Hazard, Risk and Vulnerability
nani370
Reading6.16.2.pdf
INTRODUCTION
Risk is an unavoidable part of life, affecting all people without exception, irrespective of geographic or socioeconomic limits. Each choice we make as individuals and as a society involves specific, often unknown, factors of risk, and full risk avoidance generally is impossible.
On the individual level, each person is primarily responsible for managing the risks he faces as he sees fit. For some risks, management may be obligatory, as with automobile speed limits and seatbelt usage. For other personal risks, such as those associated with many recreational sports, individuals are free to decide the degree to which they will reduce their risk exposure, such as wearing a ski helmet or other protective clothing. Similarly, the risk of dis- ease affects humans as individuals, and as such is generally managed by individuals. By employing risk reduction techniques for each life hazard, individuals effectively reduce their vulnerability to those hazard risks.
As a society or a nation, citizens collectively face risks from a range of large-scale hazards. Although these hazards usually result in fewer total injuries and fatalities over the course of each year than individu-
ally faced hazards, they are considered much more significant in that they have the potential to result in many deaths, injuries, or damages in a single event or series of events. In fact, some of these hazards are so great that, if they occurred, they would result in such devastation that the capacity of local response mecha- nisms would be overwhelmed. This, by definition, is a disaster. For these large-scale hazards, many of which were identified in Chapter 2, vulnerability is most effectively reduced by disaster management efforts collectively, as a society. For most, though not all of these hazards, it is the government’s responsibility to manage, or at least guide the management of, hazard risk reduction measures. And when these hazards do result in disaster, it is likewise the responsibility of governments to respond to them and aid in the fol- lowing recovery.
This text focuses on the management of interna- tional disasters, which are those events that over- whelm an individual nation or region’s ability to respond, thereby requiring the assistance of the inter- national body of response agencies. This chapter, therefore, will focus not upon individual, daily risks and vulnerabilities, but on the risks and vulnerabilities that apply to the large-scale hazards like those dis- cussed in Chapter 2.
3 Risk and Vulnerability
113
TWO COMPONENTS OF RISK
Chapter 1 defined risk as the interaction of a haz- ard’s consequences with its probability or likelihood. This is its definition in virtually all documents asso- ciated with risk management. Clearly defining the meaning of “risk” is important, because the term often carries markedly different meanings for different peo- ple (Jardine and Hrudey, 1997). One of the simplest and most common definitions of risk, preferred by many risk managers, is displayed by the equation stat- ing that risk is the likelihood of an event occurring multiplied by the consequence of that event, were it to occur.
RISK = LIKELIHOOD × CONSEQUENCE (Ansel and Wharton 1992)
LIKELIHOOD
“Likelihood” can be given as a probability or a fre- quency, whichever is appropriate for the analysis under consideration. Variants of this definition appear in virtually all risk management documents. “Fre- quency” refers to the number of times an event will occur within an established sample size over a specific period of time. Quite literally, it tells how frequently an event occurs. For instance, the frequency of auto accident deaths in the United States averages around 1 per 300 million miles driven (Wilson, 1979).
In contrast to frequency, “probability” refers to sin- gle-event scenarios. Its value is expressed as a number between 0 and 1, with 0 signifying a zero chance of occurrence, and 1 signifying certain occurrence. Using the auto accident example, in which the fre- quency of death is 1 per 300 million miles driven, we can say that the probability of a random person in the United States dying in a car accident equals .000001 if he was to drive 300 miles.
Disaster managers use this formula for risk to determine the likelihood and the consequences of each hazard according to a standardized method of mea- surement. The identified hazard risks thus can be
compared to each other and, therefore, ranked accord- ing to severity. (If risks were analyzed and described using different methods and/or terms of reference, it would be very difficult to accurately compare them later in the hazards risk management process.)
This ranking of risks, or “risk evaluation,” as it is often called, allows disaster managers to determine which treatment (mitigation and preparedness) op- tions are the most effective, most appropriate, and provide the most benefit per unit of cost. Not all risks are equally serious, and risk analysis can provide a clearer idea of these levels of seriousness.
Without exception governments have a limited amount of funds available to manage the risks they face. While the treatment of one hazard may be less expensive or more easily implemented than the treat- ment of another, cost and ease alone may not be valid reasons to choose a treatment option. Hazards that have great consequences (in terms of lives lost or injured or property damaged or destroyed) and/or occur with great frequency pose the greatest overall threat. Considering the limited funds, disaster man- agers generally should recommend first treating those risks that pose the greatest threat. Fiscal realities often drive this analytic approach, resulting in situations in which certain hazards in the community’s overall risk profile are mitigated, while others are not addressed at all.
