility of significant impacts

at lower levels of aggregation (census tracts, block groups, or blocks). Although it is logically possible

that disasters could affect the number of births, it does not seem likely that the effect would be large.

Moreover, as noted in the previous section on physical impacts, the number of deaths from disasters in the

United States has been small relative to historical levels (e.g., the 6000 deaths in the 1900 Galveston

hurricane were approximately 17% of the city’s population) or to the levels reported in developing

countries. The major demographic impacts of disasters are likely to be the (temporary) immigration of

construction workers after major disasters and the emigration of population segments that have lost

housing. In many cases, the housing-related emigration is also temporary, but there are documented cases

in which housing reconstruction has been delayed indefinitely—leading to “ghost towns” (Comerio,

1998). Other potential causes of emigration are psychological impacts (belief that the likelihood of

disaster recurrence is unacceptably high), economic impacts (loss of jobs or community services), or

political impacts (increased neighborhood or community conflict).

Economic impacts. The property damage caused by disaster impact creates losses in asset values

that can be measured by the cost of repair or replacement (Committee on Assessing the Costs of Natural

Disasters, 1999). Disaster losses in United States are initially borne by the affected households,

businesses, and local government agencies whose property is damaged or destroyed. However, some of

these losses are redistributed during the disaster recovery process. There have been many attempts to

estimate the magnitude of direct losses from individual disasters and the annual average losses from

particular types of hazards (e.g., Mileti, 1999). Unfortunately, these losses are difficult to determine

precisely because there is no organization that tracks all of the relevant data and some data are not

recorded at all (Charvériat, 2000; Committee on Assessing the Costs of Natural Disasters, 1999). For

insured property, the insurers record the amount of the deductible and the reimbursed loss, but uninsured

losses are not recorded so they must be estimated—often with questionable accuracy.

The ultimate economic impact of a disaster depends upon the disposition of the damaged assets.

Some of these assets are not replaced, so their loss causes a reduction in consumption (and, thus, a

decrease in the quality of life) or a reduction in investment (and, thus, a decrease in economic

productivity). Other assets are replaced—either through in-kind donations (e.g., food and clothing) or

commercial purchases. In the latter case, the cost of replacement must come from some source of

recovery funding, which generally can be characterized as either intertemporal transfers (to the present

time from past savings or future loan payments) or interpersonal transfers (from one group to another at a

given time). Some of the specific mechanisms for financing recovery include obtaining tax deductions or

162

deferrals, unemployment benefits, loans (paying back the principal at low- or no-interest), grants

(requiring no return of principal), insurance payoffs, or additional employment. Other sources include

depleting cash financial assets (e.g., savings accounts), selling tangible assets, or migrating to an area with

available housing, employment, or less risk (in some cases this is done by the principal wage earner only).

In addition to direct economic losses, there are indirect losses that arise from the interdependence

of community subunits. Research on the economic impacts of disasters (Alesch, et al., 1993; Dacy &

Kunreuther, 1969; Dalhamer & D’Sousa, 1997; Durkin, 1984; Gordon, et al., 1995; Kroll, et al., 1991;

Lindell & Perry, 1998; Nigg, 1995; Tierney, 1997a) suggests the relationships among the social units

within a community can be described as a state of dynamic equilibrium involving a steady flow of

resources, especially money. Specifically, a household’s linkages with the community are defined by the

money it must pay for products, services, and infrastructure support. This money is obtained from the

wages that employers pay for the household’s labor. Similarly, the linkages that a business has with the

community are defined by the money it provides to its employees, suppliers, and infrastructure in

exchange for inputs such as labor, materials and services, and electric power, fuel, water/wastewater,

telecommunications, and transportation. Conversely, it provides products or services to customers in

exchange for the money it uses to pay for its inputs.

It also is important to recognize the financial impacts of recovery (in addition to the financial

impacts of emergency response) on local government. Costs must be incurred for tasks such as damage

assessment, emergency demolition, debris removal, infrastructure restoration, and re-planning stricken

areas. In addition to these costs, there are decreased revenues due to loss or deferral of sales taxes,

business taxes, property taxes, personal income taxes, and user fees.

Political impacts. There is substantial evidence that disaster impacts can cause social activism

resulting in political disruption, especially during the seemingly interminable period of disaster recovery.

The disaster recovery period is a source of many victim grievances and this creates many opportunities

for community conflict, both in the US (Bolin 1982, 1993) and abroad (Bates & Peacock 1988). Victims

usually attempt to recreate preimpact housing patterns, but it can be problematic for their neighbors if

victims attempt to site mobile homes on their own lots while awaiting the reconstruction of permanent

housing. Conflicts arise because such housing usually is considered to be a blight on the neighborhood

and neighbors are afraid the “temporary” housing will become permanent. Neighbors also are pitted

against each other when developers attempt to buy up damaged or destroyed properties and build

multifamily units on lots previously zoned for single family dwellings. Such rezoning attempts are a

major threat to the market value of owner-occupied homes but tend to have less impact on renters because

they have less incentive to remain in the neighborhood. There are exceptions to this generalization

because some ethnic groups have very close ties to their neighborhoods, even if they rent rather than own.

Attempts to change prevailing patterns of civil governance can arise when individuals sharing a

grievance about the handling of the recovery process seek to redress that grievance through collective

action. Consistent with Dynes’s (1970) typology of organizations, existing community groups with an

explicit political agenda can expand their membership to increase their strength, whereas community

groups without an explicit political agenda can extend their domains to include disaster-related

grievances. Alternatively, new groups can emerge to influence local, state, or federal government

agencies and legislators to take actions that they support and to terminate actions that they disapprove.

Indeed, such was the case for Latinos in Watsonville, California following the Loma Prieta earthquake

(Tierney, et al., 2001). Usually, community action groups pressure government to provide additional

163

resources for recovering from disaster impact, but may oppose candidates’ re-elections or even seek to

recall some politicians from office (Olson & Drury, 1997; Prater & Lindell, 2000; Shefner, 1999). The

point here is not that disasters produce political behavior that is different from that encountered in normal

life. Rather, disaster impacts might only produce a different set of victims and grievances and, therefore, a

minor variation on the prevailing political agenda (Morrow & Peacock, 1997).

Emergency Management Interventions

As Figure 6-1 indicates, there are three types of preimpact interventions that can effect reductions

in disaster impacts. Hazard mitigation and emergency preparedness practices directly reduce a disaster’s

physical impacts (casualties and damage) and indirectly reduce its social impacts, whereas recovery

preparedness practices directly reduce a disaster’s social impacts. Improvised disaster response actions

also directly affect disasters’ physical impacts but, by their very nature, are likely to be much less

effective than planned interventions. Similarly, improvised recovery assistance directly affects disasters’

social impacts but is likely to be less effective than systematic recovery preparedness practices.

Figure 6-1 includes the four “phases” of emergency management—mitigation, preparedness,

response, and recovery—but makes it clear there is a complex relationship between them. In reality, these

“phases” might better be called functions, since they are neither discrete nor temporally sequential. Later

chapters will address hazard mitigation (Chapter 7), emergency preparedness (Chapter 9), emergency

response (Chapter 10), and disaster recovery (Chapter 11) in greater detail. However, this section will

provide a brief description of each of these functions and their interrelationships.

Hazard Mitigation Practices

One way to reduce the physical impacts of disasters is to adopt hazard mitigation practices. These

can be defined as preimpact actions that protect passively against casualties and damage at the time of

hazard impact (as opposed to an active emergency response). Hazard mitigation includes hazard source

control, community protection works, land use practices, building construction practices, and building

contents protection. Hazard source control acts directly on the hazard agent to reduce its magnitude or

duration. For example, patching a hole in a leaking tank truck prevents a gas from being released.

Community protection works, which limit the impact of a hazard agent on an entire community, include

dams and levees that protect against floodwater and sea walls that protect against storm surge. Land use

practices reduce hazard vulnerability by avoiding construction in areas that are susceptible to hazard

impact. The use of the term land use practices instead of land use regulations is deliberate. Landowners

can adopt sustainable practices whether or not they are required to do so. Thus, government agencies can

encourage the adoption of appropriate land use practices by providing incentives to encourage

development in safe locations, establishing sanctions to prevent development in hazardous locations, or

engaging in risk communication to inform landowners about the risks and benefits of development in

locations throughout the community.

Hazard mitigation can also be achieved through building construction practices that make

individual structures less vulnerable to natural hazards. Here too, the use of the term building construction

practices rather than building codes is deliberate because building owners can adopt hazard resistant

designs and construction materials in the absence of government intervention. Disaster resistant

construction practices include elevating structures out of flood plains, designing structures to respond

more effectively to lateral stresses, and providing window shutters to protect against wind pressure and

debris impacts. Government agencies can encourage the adoption of appropriate building construction

164

practices by providing incentives to encourage appropriate designs and materials, establishing code

requirements for hazard resistant building designs and materials, or informing building owners about the

risks and benefits of different building designs and materials. Finally, hazard mitigation can be achieved

by contents protection strategies such as elevating appliances above the base flood elevation or bolting

them to walls to resist seismic forces.

Emergency Preparedness Practices

Another way to reduce a disaster’s physical impacts is to adopt emergency preparedness

practices, which are preimpact actions that provide the human and material resources needed to support

active responses at the time of hazard impact (Lindell & Perry, 2000). The first step in emergency

preparedness is to use community hazard/vulnerability analysis (HVA) to identify the emergency

response demands that must be met by performing four basic emergency response functions—emergency

assessment, hazard operations, population protection, and incident management (Lindell & Perry, 1992,

1996b). Emergency assessment consists of actions that define the potential scope of the disaster impacts

(e.g., projecting hurricane wind speed) whereas hazard operations consists of short-term actions to

protect property through expedient hazard mitigation actions initiated during an emergency (e.g.,

sandbagging around structures). Population protection actions protect people from impact (e.g., warning

and evacuation) and incident management actions activate and coordinate the emergency response (e.g.,

communication among responding agencies). The next step is to determine which community

organization will be responsible for accomplishing each function (Federal Emergency Management

Agency, 1996b). Once functional responsibilities have been assigned, each organization must develop

procedures for accomplishing those functions. Finally, the organizations must acquire response resources

(personnel, facilities, and equipment) to implement their plans and they need to maintain preparedness for

emergency response through continued planning, training, drills, and exercises (Daines, 1991).

Recovery Preparedness Practices

Just as emergency preparedness practices are preimpact actions intended to develop the human

and material resources needed to support an active emergency response, recovery preparedness practices

are preimpact actions that are intended to develop the financial and material resources needed to support a

prompt and effective disaster recovery. First of all, households and businesses need to prepare for disaster

recovery by purchasing hazard insurance because this will provide the money they need to rebuild

damaged structures and replace destroyed contents. However, hazard insurance varies significantly in its

availability and cost—flood, hurricane, and earthquake insurance being particularly problematic

(Kunreuther & Roth, 1998). Moreover, some ethnic groups cannot afford the rates of high quality

insurance companies or are denied coverage altogether (Peacock & Girard, 1997).