The goal of risk analysis is to establish a standard and therefore comparable measurement of the likeli- hood and consequence of every identified hazard. The many ways by which likelihoods and consequences are determined are commonly divided into two cate- gories of analysis: quantitative analysis and qualita- tive analysis. Quantitative analysis uses mathematical and/or statistical data to derive numerical descrip- tions of risk. Qualitative analysis uses defined terms (words) to describe and categorize the likelihood and consequences of risk. Quantitative analysis gives a specific data point (dollars, probability, frequency, or number of injuries/fatalities, for example) while qual- itative analysis allows each qualifier to represent a range of possibilities. It is often cost and time prohib- itive, and often not necessary, to find the exact quanti-
114 Introduction to International Disaster Management
tative measures for the likelihood and consequence factors of risk. Qualitative measures, on the other hand, are much easier to determine, and require less time, money and, most importantly, expertise to con- duct. For this reason, it is often the preferred measure of choice. The following section provides a general explanation of how these two types of measurement apply to the likelihood and consequence components of risk.
Quantitative Representation of Likelihood
As previously stated, likelihood can be derived as either a frequency or a probability. A quantitative sys- tem of measurement exists for each. For frequency, this number indicates the number of times a hazard is expected to result in an actual event over a chosen time frame: 3 times per year, 1 time per decade, 10 times a week, and so on. Probability measures the same data, but the outcome is expressed as a measure between 0 and 1, or as a percentage between 0% and 100%, representing the chance of occurrence. For example, a 50-year flood has a 1/50 chance of occur- ring in any given year, or a probability of 2% or .02. An event that is expected to occur 2 times in the next 3 years has a .66 probability each year, or a 66% chance of occurrence.
Qualitative Representation of Likelihood
Likelihood can also be expressed using qualitative measurement, using words to describe the chance of occurrence. Each word or phrase has a designated range of possibilities attached to it. For instance, events could be described as follows:
● Certain. >99% chance of occurring in a given year (1 or more occurrences per year)
● Likely. 50–99% chance of occurring in a given year (1 occurrence every 1 to 2 years)
● Possible. 5–49% chance of occurring in a given year (1 occurrence every 2 to 20 years)
● Unlikely. 2–5% chance of occurring in a given year (1 occurrence every 20 to 50 years)
● Rare. 1–2% chance of occurring in a given year (1 occurrence every 50 to 100 years)
● Extremely rare. <1% chance of occurring in a given year (1 occurrence every 100 or more years)
Note that this is just one of a limitless range of qualitative terms and values that can be used to describe the likelihood component of risk. As long as all hazards are compared using the same range of qualitative values, the actual determination of likeli- hood ranges attached to each term does not necessar- ily matter (see Exhibit 3-1).
Chapter 3 Risk and Vulnerability 115
EXHIBIT 3-1 Qualitative Measurements: The Consideration of Risk Perception and Standardization
This chapter will later discuss the concept of risk perception. In brief, different people fear different hazards, for many different reasons. These differ- ences in perception can be based upon experience with previous instances of disasters, specific char- acteristics of the hazard, or many other combination of reasons. Even the word risk has different mean- ings to different people, ranging from “danger” to “adventure.”
Members of assembled disaster management teams are likely to be from different parts of the country or the world, and all have different percep-
tions of risk (regardless of whether they are able to recognize these differences). Such differences can be subtle, but they make a major difference in the risk analysis process.
Quantitative methods of assessing risk use exact measurements, and are therefore not very sus- ceptible to the effects of risk perception. A 50% likelihood of occurrence is the same to everyone, regardless of their convictions. Unfortunately, there rarely exists sufficient information to make definitive calculations of a hazard’s likelihood and consequence.
116 Introduction to International Disaster Management
The exact numeric form of measurement achieved through quantitative measurements is in- comparable. The value of qualitative assessments, however, lies in their ability to accommodate for an absence of exact figures and in their ease of use.
Unfortunately, risk perception causes different people to view the terms used in qualitative systems of measurement differently. For this reason, quali- tative assessments of risk must be based upon quan- titative ranges of possibilities or clear definitions. For example, imagine a qualitative system for mea- suring the consequences of earthquakes in a partic- ular city, in terms of lives lost and people injured. Now imagine that the disaster management team’s options are “None,” “Minor,” “Moderate,” “Major,” “Catastrophic.” One person on the team could consider 10 lives lost as minor. However, another team member considers the same number of fatalities as catastrophic. It depends on the perception of risk that each has developed over time.