In addition, the governments of hazard prone communities need to prepare for actions including

impact assessment, debris management, infrastructure restoration, housing recovery, economic recovery,

and linkage to hazard mitigation. It seems to be commonly thought that the development of disaster

recovery operations plans (ROPs) can be delayed until after disaster strikes, but practitioners and

researchers agree that community disaster recovery is faster and more effective when it is based on a plan

that has been developed prior to disaster impact (Geis, 1996; Olson, Olson & Gawronski, 1998; Schwab,

et al., 1998; Wilson, 1991; Wu & Lindell, 2004).

There are six important features of a preimpact ROP. First, it should define a disaster recovery

organization. Second, it should identify the location of temporary housing because resolving this issue

can cause conflicts that delay consideration of longer-term issues of permanent housing and distract

165

policymakers altogether from hazard mitigation (Bolin & Trainer, 1978; Bolin, 1982). Third, the plan

should indicate how to accomplish essential tasks such as damage assessment, condemnation, debris

removal and disposal, rezoning, infrastructure restoration, temporary repair permits, development

moratoria, and permit processing because all of these tasks must be addressed before the reconstruction of

permanent housing can begin (Schwab, et al., 1998).

Fourth, preimpact recovery plans also should address the licensing and monitoring of contractors

and retail price controls to ensure victims are not exploited and also should address the jurisdiction’s

administrative powers and resources, especially the level of staffing that is available. It is almost

inevitable that local government will have insufficient staff to perform critical recovery tasks such as

damage assessment and building permit processing, so arrangements should be made to borrow staff from

other jurisdictions (via pre-existing Memoranda of Agreement) and to use trained volunteers such as local

engineers, architects, and planners. Fifth, these plans also need to address the ways in which recovery

tasks will be implemented at historical sites (Spennemann & Look, 1998). Finally, preimpact recovery

plans should recognize the recovery period as a unique time to enact policies for hazard mitigation and

make provision for incorporating this objective into the recovery planning process.

Conducting Community Hazard/Vulnerability Analyses

The model described in the previous pages provides a framework for understanding how hazard

agent characteristics produce physical and social impacts that can be ameliorated by emergency

management interventions. The next section will describe how local emergency managers can use this

framework to guide assessments of their communities’ exposure to specific hazards and vulnerability to

physical and social impacts. Although this section focuses on how to conduct hazard/vulnerability

analyses (HVAs), it is important to remember who should be involved in the process. The HVAs provides

critical information for the community’s hazard mitigation, emergency preparedness, and recovery

preparedness practices so it should involve a wide range of the community’s emergency management

stakeholders (Pearce, 2003). This is an important reason for the LEMC to conduct the community HVA

through the LEMC’s Hazard/Vulnerability Analysis subcommittee.

Mapping Natural Hazard Exposure

States and local jurisdictions across the country vary in their exposure to the hazards described in

Chapter 5. Consequently, an important objective for a local emergency manager is to identify the hazards

that his or her community should set as priorities for its emergency management program. There are

many useful sources of information about the regional incidence of these hazards, one of which is the set

of maps contained in the Federal Emergency Management Agency’s (1997) Multi Hazard Identification

and Risk Assessment. This source has an extensive set of maps describing exposure to natural hazards and

also addresses some technological hazards. The maps of natural hazard exposures contained in Multi

Hazard Identification and Risk Assessment can be supplemented by visiting the websites of FEMA

(www.fema.gov), the US Geological Survey (www.usgs.gov), and the National Weather Service

(www.nws.noaa.gov). These maps provide a good start toward assessing the potential impacts of

disasters, but they have three limitations. First, many of these large scale maps are designed to compare

the relative risk of broad geographical areas (i.e., regions of the country). This information enables local

emergency managers to identify the hazards that could strike their jurisdictions, but it does not provide

enough resolution to tell them which areas within their jurisdictions are most likely to be struck by a

disaster. For example, a coastal county might be exposed to hurricanes but, as Chapter 5 indicated, only

166

part of the jurisdiction is likely to experience significant damage. Consequently, smaller scale maps are

needed to assess exposure of different areas to storm surge, inland flooding, and high wind.

Second, these maps vary from one hazard to another in terms of whether they define risk areas in

terms of event magnitudes or in terms of recurrence intervals. For example, hurricane risk area maps

identify areas that are expected to be affected by Category 1-5 hurricanes. However, these maps provide

no information about the probability that each of these different hurricane intensities would occur. By

contrast, US Geological Survey earthquake hazard maps plot the peak ground acceleration (PGA) with a

2% probability of exceedance in 50 years. Thus, these maps provide useful information about the areas in

which buildings are most likely to collapse and some indication likelihood of a disaster. However, as will

be explained in the section on physical vulnerability, trained engineers would be needed to use this

quantitative information to assess the probabilities of building failure.

Third, these maps are insufficient for local HVAs because local emergency managers also need to

assess the relative risk of different hazards for a given geographical area. That is, emergency managers

need to know whether their jurisdictions are at risk for certain hazards, but they also need to know what is

the likelihood of a flood in comparison to a tornado, an earthquake, or a toxic chemical release.

As a result of these limitations, local emergency managers must often settle for qualitative

comparisons of the relative risk of different hazards. That is, they must typically categorize the

probability of disaster impact as high, medium, or low (Federal Emergency Management Agency, 1996b).

Such categorization provides only a “rough screen” for determining which hazards require the most

attention, but this limitation is often more apparent than real because many hazards impose similar

demands on the community, especially during the emergency response phase. Specifically, the equipment

and methods used for emergency assessment and hazard operations might differ among the

meteorological, hydrological, geophysical, and hazardous materials hazards, but the equipment and

methods used for population protection and incident management will be quite similar. Consequently,

differences in hazard probability are often unimportant because preparedness for one provides

preparedness for many, if not all, other hazards. Nonetheless, many mitigation measures are hazard-

specific, so an understanding of the relative risk from different hazards can provide valuable guidance in

determining which investments are most likely to reduce a community’s vulnerability.

Mapping Hazmat Exposures

Incidents involving fires, explosions, or chemical releases can be initiated by internal (accident or

sabotage) or external (geophysical, meteorological, or hydrological events, or terrorist attacks) causes.

The types of hazards that can occur at a chemical facility, their initiating events, their consequences, and

their likelihoods of occurrence can be assessed using hazard analysis (Huebner, et al., 2000). This process

begins by identifying dangerous chemicals (i.e., those that are threats because of their flammability,

reactivity, or toxicity), their locations, and the quantities stored at those locations. Once the chemical

inventory has been developed, this information can be used to assess the threats these chemicals pose to

the facility, its workers, its neighbors, and the environment. In the case of Extremely Hazardous

Substances (EHSs) defined under SARA Title III, Vulnerable Zones (VZs) can be computed using data

on the chemical’s toxicity, its quantity available for release, the type of spill (liquid or gaseous), the

postulated release duration (e.g., 10 minutes), assumed meteorological conditions (wind speed and

atmospheric stability), and terrain (urban or rural). Available methods include manual computations (US

Environmental Protection Agency, 1987), ALOHA (Federal Emergency Management Agency, no date, a,

see information about CAMEO at www.epa.gov/ceppo/cameo), or RMP*Comp at

167

yosemite.epa.gov/oswer/ceppoweb.nsf/content/ rmp-comp.htm). Once the radii of the VZs for the

different chemicals have been computed, these can be overlaid onto a map with the release point in the

center of the circle and the radius drawn around it (see Figure 6-2).

Emergency managers also should work with their LEPCs to identify the highway, rail, water, and

air routes though which hazardous materials are transported. Once these routes have been identified, the

number of tank trucks, railroad tankcars, and barges carrying each type of hazardous material can be

counted in a commodity flow study. Information about hazardous materials transportation can be found on

the US DOT website (hazmat.dot.gov) and specific guidance for commodity flow studies is found at

hazmat.dot.gov/hmep/guide_flow_surveys.pdf. Once the chemicals being transported have been

identified, analysts can use the same procedures that were used for fixed site facilities. As Figure 6-2

indicates, this will lead to the construction of rectangular VZs surrounding the transportation routes. The

facility and transportation route VZs can then be examined to identify areas of residential, commercial,

and industrial land use (see Lindell, 1995, for an analysis of hazardous waste transportation to an

incinerator).

Figure 6-2. Vulnerable Zones around fixed-site facility and transportation route.

Source: Adapted from Lindell (2006)

If VZs for the transportation routes have not been prepared before an incident occurs, the North

American Emergency Response Guidebook (hazmat.dot.gov/ohmform.htm#erg) can be used to

approximate them. This document contains a table listing the principal chemicals commonly found in

transportation, each chemical’s identification number, and the emergency response guide that should be

used to provide technical assistance in responding to a spill. In addition, the table identifies the most

dangerous chemicals and directs emergency responders to a Table of Initial Isolation and Protective

Action Distances that should be used to determine where protective action (either evacuation or shelter in-

place) should be implemented. Emergency managers should understand that the Emergency Response

Guidebook classifies releases only as small or large, so use of the procedures in the Technical Guidance

for Hazards Analysis or more advanced methods identified in the Handbook of Chemical Hazard Analysis

Procedures is preferred to the table in the Emergency Response Guidebook.

Town of Buckley

Hazmat Inc. H ig

h w

a y 1

0 1

Fixed site facility

Facility Vulnerable Zone Transportation route Vulnerable Zone

Hazmat transportation

route

H ig

h w

a y

1 0

1

168

A special case of hazmat exposures arises in connection with nuclear power plants. The NRC

conducted extensive analyses to define the emergency planning zones (EPZs) that should be designated

around these facilities. The NRC’s analyses led to the establishment of a 10 mile radius plume inhalation

EPZ in which state and local authorities should be prepared for people to evacuate or shelter in-place to

avoid inhalation exposure and direct radiation from a radioactive plume (US Nuclear Regulatory

Commission, 1978). In addition, there is a 50 mile radius ingestion pathway EPZ in which authorities

should prepare to monitor water, milk, and food (especially leafy green vegetables) for contamination.

Mapping Exposure to Secondary Hazards

Emergency managers should recall that some disaster impacts can initiate others. As noted in

Chapter 5, earthquakes can cause surface faulting, ground failure, landslides, fires, dam failures, and

hazardous materials releases in addition to the expected structural failures caused by ground shaking. One

method of identifying areas exposed to multiple hazards is to use a GIS to overlay the areas subject to

these different hazards. This is accomplished by entering all of the data on primary and secondary hazard

exposures and special facilities into a GIS that creates separate layers for fault lines; areas prone to the

highest levels of ground shaking, subsidence, and landsliding; hazardous facility Vulnerable Zones; and

locations of sensitive facilities. Next, these layers are intersected to produce composite maps displaying

the areas subject to multiple hazards. Finally, the layers identifying the locations of residential,

commercial, and industrial areas, and sensitive facilities are overlaid to produce the final maps. The most

common secondary hazards of the events described in Chapter 5 are listed in Table 6-1.

Table 6-1. Secondary Hazards. Primary Hazard Secondary Hazards

Severe storms Floods, tornadoes, landslides

Extreme summer weather Wildfires

Tornadoes Toxic chemical or radiological materials releases

Hurricane wind Toxic chemical or radiological materials releases

Wildfires Landslides (on hillsides in later rains)

Floods Toxic chemical or radiological materials releases

Storm surge Toxic chemical or radiological materials releases

Tsunamis Toxic chemical or radiological materials releases

Volcanic eruptions Floods, wildfires, tsunamis

Earthquakes Fires, floods (dam failures), tsunami, landslides, toxic chemical or radiological materials releases

Landslides Tsunami

Assessing Physical Vulnerability

Information on hazard exposure needs to be supplemented with information on physical

vulnerability of structures and people. This makes it necessary to identify the types of structures and

populations that are located in the areas exposed to environmental hazards.

Structural Vulnerability to Wind, Seismic, and Water Forces

Structures can be vulnerable to environmental hazards because of inadequate designs, inadequate

construction materials, or both. Older homes have been constructed under earlier building codes and most

neighborhoods are relatively homogeneous with respect to age, so identifying older neighborhoods in

hazard-prone areas will help set priorities of emergency management interventions. For example, many

areas of the country have homes, built during the early part of the 20 th

Century using unreinforced

masonry (brick walls constructed without steel reinforcing rods), that are especially vulnerable to

169

earthquakes. Similarly, older homes are usually less weathertight, so they are much more prone to

infiltration of hazardous materials.

There are three major issues in assessing structural vulnerability. First is the question of whether

the structure has the strength or resilience to withstand environmental forces such as wind, seismicity, or

water. In this case, the concern is about the impact on the structure itself and, consequently, the loss of

function and the time and cost of rebuilding. The second issue concerns the ability of the structure to

protect the contents. This issue is distinct from the first one because in earthquakes, for example,

buildings that survive ground shaking without damage can transmit the motion to light fixtures, cabinets,

and furniture—possibly damaging these items. The third issue concerns the ability of the structure to

protect the occupants. This is especially important in connection with hazardous materials because they

can infiltrate into a structure and kill the occupants without damaging the building.

Once the areas at risk from environmental hazards have been identified, emergency managers

should identify the types of residential, commercial, and industrial land uses located within them. It is

particularly important to determine if there are any facilities within each VZ that have highly vulnerable

populations. A list of such facilities is listed in Table 6-2 and important characteristics of the facility users

are listed in Table 6-3. These characteristics make it much more problematic to evacuate facility

occupants than households (Vogt, 1991). In addition, it is important to identify the location of critical

infrastructure facilities (see Table 6-4).

Table 6-2. Facilities With Highly Vulnerable Populations. HEALTH RELATED

Hospitals

Nursing homes

Halfway houses (drug, alcohol, mental retardation)

Mental institutions

PENAL

Jails

Prisons

Detention camps

Reformatories

ASSEMBLY & ATHLETIC

Auditoriums

Theaters

Exhibition halls

Gymnasiums

Athletic stadiums or fields

AMUSEMENT & RECREATION

Beaches

Camp/conference centers

Amusement parks/fairgrounds/race courses

Campgrounds/Recreational Vehicle parks

Parks/lakes/rivers

Golf courses

Ski resorts

Community recreation centers

RELIGIOUS

Churches/synagogues

Evangelical group centers

HIGH DENSITY RESIDENTIAL

Hotels/motels

Apartment/condominium complexes

Mobile home parks

Dormitories (college, military)

Convents/monasteries

TRANSPORTATION

Rivers/lakes

Dam locks/toll booths

Ferry/railroad/bus terminals

COMMERCIAL

Shopping centers

Central business districts

Commercial/industrial parks

EDUCATIONAL

Day care centers

Preschools/kindergartens

Elementary/secondary schools

Vocational/business/specialty schools

Colleges/universities

Source: Lindell and Perry (1992).

170

Of course, the assessment of structural vulnerability usually involves all three issues. For riverine

flooding and hurricane storm surge, structures—especially concrete structures with well-anchored

foundations—resist battering waves to protect the structure and provide the height to escape the rising

water that could threaten contents and occupants. In other cases, it is the strength of construction in

resisting wind loads (tornadoes and hurricanes), blast forces (explosions and volcanic eruptions) and

ground shaking (earthquakes) that protects the structure, contents, and occupants. For chemical,

radiological, and volcanic ash threats, it is the tightness of construction in preventing the infiltration of

outside (contaminated) air into the structure that is the important protective feature. Finally, in the case of

exposure to a cloud of radioactive material, the construction material can provide shielding from

penetrating radiation and from surface contamination.

Table 6-3. Characteristics of Special Facility Users.

Characteristics Of Users Special Considerations

Mobility of users Ambulatory

Require close supervision

Nonambulatory

Require life support

Permanent residence of users

Facility residents

Residents of the impact area, but not of the facility (e.g., prison guards)

Transients

Periods of use Days of week/hours of day

Special events

User density Concentrated

Dispersed

Sheltering in place Highly effective

Moderately effective

Minimally or not effective

Transportation support Would use own vehicles

Require buses or other high occupancy vehicles

Require ambulances

Source: Lindell and Perry (1992).

In high wind (including tornadoes and hurricanes) and explosions (usually technological in

origin, but also including some volcanic eruptions), a substantial increase in air pressure can cause

structures to collapse. Such structural failures are caused by deficiencies in either design or materials, or

both (American Institute of Architects, 1995; Institute for Business and Home Safety, 1997).

Fortified homes (see www.ibhs.org) provide for installation of connections and braces to

reinforce roofs and gable-end walls against wind attack. In addition to positive pressure on upwind walls,

high wind creates negative pressure, or suction, as it is forced to flow up and over the roof. This suction is

greatest in flat roofs, intermediate in gable-end roofs (which slope in two directions), and least in hipped

roofs (which slope in four directions). Suction tends to lift the roof from the walls unless resisted by

adequate connections to the walls—which in turn, must have adequate connections to the foundation.

Adequate designs and materials also provide protection to building openings such as windows and doors,

thus preventing the wind from pressurizing the interior and adding to the stress on roof and walls. The

need for window and sliding glass door shutters is widely recognized because the shutters resist the direct

pressure of the wind and the impact of flying debris. However, door reinforcement is also important—

especially for double-wide (two-car) garage doors that are highly susceptible to failure because their great

width allows the wind to deflect them inward and pull the rollers out of their tracks.

171

Table 6-4. Infrastructure Facilities.

Facility Type Examples

Water Pumping stations and pipelines

Sewer Pumping stations and pipelines

Electric power Generating stations and power lines

Liquid (e.g., oil) and gas (e.g. natural gas) fuels Pumping stations and pipelines

Telecommunications Broadcast studios, transmission towers

Cellular telephone towers

Telephone switching centers and telephone lines

Transportation Bus/truck terminals and roads

Rail yards and rail lines

Sea and inland marine ports

Airports

Public safety and health Police and fire stations

Ambulance garages

Hospitals

Similar observations apply to earthquakes; building damage typically results from the lateral

pressures (ground shaking, surface faulting, and soil failure) exerted against a structure that was designed

principally to resist the vertical loads resulting from the weight of the occupants, furniture, upper stories,

and roof (American Institute of Architects, 1992). Rigid structures such as unreinforced masonry are

extremely vulnerable, whereas wood frame dwellings are much safer. In the latter case, there might be

rigid portions of a structure, such as brick chimneys, that separate from the rest of the structure and

collapse into the living area. In addition, glass from broken windows, falling pictures and mirrors, the

toppling of unsecured furniture, and other flying debris are safety hazards. For hurricanes, structures on

the open coast must be of sufficiently sturdy construction that they can resist the direct impact of storm

surf as well as the force of extremely high winds. However, both riverine flooding and hurricane storm

surge also require the structure to have foundations anchored well enough to resist scouring by water

currents that can undermine building foundations and cause structural collapse. Additional protection can

be provided by expedient floodproofing that uses waterproof construction materials, sealing of cracks,

provision of valves on sewer lines, steel bulkheads for lower-level openings, and sump pumps to eject

seepage (US Office of Emergency Preparedness, 1972).

As with hurricane surge and riverine flooding, volcanic mudflows and floods present the

challenge of maintaining the integrity of buildings and their foundation. However, flooding generated by

volcanic eruption commonly contains a substantial volume of rock and ash, resulting in mudflows that

have substantial carrying power (Hays, 1981). Moreover, as was found during the eruption of Mt. St.

Helens, silt buildup can significantly raise the bed of the river channel (Perry & Lindell, 1990).

Tsunami impact poses an even greater threat than inland flooding, surge from hurricanes and

coastal storms, or volcanic mudflows. As noted in Chapter 5, these seismic sea waves can threaten areas

as much as 100 feet above sea level, so destruction is highly likely for most structures located very near

the shoreline. However, properly designed steel-reinforced concrete structures located a short distance

inland are likely to survive even the largest tsunamis.

Human Vulnerability to Inhalation Exposure

In the case of radiological or toxic materials, the principal public health hazard arises from

inhalation of airborne materials that have an adverse health effect (US Environmental Protection Agency,

172

1987). Inhalation exposures can also result from the dispersion of airborne debris such as ash and gas

from volcanic eruptions. Ideally, an enclosed space will provide a barrier if it can be closed tightly

enough to keep out the hazardous material and has enough oxygen to sustain those within it until the

danger has passed. Unfortunately, most structures are leaky, allowing contaminated air to infiltrate even

when the doors and windows are closed. The rate of air exchange increases with the amount of leakage

area, the wind speed, and the temperature differential between the indoor and outdoor air.

The rate at which indoor and outdoor air are exchanged is commonly measured in air changes per

hour (ACH). However, emergency managers will find it more useful to think of air exchange in terms of

turnover time, which the reciprocal of the air exchange rate, or tB = 1/ACH, where ACH is the number of

air changes per hour. As Wilson (1987) emphasizes, an infiltration rate of 1.0 ACH does not imply that all

the clean air will be gone in one hour. Rather, the proportion of contaminated air gradually rises until at

the end of 1.0 tB hours, 63% of the original air has been replaced by contaminated air, while 95% of the

original air has been replaced by the end of 3.0 tB hours. Thus, for the case of 1.0 ACH, it will take over

three hours (not just one hour) for the indoor air to become almost completely contaminated. This result is

extremely important because it indicates that in-place sheltering is more effective than most people might

infer from the apparent implication of the number of air changes per hour. The reason for the difference

between the apparent result and the correct result can best be illustrated by examining the difference

between the apparent, but incorrect, mechanism of air exchange and the actual mechanism. It would only

take one hour to replace the clean air (the incorrect result) if the contaminated air somehow “pushed out”

the clean air, but this is not what happens. Rather, the contaminated air that infiltrates into the structure

mixes with the clean air rather than “pushing it out”. Clearly, exfiltration of a mixture of clean air and

contaminated air will take longer to exhaust the clean air in a structure than will exfiltration of (“pushing

out”) clean air alone. Consequently, sheltering in-place is at least three times as effective in reducing

inhalation exposure as it first appears to be.

While this time lag effect is important, it is not the only mechanism by which sheltering in-place

can reduce adverse health effects. It is also important to recognize the impact of a damping effect in

reducing the fluctuations in plume concentrations. As was discussed in the previous section, these

fluctuations arise from irregularities in meteorological conditions and local terrain. One way of measuring

peak concentration is by estimating the value that is exceeded approximately 1% of the time. Wilson

(1987) reports that in the outdoor (contaminated) air, such 1% peak concentrations are 400% as large as

the mean concentration. For indoor air, the equivalent peak concentration is only 50% larger than the

mean. As he notes, even when the indoor concentration (100 parts per million [ppm], for example) has

risen after six hours to match the outdoor concentration, the indoor peaks would be expected to be

approximately 150 ppm when the outdoor peaks would be 400 ppm. This is, of course, of considerable

significance when peak concentrations are the principal health threats.

As Wilson (1987, 1989) and others have observed, a major problem in assessing the effectiveness

of sheltering in-place is uncertainty about whether indoor air concentrations will remain sufficiently low

for a sufficiently long period of time. This can be answered definitively only if there is information about

the hazardous material being released (especially the identity of the material released, and the rate and

duration of the release), the meteorological data needed for a computerized plume dispersion model (wind

speed, wind direction, and atmospheric stability), and the air exchange rates for the structures in the

hazard impact area. Data on the release and the meteorological conditions will not be available until an

incident occurs, but data on the efficacy of sheltering in-place can be collected in advance. First, Rogers,

173

et al. (1990) report that energy conservation research has shown air exchange in most US dwellings

ranges from 0.5 to 1.5 ACH. Second, Wilson (1989) reports that the most important factor affecting

leakage area is the presence of a vapor barrier in the walls and ceiling of a structure, a feature that is most

common in houses built in cold climates after 1960. Thus, emergency managers could estimate the

effectiveness of sheltering in-place by obtaining access to local data on the age of the housing stock

within different neighborhoods within their jurisdiction. Finally, the fact that so much of the research and

data on infiltration of hazardous materials has been developed from studies of energy conservation

suggests that emergency planners consult with their local utilities to determine what information is

available regarding the air exchange rates of different types of structures (e.g., residences, schools, and

commercial buildings) in their communities. Special facilities, especially those such as hospitals that have

low mobility residents, should be examined individually to assess their air exchange rates.

Human Vulnerability to Radiological Materials

Although both toxic and radiological materials present an inhalation hazard, a plume of

radioactive material released from a nuclear power plant or during a transportation accident also can

cause harm by means of external gamma radiation from the cloud and from ground contamination. Dense

building materials such concrete, brick, and stone provide shielding from external gamma radiation and,

thus, can provide a basis for in-place sheltering during radiological emergencies. The effectiveness of

structures made from different types of building materials has been examined in studies by Burson and

Profio (1977), Anno and Dore (1978a; 1978b), Aldrich, Ericson and Johnson (1978), and Aldrich, et al.

(1982). These studies calculated the dose reduction factors (the ratio of the dose received while sheltering

to the unprotected dose) for three exposure routes: external gamma radiation from the cloud, external

gamma radiation from ground contamination, and inhalation of radioactive materials infiltrating into the

structure. Burson and Profio (1977) found that sheltering in a wood frame dwelling provides little more

protection from cloud and ground exposure than does “sheltering” in a vehicle while evacuating.

Sheltering on the ground floor of a masonry home with no basement or in the basement of a wood frame

home gave considerably higher levels of protection: about 50% of the unprotected exposure to the cloud

and less than 20% of the unprotected exposure to ground. As one might expect, the basement of a

masonry house was even more effective: 40% of the cloud exposure and 5% of the ground exposure. A

large office building was the most effective shelter of all, reducing cloud exposure to about 20% and

ground exposure to 1%.

The importance of the construction materials is underscored by Burson and Profio's (1977) work

indicating it is the cloud exposure that produces most of the whole body radiation dose received by those

sheltering in a home. Infiltration into the structure would account for only about 5% of the gamma

radiation dose. Anno and Dore (1978a) calculated cloud dose reduction factors for single family

dwellings and large structures (e.g., office buildings, multistory apartment complexes). They considered

0.125 to 3 ACH to define the range of infiltration rates for single family dwellings and other structures

that could be used as temporary public shelters. For single family dwellings, whole body dose reduction

factors for low air exchange rates (0.125 ACH) were calculated to be 0.40-0.33 compared to 0.43 for

more representative air change rates (3 ACH). For large structures, whole body dose reduction factors for

low air change rates were calculated to be 0.08 compared to 0.17-0.11 for the more representative air

change rates. These investigators also estimated thyroid (inhalation) dose reduction factors to be about

0.05 to 0.01 for low air change rates and from 0.25 to 0.10 for more representative air change rates for

either single family dwellings or large structures.

174

More recent analyses by the NRC provide an indication of vulnerability to radiological materials

releases. The results of these analyses can be seen in Figure 6-3, which displays the relative effectiveness

of continuing normal activity, sheltering in a home basement or large building, or evacuating before or

after plume arrival at three distances from the plant. According to this figure, which is adapted from

McKenna (2000), even a late evacuation is better than home shelter, large building shelter is better than

late evacuation, and early evacuation is best of all. Thus, those who are within five miles of a nuclear

power plant have high vulnerability if they remain in their homes during a release. Only large building

shelter provides as much safety as early evacuation.

Figure 6-3. Effectiveness of Protective Actions in a Nuclear Power Plant Emergency.

1 Normal activity 4 Evacuation 1 hour before plume arrival 2 Home basement shelter 5 Evacuation after plume arrival 3 Large building shelter

Source: Adapted from McKenna (2000).

Assessing Agricultural and Livestock Vulnerability

Assessing the physical vulnerability of crops and livestock is a task that is rarely considered to be

the responsibility of emergency managers. One reason for giving minimal emphasis to the agricultural

sector is that it accounts for a relatively small part of the total vulnerability in many jurisdictions. In those

cases where the agricultural sector is a significant part of the local economy, emergency managers should

consult agricultural experts such as those from the US Department of Agriculture because, as noted

earlier, there is substantial variation among animal and plant species in their susceptibility to extreme

environmental conditions. For example, fruit orchards can be devastated by wind speeds that have no

impact whatsoever on rangeland. Moreover, the damage to many crops depends on the stage in growth

cycle—with some crops having minimal susceptibility to wind damage until just before harvest.

Assessing and Mapping Social Vulnerability

In contrast to physical vulnerability, which arises from the potential for environmental extremes

to create adverse physiological changes, social vulnerability arises from the potential for these extreme

P ro

b a

b il it y o

f e

x c e

e d

in g

2 0

0 r

e m

(2

S v )

175

events to cause changes in people’s behavior. People can vary in their potential for injury to themselves

and their families. They also vary in the potential for destruction of their homes and workplaces, as well

as the destruction of the transportation systems and locations for shopping and recreation they use in their

daily activities. The discussion below emphasizes census data but it also is important to examine other

archival sources such as school records, immigration services, local aging agency, special needs

registries, property tax records, facilities locations (Enarson, et al., 2003). In addition, consult local social

service providers (government and NGO) and churches to identify vulnerable populations.

Assessing Psychosocial Vulnerability

One important component of psychological vulnerability is personal fragility—that is, a lack of

emotion-focused coping skills. Another component of psychological vulnerability is rigidity—that is, a

lack of problem-focused coping skills defined by an inability to develop adaptive strategies for

responding to altered conditions. Ozer and Weiss’s (2004) summary of research on post-traumatic stress

disorder (PTSD) concluded the four categories of PTSD predictors were

• A person’s pre-existing characteristics (e.g., intelligence, previous psychological trauma),

• The severity of the personal impact of the disaster,

• Psychological processes immediately after the impact, and

• Life stress and social support after the traumatic event.

Quite obviously, only the first of these categories can measure psychological vulnerability that

exists before a disaster strikes and none of the variables in this category is routinely available through

secondary sources such as Census data. Because direct measures of the incidence of PTSD predictors

(e.g., through community surveys) are prohibitively expensive, psychological vulnerability must be

measured indirectly, as discussed in a later section.

A major concern is social isolation. Thus, vulnerability is also measured by the infrequency and

superficiality of social contacts with peers such as kin (extended family), neighbors, and coworkers.

Routine measures of social vulnerability are rarely available through surveys conducted using

representative samples of community members. However, there are proxy variables that have statistically

significant—although admittedly small—correlations with social isolation. Suitable proxy variables that

are routinely available through Census files include age, income, and ethnicity. Specifically, increasing

age is associated with reduced levels of community participation (involvement in voluntary associations)

and immersion in kin and friendship networks (Perry, 1985; Perry, et al., 1981). By contrast,

socioeconomic status is positively associated with participation in community organizations (Alvirez &

Bean, 1976; Tomeh, 1973) and minority ethnicity is positively associated with immersion in kin and

friendship networks (Staples, 1976; Wilkinson, 1999). Accordingly, the use of age, socioeconomic status,

and ethnicity as proxy measures of psychosocial vulnerability will also be discussed below.

Assessing Demographic Vulnerability

Vulnerability to demographic changes follows from the demographic balancing equation

discussed earlier. Until 2005, recent trends in disaster casualties had indicated the number of deaths

would be relatively small for any North American disaster. It is unlikely that Hurricane Katrina marks a

reversal of that trend because New Orleans is the only major coastal city with a significant portion of its

land area (and, thus, its population) below sea level. In most disasters, it is the number of in-migrants and

out-migrants cause significant changes in its demographic composition. Once again, census data can be

used to provide indicators such as age, income, homeownership, and ethnicity. Older, more affluent

homeowners are likely to have high levels of community bondedness (Turner, et al., 1986) and seek

176

permanent housing in the community even if their homes have been destroyed. Similarly, ethnic

minorities have tightly integrated kin networks that make them stay. However, even otherwise stable

communities are likely to experience short term changes in their demographic composition if there are

few rental vacancies after a disaster. This is because a local housing shortages would require residents to

move farther away for temporary housing. Similarly, a community’s demographic composition is likely to

change if local businesses have high levels of physical vulnerability to disaster impacts (making them

likely to shut down for extended periods). A declining local economy will make them financially

vulnerable and, thus, more likely to cease operations altogether.

Assessing Economic Vulnerability

It is obvious that wealth is a major component of economic vulnerability, but the assets

comprising wealth vary in their vulnerability to disasters. Tangible assets such as buildings, equipment,

furniture, and vehicles that are located in the disaster impact area are more vulnerable than financial

assets such as bank accounts, stocks, and bonds that are recorded electronically. Households and

businesses both have tangible and financial assets, so both are vulnerable to the loss of their tangible

assets and both have financial assets that can be used to support disaster recovery. Of course, there are

substantial variations among households in their assets and the same is true for businesses.

One noteworthy difference between households and businesses is that the latter also have

operational vulnerability arising from dependency upon those who supply its inputs (suppliers and labor)

as well as those who purchase its outputs (distributors and customers). Evidence of businesses’

operational vulnerability to input disruptions can be seen in data provided by Nigg (1995), who reported

that business managers’ median estimate of the amount of time that they could continue to operate

without infrastructure was 0 hours for electric power, 4 hours for telephones, 48 hours for water/sewer,

and 120 hours for fuel. If this infrastructure support is unavailable for time periods longer than these, then

businesses must suspend operations even if they have suffered no damage to their structures or contents.

Measures of household wealth are not available, but data on household income are available in

census files. Available census data on businesses are more limited in their relevance to economic

vulnerability. The Census Bureau’s website (censtats.census.gov) provides ZIP code-level data on the

number of businesses in each economic sector, broken down by number of employees. These data can be

overlaid onto risk areas for different hazards such as 500-year floodplains, hurricane surge zones, or

earthquake seismic zones to develop estimates of the community’s economic vulnerability to disaster

impact.

Assessing Political Vulnerability

As will be discussed later, political impacts of disasters often arise from conflicts over the

management of the emergency response and disaster recovery. Accordingly, political vulnerability arises

from inadequate emergency management interventions—which create situations that pit one group of

stakeholders against another—and inadequate mechanisms for managing this conflict when it does arise.

The adequacy of emergency management interventions to reduce hazard vulnerability will be addressed

in Chapters 9-12, but the adequacy of mechanisms for managing conflict is a crucial part of civil

governance. In particular, government agencies that are believed to lack legitimacy, expertise, and

adequate information for making decisions about the allocation of public resources will prove vulnerable

in the aftermath of disaster. As is the case with psychological, demographic, and economic vulnerability,

there currently are no direct measures of political vulnerability that are readily available for use by

177

emergency managers. Consequently, the use of age, socioeconomic status, and ethnicity as proxy

measures of political vulnerability will also be discussed in the next section.

Predicting Household Vulnerability

As noted earlier in this chapter, it is important to recognize that social vulnerability is not

randomly distributed either demographically or geographically. In particular, the social vulnerability

arising from a lack of psychological resilience, social network integration, economic assets, and political

power vary across demographic groups. Some of these components of social vulnerability can be

predicted by demographic characteristics such as gender, age, education, income, and ethnicity.

Moreover, these demographic groups tend to be distributed relatively systematically across the landscape

of each community. Even though there might not be sharp geographic lines of demarcation between the

locations of different demographic groups within a community, there are variations in the concentration

of these groups in different neighborhoods. Thus, GISs can be used to conduct disaggregated (e.g., census

tract-level) spatial analyses to identify the demographic segments most likely to be vulnerable to disaster

impacts.

The demographic predictors (e.g., gender, age, education, income, and ethnicity) of social

vulnerability are frequently also associated with hazard exposure because the population segments with

the fewest psychological, social, economic, and political resources often disproportionately occupy the

most hazardous geographical areas. Similarly, demographic predictors of social vulnerability are often

associated with structural vulnerability because those same population segments disproportionately

occupy the oldest, most poorly maintained buildings. Thus, those who are most socially vulnerable are

also likely to experience the greatest physical impacts such as casualties and property loss.

Because emergency managers rarely have access to direct measures of social vulnerability,

geographic analyses of social vulnerability are conducted on Census data, preferably at the lowest

possible level of aggregation (e.g., block-group or tract). Recent research has shown these aggregated

indicators of social vulnerability are strongly correlated, so it is advisable to use either a composite

measure of social vulnerability or a subset of these indicators. Table 6-5 lists a sample set of social

vulnerability indicators recently used in analyses of social vulnerability to the impacts of earthquakes in

Shelby County (Memphis), Tennessee (Prater, et al., 2004).

Based on the recognition, described above, that hazard exposure, structural vulnerability, and

social vulnerability tend to be related, Prater and her colleagues advocated identifying vulnerability

hotspots—the geographic areas occupied by demographic segments that are most vulnerable to disaster

impacts. These vulnerability hotspots can be identified by using a GIS to either overlay or mathematically

combine data on hazard exposure (e.g., ground motion and ground failure from earthquakes), structural

vulnerability (e.g., due to dilapidated housing), and lifeline vulnerability (e.g., due to old and poorly

maintained water, sewer, natural gas, electric power, and telephone lines and streets, viaducts, and

bridges). This concept of vulnerability hotspot analysis is illustrated in Figure 6-4, below.

Vulnerability Dynamics

A major challenge for emergency managers is to understand the processes by which communities

increase or decrease their hazard exposure, physical vulnerability, and social vulnerability. According to

economic theory, excessive hazard exposure and structural vulnerability arise from systemic complexities

that can be characterized as market failures such as inadequate information, barriers to market entry and

exit, and capital flow restrictions (Kunreuther, 1998; Lindell, et al., 1997). An ideal pattern of economic

178

development would be one in which risk area occupants purchase property on the basis of adequate

information about hazard exposure and structural vulnerability. Moreover, they would locate only where

it was economically advantageous in the long term as well as in the short term, and would diversify their

assets over other locations and other forms of financial (e.g., savings accounts, insurance, stocks/bonds)

and social (e.g., extended family) recovery assistance. Finally, risk area occupants would adopt hazard

adjustments to limit their losses if a disaster were to strike. These adjustments would include hazard

mitigation (e.g., land use practices and building construction practices), emergency preparedness practices

(e.g., detection and warning systems), and recovery preparedness practices (e.g., diversified investments

and hazard insurance) to avoid casualties and property damage.

Table 6-5. Indicators of Social Vulnerability. Vulnerable Groups Vulnerability Indicators

Female headed households Percent female headed households

Elderly Percent individuals over 65

Percent of elderly households

Low income/high poverty Percent of households below poverty level

Percent of households below HUD standards

Renters Percent of households residing in rental housing

Percent of households residing in rental housing by type of dwelling units

Ethnic/racial/language minorities Percent of individual from Black, Hispanic, and other minorities

Percent of non-English speakers

Children/youth Percent of population in selected age groupings

Percent of households with dependency ratios above a specified level

Social vulnerability hot spot analysis Areas with combined social vulnerabilities

Actual patterns of development are significantly different from the ideal. In many cases, there is

migration to hazard-prone areas because of beneficial land uses for agriculture, transportation, and

recreation (i.e., people are “pulled in”, Bolin & Bolton, 1986). This is compounded by a lack of

accountability for investment decisions. Developers are at risk for only a short period of time before they

pass an investment on to others (homeowners, insurers, mortgage holders) who will ultimately experience

the disaster impact. Such transactions can occur because many risk area residents are new arrivals who

are unaware of the hazard. Even long-term residents of risk areas sometimes have little or no information

about hazards and adjustments to those hazards because such information is suppressed by those with a

major stake in the community’s economic development (Meltsner, 1979). Even when there is local

knowledge about hazards, there often is a lack of hazard intrusiveness because events that are not recent

or frequent tend not to be thought about or discussed (Lindell & Prater, 2000). Moreover, many people

ignore low probability events, think of them as occurring far in the future, or have an optimistic bias that

the negative consequences of these events will not happen to them (Weinstein, 1980). In particular,

politicians tend to ignore consequences that they expect to occur only after their term of office is over, so

only frequent, recent, or major impacts lead to increased adoption of community-wide hazard adjustments

such as land use controls or more stringent building codes. Even then, the window of opportunity for the

adoption of these adjustments is open only temporarily (Birkland, 1997; Prater & Lindell, 2000).

179

Figure 6-4. Disaster Impact Vulnerability Assessment Model.

Source: Prater, et al. (2004)

Increased hazard exposure also is caused by displacement from safer areas due to population

pressures (i.e., people are “pushed in”). When this occurs, the demographic distribution of risk tends to be

inequitable because geographical locations often are systematically related to their residents’

demographic characteristics—especially their (lack of) economic and political power to decrease hazard

vulnerability. This pattern is very common in developing countries such as Brazil, where favelas are

located in flood plains and on landslide-prone slopes because the residents cannot afford to purchase

homes in safer areas.

There also are problems in the adoption of effective hazard adjustments. One of these arises from

households’ and businesses’ concentration of hazard exposure (i.e., having physical and financial assets

located only in the risk area). Diversification is an effective way of avoiding concentration of hazard

vulnerability, but low-income households and small businesses often have so few physical or financial

assets that they cannot afford to locate some of them in safer areas. Hazard insurance is problematic

because it tends to suffer from adverse selection, which means that only those who are at the greatest risk

are likely to purchase it (Kunreuther, 1998). Moreover, the actions of one party can sometimes increase

the vulnerability of another. In floodplains, upstream development cuts down trees and replaces it with

hardscape, thus increasing the speed of rainfall runoff and downstream flooding. Technological protection

works such as dams and levees can offset such increases in hazard exposure, but many risk area occupants

overestimate the effectiveness of such hazard adjustments (Harding & Parker, 1974). This can cause

further development of floodplains and, thus, increased hazard exposure that exceeds the risk reduction

provided by the adjustment that was adopted.

Conducting HVA with HAZUS-MH

HAZUS-MH (HAZards US-Multi Hazard) is a computer program that models potential losses

from earthquakes, floods, and hurricane winds. HAZUS-MH uses GIS software to analyze and display

data on estimated structural damage and economic loss estimates for buildings and infrastructure. It also

provides estimates of the casualties resulting from earthquakes. HAZUS-MH can be used to conduct

analyses in support of hazard mitigation, emergency preparedness, and recovery preparedness planning.

Area with highest potential for

social impacts: Social vulnerability “hotspots”

High social and

governmental vulnerability

High hazard exposure

High structural and lifeline vulnerability

180

In addition, HAZUS-MH can be used to conduct rapid analyses in support of postimpact emergency

response and disaster recovery operations.

HAZUS-MH supports three levels of analysis. A Level 1 analysis uses national average data to

produce approximate results. Accordingly, a Level 1 analysis is best considered to be an initial screen that

identifies the communities at highest risk. A Level 2 analysis takes refined data and hazard maps provided

by the user to produce more accurate estimates. Input for a Level 2 analysis is obtained from local

emergency managers, urban and regional planners, and GIS professionals. A Level 3 analysis uses

community-specific parameters to produce the most accurate loss estimates. Input for a Level 3 analysis is

obtained from structural and geotechnical engineers, as well as other technical experts to examine threats

such as dam breaks and tsunamis.

Data input to HAZUS-MH is supported by the Inventory Collection Tool (InCAST), a Building

Inventory Tool (BIT), and Flood Information Tool (FIT). InCAST is a database that is designed to

support the management of local building data needed for Level 2 and Level 3 analyses. BIT supports the

importation of building data from large files (e.g., over 100,000 records from a tax assessor data file). FIT

allows users to transform flood data to the HAZUS flood model’s required format.

HAZUS-MH has separate models for earthquakes, floods, and hurricane winds. The earthquake

model accounts for ground motion and ground failure; the flood model accounts for flood frequency,

depth, and discharge velocity. The hurricane model accounts for wind pressure, missile damage, and rain.

Direct damage can be calculated for the general building stock, essential facilities, high potential

loss facilities, transportation facilities, and lifelines. Induced damage can be estimated for fire following,

hazmat release, and debris generation. Direct losses can be estimated for the cost of repair, income loss,

crop damage, casualties, shelter needs, and recovery needs. Indirect losses include supply shortages, sales

declines, opportunity costs, and economic losses. These impact modules are most complete for earthquake

(only the crop loss module is unavailable) and floods (only the fire following and casualties modules are

unavailable). The hurricane model has the fewest features (direct damage to the general building stock,

essential facilities and high potential loss facilities, induced damage from hazmat release, debris

generation, direct losses from cost of repair/replacement, shelter needs, and recovery needs).

HAZUS-MH can be used in multihazard analyses that provide average annualized loss and

probabilistic results from the three hazard models (earthquake, flood, and wind). HAZUS-MH can also

link to external models for blast, radiological, chemical, and biological hazards. Further information about

HAZUS-MH is available from the HAZUS Resource Center located at www.fema.gov/hazus. This source

includes data on HAZUS-MH hardware and software requirements, manuals, case studies, and contacts

for membership in users groups.

The Community Vulnerability Assessment Tool

The NOAA Coastal Services Center also provides guidance on conducting community

vulnerability assessment on their website www.csc.noaa.gov/products/nchaz/startup.htm. First, the hazard

identification module asks the user to list the community’s hazards and rate each of them on a 1-5 scale in

terms of its frequency and area impact (these scores are added together) and potential damage magnitude

(this score is multiplied by the sum of the previous two factors) to produce a total score for each hazard.

Second, the hazard analysis module recommends using risk maps to identify areas within the user’s

jurisdiction that vary in their degree of risk from each hazard and to rate each of these areas on a 1-5 scale

in terms of its risk. Third, the critical facilities analysis module prompts the user to identify the

181

community’s critical facilities and develop an inventory listing each facility’s type, name, address, and

other important information. This module concludes by identifying facilities located in the highest risk

areas (which were identified in the hazard analysis module). Fourth, the societal analysis module is

designed to identify neighborhoods requiring special consideration because they are high in social

vulnerability (refer to Table 6-5 for sample criteria). This module continues with an analysis of the

neighborhoods located in the highest risk areas (which were identified in the hazard analysis module).

The module concludes with an inventory of the special consideration/high risk locations listing the

neighborhood name, location, and other important information such as the number of households and the

nature of their needs. Fifth, the economic analysis module identifies the principal economic sectors in the

community and their geographic locations. This module continues with an analysis of the identification of

the businesses that are located in the highest risk areas (which were identified in the hazard analysis

module). The module concludes with an inventory of economic sector/high risk locations listing the

economic sector name, location, and other important information such as the number of businesses,

individual businesses having the most employees, and the nature of these businesses’ vulnerabilities (e.g.,

the length of time they can operate without infrastructure). Sixth, the environmental analysis module

prompts users to identify locations of secondary hazards (e.g., hazardous materials) and key

environmental resources (e.g., wetlands). This module continues with an analysis of the identification of

the secondary hazards and key environmental resources that are located in the highest risk areas

(identified in the hazard analysis module). The module concludes with an inventory of locations at which

secondary hazards located in the primary hazard areas could affect key environmental resources by their

proximity. The module concludes with an inventory of secondary hazard sites and their addition to the

critical facilities list. Seventh, the mitigation opportunities analysis identifies undeveloped land in high

hazard areas to support development of mitigation strategies focused on new development. It also

encourages review of the community’s status in the National Flood Insurance Program (NFIP).

The Coastal Services Center website also contains an overview of LIDAR (LIght Detection And

Ranging) beach mapping to obtain highly accurate elevation data. It also describes a damage assessment

tool for rapid postimpact that allows personnel to retrieve parcel data in a GIS database and integrate it

with FEMA damage assessment forms. Finally, the website also explains how remote sensing can be used

to provide broad area views of the impact area after a disaster strikes.

Analyzing and Disseminating Hazard/Vulnerability Data

The widespread availability of powerful desktop computers provides an important method for

conducting hazard/vulnerability analyses in identifying areas at risk (Dash, 1997; Griffith, 1986; Berke,

Larsen and Ruch, 1984) and projecting the damages resulting from a major incident (French, 1986;

Haney, 1986; Scawthorne, 1986). To accomplish these tasks, emergency managers can use software such

as GIS (Environmental Systems Research Institute, 2000), CAMEO (National Safety Council, 1995),

ALOHA (FEMA, no date, a), and HAZUS (National Institute for Building Sciences, 1998).

In addition, desktop computers provide emergency managers with a powerful tool for obtaining

hazard/vulnerability data from the websites of many government agencies, university research centers,

private sector organizations, and NGOs. In particular, the Internet has become an important means of

obtaining the data needed for conducting HVAs and disseminating their results—a technological

development that is important for three reasons. First, federal and state agencies have generated many

hazard analysis documents, maps, and databases that already are in digital form and are available to put

182

onto websites. Second, SEMA websites can be linked electronically to other organizations’ websites, thus

allowing users to immediately access additional hazard analysis information that might otherwise take

months to obtain if they were to request it in paper copy. Third, hazard analyses disseminated over the

Internet can be updated frequently and, by avoiding the printing costs associated with hundreds of paper

copies, can be disseminated less expensively.

Analyzing Hazard/Vulnerability Data

Despite the great promise of computers in analyzing and disseminating hazard/vulnerability data,

there is little documentation of the extent to which these tools are actually being used. Some indication of

the degree to which progress remains to be made can be seen in Lindell and Perry’s (2001) data from

LEPC Chairs in Illinois, Indiana, and Michigan indicating that only 59% of the LEPCs had calculated

VZs around their communities’ hazmat facilities. Of those who had calculated VZs, only 36% had used

computer models such as CAMEO (National Safety Council, 1995) or ARCHIE (Federal Emergency

Management Agency, no date, a) to perform the analyses. Thus, only a small fraction of the LEPCs used

computer-based methods to calculate VZs. In addition, there were differences among types of computer

use, with some LEPCs using computerized databases more extensively for the management of data on

chemical hazards (i.e., chemical inventories at local plants) and community emergency response

resources than did other jurisdictions. Thus, these data, though very limited in scope, indicate local

emergency management agencies have a long way to go in using emergency management information

technology to its fullest advantage.

SEMA Dissemination of Hazard/Vulnerability Data Via Websites

A recent examination of SEMAs’ websites revealed that most SEMAs provide some hazard

analysis information (Hwang, et al., 2001). The most commonly addressed hazards on SEMA websites

were hurricane, earthquake, flood, fire, tornado, hazardous material, storm, terrorism, drought, and

radiological material (see Table 6-6). This list includes some of the most significant hazards, as indicated

by the fact that of the 468 Presidential Disaster Declarations between January, 1992 and September 1999,

172 were for storms, 170 for floods, 58 were for tornadoes, 37 were for hurricanes, 17 were for blizzards,

5 each were for fires and earthquakes, and 4 were for landslides. However, many states that are vulnerable

to these hazards failed to address them and there are other hazards that were not addressed on SEMAs’

websites that also should receive attention.

Table 6-6. Hazard Agents and their Frequency of Mention on SEMA Websites

Hazard Agent Records Percent Hazard Agent Records Percent

Hurricane 50 15.6% Heat 8 2.5%

Earthquake 40 12.5% Structure failure 6 1.9%

Flood 33 10.3% Tsunami 4 1.2%

Fire 29 9.0% Landslide 3 0.9%

Tornado 29 9.0% Explosion 3 0.9%

Hazardous material 27 8.4% Avalanche 2 0.6%

Storm 22 6.9% Erosion 2 0.6%

Terrorism 15 4.7% Blight 1 0.3%

Drought 14 4.4% Freeze 1 0.3%

Radiological material 13 4.0% Meteor 1 0.3%

General 9 2.8% Pollution 1 0.3%

Volcano 8 2.5% Total 321 100.00

Source: Hwang, et al. (2001).

183

This study also identified 174 linkages from SEMA websites to secondary sources. Table 6-7

shows the websites for FEMA (28 links), National Oceanographic and Atmospheric Administration (18

links) and US Geological Survey (13 links) received the largest number of links. Another 19 sites

received between two and five links from SEMA websites and just over 40% of the secondary sources

identified in the study received links from only a single SEMA website. Taken together, the data from

Tables 6-6 and 6-7 indicate that there is a very erratic pattern to the dissemination of hazard analysis

information on SEMA websites.

Table 6-7. Percentage of all Links Accounted for by Each Secondary Source.

Secondary Source Web Address (URL) Percent

Source: Hwang, et al. (2001).

Federal Emergency Management Administration www.fema.gov 16.09

National Oceanographic and Atmospheric Administration www.noaa.gov 10.34

US Geological Survey www.usgs.gov 7.47

American Red Cross www.redcross.org 2.87

US Department of Health and Human Services www.hhs.gov 1.72

US Forest Service www.fs.fed.us 1.72

Weather Underground www.wunderground.com 1.72

Centers for Disease Control and Prevention, www.cdc.gov 1.15

Department of Transportation hazmat.dot.gov 1.15

Environmental Protection Agency www.epa.gov 1.15

Florida Division of Forestry fl-dof.com 1.15

Institute of Global Environment and Society grads.iges.org 1.15

National Drought Mitigation Center drought.unl.edu 1.15

National Fire Protection Association www.nfpa.org 1.15

New England States Emergency Consortium www.nesec.org 1.15

North Carolina Emergency Management Agency www.dem.dcc.state.nc.us 1.15

US Nuclear Regulatory Commission www.nrc.gov 1.15

Oklahoma Mesonetwork www.mesonet.ou.edu 1.15

Stormdisplay.com www.stormdisplay.com 1.15

Tornado Project www.tornadoproject.com 1.15

University of Illinois Dept. of Atmospheric Sciences www.atmos.uiuc.edu 1.15

Weather Channel www.weather.com 1.15

Total 59.20

Curiously, there was not a very strong correspondence between states’ hazard exposure and their

SEMA websites’ dissemination of information about those hazards. Specifically, analysis of six highly

probable or highly damaging hazards (storm, flood, tornado, hurricane, earthquake, and landslide) shows

substantial variation across states and hazards in the degree to which hazard analysis information on

SEMA websites corresponded to those states’ hazard vulnerability and recent disaster experience. Table

6-8 indicates most SEMAs failed even to address many common hazard agents on their websites and less

common hazards such as tsunami, structure failures, landslides, and avalanche are generally neglected

even though they also have the potential for significant impacts. Some plausible explanations for the

inadequate coverage of these hazard agents in SEMA websites are addressed below.

Some EMAs may lack the resources to deliver information through the Internet. Hwang and his

colleagues (2001) found two SEMAs had no websites, but this is likely to be a problem affecting other

EMAs as well because Drabek (1991d) found insufficient technical staff, as well as hardware and

184

software limitations, are impediments to emergency management agency adoption of emergency

management information technology. Drabek’s findings are consistent with the data from Hwang, et al

(2001), which indicate SEMAs with little or no hazard information on their websites tend to be those

from rural states with small budgets. Thus, they have few financial resources for purchasing computer

hardware and software and either hiring computer support staff or contracting with outside organizations

for website development. One possible way of overcoming the problem of staff limitations would be for

national- or state-level professional organizations such as NEMA or IAEM to establish basic guidance for

the development and maintenance of hazard analysis websites. The availability of hazard analysis

information on SEMA websites could be significantly improved if one of these professional organizations

were to establish a list of qualified contractors that could work with the states to upgrade their websites. A

longer-term solution would be for universities offering programs in emergency management to emphasize

emergency management information technology within their curricula so that state and local agencies

would have a pool of applicants who are technically qualified in this area.

Table 6-8. Relationship between Hazard Exposure and Website Records.

Hazard Number of exposed

states

Number of states with major disaster

declarations

Number of states with hazard analysis

records

Correlation of exposure with

hazard analysis records

Correlation of disasters with hazard

analysis records

Storm 28 47 15 -0.12 -0.04

Flood 25 46 18 0.03 -0.08

Tornado 23 24 17 0.22 0.17

Hurricane 13 17 12 0.53** 0.85**

Earthquake 16 2 16 0.38* 0.16

Landslide 22 3 1 -0.12 -0.03

Source: Lindell, et al. (2002). *p < .05; ** p < .01

Some EMAs may think a hazard analysis website is unnecessary because people already know

about most common hazards such as storms and floods. This rationale is entirely inappropriate because

people frequently have inaccurate beliefs about hazards, misjudge their personal vulnerability, and lack

information about methods of protecting themselves (Lindell & Perry, 2000; Lindell & Barnes, 1986;

Slovic, et al., 1974; Whitney, et al., 2004). Indeed, it is precisely because local emergency managers and

the public lack accurate hazard information that federal agencies such as FEMA and USGS disseminate

this information.

Some EMAs may not believe that the Internet is an effective method of disseminating hazard

analysis information. There is some validity to this belief because Internet access, though extensive, is far

from universal for local emergency managers and the public. Nonetheless, Internet access is rapidly

becoming more widespread and the number of people who can be reached by this communications

medium is becoming increasingly large. This is especially true among younger population segments that

have spent their entire lives using computers. Moreover, cost-effectiveness is a major incentive for EMAs

to increase their use of the Internet. EMAs bear essentially no reproduction or distribution costs because

users pay their Internet Service Providers for access to the website and there is a cost of printing only if

the users decide hard copy is needed. Finally, electronic dissemination is advantageous for local

emergency managers because they can “cut and paste” portions of the hazard analysis information from

the SEMA website into their own local hazard/vulnerability analyses.

185

Some EMAs may provide little hazard information on their websites because they already

distribute this information through other media. This rationale has some validity because, for example,

many states that are vulnerable to hurricanes distribute brochures containing maps of risk areas and

hurricane survival tips. However, this explanation ignores the fact that hazard analysis information

developed for distribution through other channels can be adapted quite readily and inexpensively to a

website. Similarly, hazard analysis information received from other sources (e.g., Red Cross disaster

preparedness brochures) is becoming available on those organizations’ websites and EMAs only need to

establish a link to that information. A cost-effective method for disseminating hazard analysis information

would be to provide detailed information on the state website and publicize the address of this website

through other media such as one-page maps and leaflets, newspaper articles, and radio and television

public service announcement.

EMAs might overlook hazard agents with which they have little or no recent disaster experience.

This does not appear to be the case because more commonly experienced hazards such as storms receive

less attention than more dramatic agents such as hurricanes and earthquakes. The problem seems to be

that EMAs simply are not providing enough hazard information themselves, or linking to others who do

provide that information.

Conclusions. Most EMAs are underutilizing an important channel for delivering hazard analysis

information to local emergency managers and the public. As mentioned earlier, it is likely that the

demand for Web-based hazard information will increase over the time. According to FEMA

(www.fema.gov), Internet users visited FEMA’s hurricane-related websites more than 1.25 million times

in the week after Hurricane Bertha hit the US in 1996. Recognized authorities, such as emergency

management agencies and major research centers, need to play an important role in disseminating hazard-

related information (Fischer, 1998). SEMA websites can play an important role in meeting this need by

helping local emergency planners collect needed information rapidly and easily; LEMA websites can add

value by providing information that is specific to their jurisdictions. In both cases, EMA websites can

help local residents recognize their exposure to natural and technological hazards. This can motivate them

to adopt hazard adjustments that would reduce their vulnerability to these threats.

In the course of examining SEMA websites, Hwang, et al. (2001) observed a number of recurring

deficiencies. Some of these related to the content of the hazard analysis information, while other

deficiencies arose from the way in which the web pages were structured. Table 6-9 lists 18

recommendations for improving the presentation of hazard analysis information on EMA websites. These

recommendations are specifically directed toward improving the usability of those portions of a website

accessed by the public, but do not address many of the general issues of website design that are covered in

great detail elsewhere (e.g., Nielsen, 2000).

Local Utilization of Hazard Analysis Websites

In addition to knowing what is on SEMA websites, it is important to be aware of the sources local

emergency managers use to obtain their hazard analysis information and the extent to which local

government agencies utilize different emergency management information technology applications. A

recent survey of Texas emergency managers and land use planners found over one-third of the responding

local government agencies use few sources of hazard analysis information and that nearly one-third use

no hazard analysis information at all (Lindell, et al., 2002). Of those who do use hazard analysis

information, two-thirds of the materials used are printed documents and only one-third of the materials

186

used were obtained from the Internet. The relative importance of different websites can be seen in Table

6-10, which shows respondents’ ratings of the extent to which they used each agency’s website.

Table 6-9. Recommendations for Hazard Analysis Websites. • Recognize that your website will be considered an authoritative source by users, which means that you have a

responsibility for ensuring the accuracy of the information you provide.

• Recognize that a website transmits information directly to the public without passing through the usual print (newspapers and magazines) and electronic (television and radio) news media. This reduces message distortion, but requires a high standard of clarity and organization.

• Coordinate the information provided through your website with the guidelines in the Red Cross’s Talking about Disaster: Guide for Standard Messages (www.redcross.org/disaster/safety/guide.html). This guide—a joint effort of the Red Cross, FEMA, National Weather Service, and other organizations—is designed to standardize disaster information provided by authoritative sources so they do not issue conflicting messages.

• Address all significant hazards to which your jurisdiction is vulnerable, but also provide information about the likelihood of major events so that people can judge which ones deserve the greatest priority. Display this information on maps to show where these hazards are most likely to occur.

• Provide non-technical explanations of the physical aspects of hazards (e.g., how hurricanes form and how wind, rain, and surge behavior affects the built environment) to help people understand what will happen and why it will happen.

• Provide information about hazard impacts so users can understand how a disaster will affect their communities. Important information about hazard impact includes the speed of onset, scope and duration of impact, and the magnitude of different types of consequences such as casualties (deaths and injuries), property damage, and economic impacts (disruption to industrial, commercial, agricultural, and governmental activity).

• Provide information that personalizes the potential consequences for the viewers. These include the cumulative probability of being affected during the different periods of time that a person would live in a risk area (e.g., 10, 20, and 30 year intervals).

• Provide a website index or table of contents to help users find needed information quickly and effectively. If your website is very large, provide a search engine for locating topics of interest.

• Provide links to other emergency-related information such as situation reports about current incidents and information available from other local, state and federal emergency-related organizations.

• Provide information about hazard adjustments—actions people can take to protect themselves, their families, and their property from environmental hazards. Explain how effective these actions are in protecting persons and property, whether they are useful for other purposes, how much they cost, what knowledge, skills, tools, and other resources they require. Describe the specific steps required to perform any unfamiliar actions.

• Organize any links to other websites by referring the user to the page that addresses a specific topic, not to an organization’s home page. For example, nonspecific links to FEMA and Red Cross home pages are of little help because these sites contain thousands of pages of information.

• Keep text clear and succinct. Use suitably large and legible fonts and simple color design schemes so the information is easy to read.

• Provide enough figures and pictures to explain the text and maintain interest, but avoid overuse of pictorial materials because this can cause the information to download so slowly that users become frustrated and abandon their information search.

• Ensure your website is compliant with the Americans with Disabilities Act, which requires pictures and graphs to be described in words and that your site be navigable without a mouse.

• Make it easy for viewers to download information by attaching documents in PDF or major word processor (e.g., Word Perfect

® or MS-Word

® ) format.

• Include contact information with postal and e-mail addresses, telephone numbers, and fax numbers of persons from whom users can obtain additional information or to whom they can offer suggestions.

• Publicize your website by seeking cross-links with news media websites, by notifying other state and local organizations, and listing the website URL on hazard awareness materials such as hurricane maps, refrigerator magnets, or preparedness checklists.

• Verify that your server can handle elevated demand for information during an emergency. Ensure that high demand will not degrade the functionality of host organization’s internal emergency management information system or of any emergency management information network residing on that server.

These data indicate FEMA, SEMAs, and EPA are the most popular sources of hazard analysis

information, but the sources differ in the ways that they are used. The respondents used slightly more

print than electronic resources from EPA and noticeably more print than electronic resources from

FEMA. However, the preference is reversed for the other federal agencies—NOAA/NWS, and USGS—

that are accessed only in electronic form. One plausible explanation for this pattern of results is that

FEMA disseminates a large volume of hazard analysis information in print and targets much of this

information for state and local government agencies. In addition, FEMA has an extensive website that

187

contains information on all phases of emergency management and also real-time emergency and disaster

information. By contrast, NOAA and NWS tend to be better known among local government agencies for

real-time weather and flood forecasts than for long-term hazard analysis information. These uses,

obviously, would tend to promote electronic rather than print access.

Table 6-10. Mean Extent of Internet Site Access by Local Government Agencies.

Website Extent of

use Website

Extent of use

Federal Emergency Management Agency 2.90 US Geological Survey 1.91

Nat. Ocean. and Atmos. Admin. 2.92 Int. Assn. of Emergency Managers 1.89

State Division of Emergency Management 2.84 US Department of Agriculture 1.86

State Department of Health 2.61 Emer. Plan. Info. Exchange 1.85

US Environmental Protection Agency 2.57 Bureau of the Census 1.83

State Nat. Res. and Cons. Comm. 2.55 Nat. Emer. Mgmt Assn. 1.74

Emer. Mgrs. Weather Information Network 2.23 State Department of Agriculture 1.68

US Department of Transportation 2.20 Salvation Army 1.57

National Hurricane Center 2.07 Small Business Administration 1.41

American Red Cross 1.97 US Nuclear Regulatory Commission 1.27

Source: Hwang, et al. (2001).

By contrast, USGS tends to be better known among local government agencies for long-term

hazard analysis information than for real-time forecasts. This might seem to favor print over electronic

access, but USGS is not as active as FEMA in disseminating its reports to local government agencies.

Consequently, its reports tend to be rather difficult to access in print, but much of the essential

information is available on the agency’s website.

Given the increasing popularity of the Internet, with its advantages of low cost dissemination and

ease of frequent updates, it would seem advisable for federal and state agencies to increase their reliance

on the Internet to disseminate hazard analysis information. However, the data from this survey indicate

that hazard analysis information should not be distributed exclusively through electronic channels because

there is a significant portion of the audience for this information that still uses print media. Further

research is needed to determine whether local government agencies prefer to use paper documents even

though they could just as easily access them electronically; whether they lack reliable hardware, software,

or trained personnel to access these documents on the Internet; or whether the information on the Internet

is too poorly organized for them to make effective use of it.

Other data from Lindell, et al. (2002) reveal a hierarchy of computer applications ranging from

the most popular, word processing, to the least popular, hazard modeling (Table 6-11). Moreover, these

computer applications can be divided into three categories. The most frequently used applications are

word processing, email, Internet connection, databases, and spreadsheets. The second category is

graphics, desktop publishing, project accounting, presentation graphics, and statistical analysis. The least

frequently used applications are GIS and mapping, CAMEO, Web page design, and hazard modeling.

Interestingly, agencies become increasingly involved with all members of one computer

application type as soon as they begin to use one member of that type. For example, agencies do not

appear to use applications in Category 3 until after they have begun to use Categories 1 and 2 extensively.

This finding is consistent with the notion of “training transfer” (Ford & Schmitt, 2000) in which the

mastery of less difficult technologies provides the basis for learning how to use more complex

technologies. This suggests state emergency management agencies can increase the utilization of more

188

complex computer applications such as GIS and hazard modeling by first ensuring that local government

agencies have mastered more basic applications such as databases and spreadsheets. In addition, the

statistically significant differences between local emergency managers and land use planners in the

utilization of GIS, mapping, and website design suggest emergency managers should be encouraged to

seek partnerships with land use planners to obtain the products of these technologies. Indeed, the fact that

land use planners have consistently greater experience with all forms of information technology suggests

that such partnerships could be beneficial in many ways.

Table 6-11. Frequency of Computer Application Use by Emergency Management Coordinators and Land Use Planners.

Type Of Computer Application

Emergency Management Coordinators

Land Use Planners

Statistical Significance Of Difference

1. Word processing 4.23 4.85 t89 = 2.31, p < .05

2. Email 4.05 5.00 t90 = 3.14, p < .01

3. Internet connection 3.70 4.42 t88 = 2.10, p < .05

4. Databases 3.56 4.35 t88 = 2.41, p < .05

5. Spreadsheets 3.44 4.27 t88 = 2.58, p < .05

6. Graphics 2.83 4.08 t88 = 3.61, p < .001

7. Desktop publishing 2.74 3.20 t88 = 1.25, ns

8. Project accounting 2.49 3.04 t85 = 1.46, ns

9. Presentation 2.25 3.46 t88 = 3.81, p < .001

10. Statistical analysis 2.24 2.81 t87 = 1.76, ns

11. GIS and mapping 1.98 3.54 t86 = 4.17, p < .001

12. CAMEO 1.96 1.97 t89 = 0.27, ns

13. Webpage design 1.77 2.46 t88 = 2.39, p < .05

14. Hazard modeling 1.75 1.75 t86 = 0.00, ns

Source: Lindell, et al. (2002).

In summary, these data suggest that federal and state agencies should provide increasing amounts

of hazard analysis information on their websites, but that the transition from print to electronic media

should not exceed the ability of local government agencies to access the Internet. This is particularly

likely to be a problem for smaller and poorer jurisdictions (Quarantelli, 1997) Moreover, federal and state

agencies should consider facilitating local emergency managers’ access to hazard analysis information by

utilizing the computer applications with which their target audience is most familiar. At present, these are

databases, spreadsheets, and Internet connections to agency websites. Local emergency managers are

more likely to make use of sophisticated hazard modeling applications such as ALOHA (Federal

Emergency Management Agency, no date, a) and HAZUS (National Institute of Building Sciences. 1998)

if they have successfully mastered more basic computer applications or if they have partnered with land

use planners in their communities who have mastered the advanced computer applications.

Assessing Risk

A major problem in applying the methods of HVA is risk and uncertainty. FEMA’s (1997) Multi-

Hazard Identification and Risk Analysis, together with information from state and federal agency

websites, can be used to identify the hazard agents to which a community is exposed. This is a good start,

but it seems insufficient. Logic suggests that an emergency manager should consider the probability of a

disaster rather than just the possibility that it could occur. After all, an event that has a 50% chance of

occurrence should be given more attention than one that has only a 1% chance of occurrence.

189

Unfortunately, there are many difficulties in analyzing risk (Frosdick, 1997). The desired probabilities

frequently have not been calculated (e.g., the probability of a hazardous materials transportation accident

in a given jurisdiction), are highly uncertain (e.g., the probability of a Category 5 hurricane strike is

difficult to estimate because the 150 year historical record is too short), are unreliable (e.g., the true

probability of a major flood is often higher than its nominal value because upstream development is

increasing the probabilities to an unknown extent), or are not available for disclosure even if they have

been calculated (e.g., the probability of a release from a toxic chemical plant is information the owner

won’t share, in part for security reasons).

Moreover, Figure 6-1 indicates that even if one knew the probability of an event, there is

significant uncertainty about the magnitude of its physical impacts. That is, knowledge that there is a 20%

probability of a hurricane strike in the coming year is of limited use because different hurricane categories

have substantially different impacts and, indeed, disaster impacts can vary significantly within hurricane

categories. In general, the number of casualties from a hurricane depends not only on storm

characteristics (including forward movement speed and track stability), but also on the size and

distribution of the population at risk (which is greater on holiday weekends than midweek during the fall)

and authorities’ ability to evacuate the risk area. Similarly, the amount of damage to structures depends

not only on the wind speed and surge depth (which can be estimated from a hurricane’s Saffir-Simpson

category), but also on the amount of inland flooding (which is unrelated to its Saffir-Simpson category).

For example, Tropical Storm Allison had a devastating impact on Houston because it stalled over the city

and rained for days even though it never reached hurricane wind speed. Finally, Figure 6-1 indicates that

even if one knew the probability of an event’s physical impacts, there is significant uncertainty about the

magnitude of its social impacts. Some communities recover rapidly from disaster impact whereas

others—sometimes ones with less damage and fewer casualties—never recover.

In summary, local emergency managers cannot currently obtain precise information about their

communities’ hazard vulnerability and such information is unlikely to be available in the near future. This

might seem to be a very pessimistic view of the usefulness of HVAs but, in fact, it is not. Rather, it

simply recognizes the current limitations in the state of HVA technology and (quite probably) inherent

limitations in the resources local emergency managers can devote to this activity.

However, it is important to recognize that emergency managers can still do their jobs effectively

even if they lack extremely precise HVA data. This is because managers in general, and emergency

managers in particular, only need enough information to make decisions about the allocation of the

resources under their control or to justify obtaining more resources to do their jobs better. Thus,

emergency managers need enough HVA data to decide how to allocate their existing resources to

different emergency management activities such as hazard mitigation, emergency response preparedness,

and disaster recovery preparedness. Indeed, HVA data alone would be insufficient to make these resource

allocation decisions because different hazard agents to which a community is exposed impose similar

demands. Consequently, investments in those emergency management activities have payoffs for many

different hazards. For example, all hazards require a capability for incident management actions such as

establishing coordination among organizations that are normally independent of each other and most

hazards require a capability for population protection actions such as evacuating local residents. Thus,

investments in the community’s capability for evacuation might reduce vulnerability more than

investments in hazard mitigation activities (which generally affect vulnerability to only a single hazard).

190

If so, the HVA only needs to be specific enough for emergency managers to persuade themselves that the

allocation of resources among emergency management activities is appropriate.

Emergency managers also need HVA data to lobby for increased budget allocations to emergency

management rather than to other community activities such as education, health care, and transportation.

In this case, the HVA should be specific enough for emergency managers to successfully persuade other

stakeholders that an increase in the emergency management budget will benefit the community more than

allocating that money to some other activity. This is not a trivial problem because, as noted in previous

chapters, emergency managers must lobby for money to solve future problems, which is inherently more

difficult than lobbying for money to solve current problems. In summary, emergency managers need to

allocate just enough time and money to HVA to achieve the precision needed to achieve their budget

allocation and budget augmentation objectives.

Case Study: Hurricane HVA for the Texas Gulf Coast

For many years, the Texas A&M University Hazard Reduction & Recovery Center conducted

hurricane HVAs for the Texas Governor’s Division of Emergency Management. The results of these

HVAs have been published in the form of hurricane storm atlases ((Texas Governor’s Division of

Emergency Management, 2004), contingency planning guides (Texas Governor’s Division of Emergency

Management, 2002a, 2002b), and evacuation maps. These hurricane HVAs include the analyses of hazard

exposure and physical vulnerability described in previous sections of this chapter.

Each hurricane HVA begins with data generated by the National Hurricane Center’s Sea, Lake,

and Overland Surges from Hurricanes (SLOSH— Jelesnianski, Chen & Shaeffer, 1992) model. SLOSH

uses data on coastal topography and bathymetry (the topography of the sea floor) and storm parameters

(Saffir-Simpson Category, forward movement speed, and storm track) as input to a complex computer

model of hurricane behavior. SLOSH output comprises tables of surge height over time for a series of

locations spaced approximately equidistantly along the coast. The data in these tables represent average

surge heights over a large number of simulated hurricane scenarios within those storm parameters.

Analysts import the SLOSH output data into a GIS that already contains data on coastal

topography, political boundaries, physical features (e.g., rivers and lakes), census data, and facility

locations. The SLOSH data are first overlaid onto the topographical data layer to compute surge depths.

This calculation subtracts the ground elevation (in feet above sea level) from the surge height (also in feet

above sea level) to compute surge depth (in feet above the ground’s surface). Computation of the surge

depth data allows the analyst to create a map that identifies the expected surge inundation boundary and

plot surge depth contours. As will be discussed below, these maps allow emergency managers to identify

populations and facilities at risk from surge flooding. The storm surge data are also used to generate

hydrographs that plot the hourly surge heights (from 24 hours before to 12 hours after landfall) at key

coastal locations for hurricanes of varying Saffir-Simpson category, forward movement speed, and track

direction. This information allows local officials to identify locations where early flooding is likely to

interrupt evacuations.

The analysts also used Kaplan and DeMaria’s (1995) Inland Wind Decay Model to generate

contours of the one-minute sustained surface wind speeds at different distances from the coastline

(Lindell, Prater & Zhang, 2001). These sustained surface wind speed contours, which are mapped as a

function of hurricane category and forward movement speed, are translated into expected peak gusts by

191

applying wind gusting factors. These maps allow emergency managers to identify populations and

facilities at risk from wind damage.

Once the surge and wind contours have been generated, they are used to define exposure zones

that vary in their threat to facilities that provide essential services or key care services. Such facilities

include those performing communications, power generation, water/sewer treatment, transportation,

public health and safety, and postdisaster sheltering functions. In addition, the exposure zones are used to

identify facilities about which there are special concerns, especially industrial and hazardous materials

facilities. Each facility is located on a map and its position indicated by a facility code (indicating its

function). In addition, analysts prepare a table listing each of the facilities, its location defined by an

alpha-numeric code (e.g., H6) indicating its location on the map grid, its latitude and longitude, its

elevation, and the lowest hurricane category at which it is likely to be affected by surge flooding, waves

over 3 feet, surge over 6 feet, or dangerous winds.

Other data are collected on highway capacity, population size and distribution, and coastal

residents’ evacuation expectations (Lindell, et al., 2001) and integrated within an evacuation time

estimation model (Lindell & Prater, 2005; Lindell, Prater, Perry & Wu, 2002) to produce evacuation time

estimates for different hurricane categories (Lindell, Prater & Wu, 2002). Finally, data are collected on

inland counties’ evacuation guidance and transportation support for hurricane evacuees (Prater & Lindell,

2002). All of these data can be used to estimate the level of damage and casualties that would be expected

from hurricanes. For further information, see the Texas Governor’s Division of Emergency Management

website (www.txdps.state.tx.us/dem/pages/planning.htm).