This confusion is significantly alleviated when detailed definitions are used to determine the assig-
nation of consequence measurements for each haz- ard. Imagine the same scenario, using the following qualitative system of measurement (adapted from EMA, 2000):
1. None. No injuries or fatalities 2. Minor. Small number of injuries but no fatal-
ities; first aid treatment required 3. Moderate. Medical treatment needed but no
fatalities; some hospitalization 4. Major. Extensive injuries, significant hospi-
talization; fatalities 5. Catastrophic. Large number of severe
injuries; extended and large numbers requir- ing hospitalization; significant fatalities
This system of qualitative measurement, with defined terms, makes it more likely that people of different backgrounds or beliefs would choose the same characterization for the same magnitude of event. Were this system to include ranges of values, such as “1–20 fatalities” for “Major”, and “over 20 fatalities” for “Catastrophic,” the confusion could be alleviated even more.
CONSEQUENCE
The consequence component of risk describes the effects of the risk on humans, built structures, and the environment. There are generally three factors examined when determining the consequences of a disaster:
1. Deaths/fatalities (human) 2. Injuries (human) 3. Damages (cost, reported in currency, generally
US dollars for international comparison)
Although attempts have been made to convert all three factors into monetary amounts in order to derive a single number to quantify the consequences of a dis- aster, doing so can be controversial (How can one place a value on life?) and complex (Is a young life
worth more than an old life? By how much?). There- fore, it is often most appropriate and convenient to maintain a distinction between these three factors.
Categories of consequence can be further divided, and often are to better understand the total sum of all disaster consequences. Two of the most common dis- tinctions are direct and indirect losses, and tangible and intangible losses.
Direct losses, as described by Keith Smith in his book Environmental Hazards are “those first order consequences which occur immediately after an event, such as the deaths and damage caused by the throwing down of buildings in an earthquake” (1992). Examples of direct losses are:
● Fatalities ● Injuries (the prediction of injuries is often more
valuable than the prediction of fatalities, because
the injured will require a commitment of medical and other resources for treatment [UNDP, 1994])
● Cost of repair or replacement of damaged or destroyed public and private structures (build- ings, schools, bridges, roads, etc.)
● Relocation costs/temporary housing ● Loss of business inventory/agriculture ● Loss of income/rental costs ● Community response costs ● Cleanup costs
Indirect losses (also as described by Smith, 1992) may emerge much later and may be much less easy to attribute directly to the event. Examples of indirect losses include:
● Loss of income ● Input/output losses of businesses ● Reductions in business/personal spending (“rip-
ple effects”) ● Loss of institutional knowledge ● Mental illness ● Bereavement
Tangible losses are those for which a dollar value can be assigned. Generally, only tangible losses are included in the estimation of future events and the reporting of past events. Examples of tangible losses include:
● Cost of building repair/replacement ● Response costs ● Loss of inventory ● Loss of income
Intangible losses are those that cannot be expressed in universally accepted financial terms. This is the primary reason that human fatalities and human injuries are assessed as a separate category from the cost measurement of consequence in disaster management. These losses are almost never included in damage assessments or predictions. Examples of intangible losses include:
● Cultural losses ● Stress ● Mental illness
● Sentimental value ● Environmental losses (aesthetic value)
Although it is extremely rare for benefits to be included in the assessment of past disasters or the pre- diction of future ones, it is undeniable that they can exist in the aftermath of disaster events. Like losses, gains can be categorized as direct or indirect, tangible or intangible. Examples of tangible, intangible, direct, and indirect gains include:
● Decreases in future hazard risk by preventing rebuilding in hazard-prone areas
● New technologies used in reconstruction that results in an increase in quality of services
● Removal of old/unused/hazardous buildings ● Jobs created in reconstruction ● Greater public recognition of hazard risk ● Local/state/federal funds for reconstruction or
mitigation ● Environmental benefits (fertile soil from a vol-
cano, for example)
As with the likelihood component of risk, the con- sequences of risk can be described according to quan- titative or qualitative reporting methods. Quantitative representations of consequence vary according to deaths/fatalities, injuries, and damages:
● Deaths/fatalities. The specific number of people who perished in a past event or who would be expected to perish in a future event; for example, 55 people killed
● Injuries. The specific number of people who were injured in a past event or who would be expected to become injured in a future event. Can be expressed just as injuries, or divided into mild and serious; for example, 530 people injured, 56 seriously
● Damages. The assessed monetary amount of actual damages incurred in a past event, or the expected amount of damages expected to occur in a future event. Occasionally, this number includes insured losses as well; for example, $2 billion in damages, $980 million in insured losses
Chapter 3 Risk and Vulnerability 117
Qualitative Representation of Consequence
As with the qualitative representation of likeli- hood, words or phrases can be used to describe the effects of a past disaster or the anticipated effects of a future one. These measurements can be assigned to deaths, injuries, or costs (the qualitative measurement of fatalities and injuries often are combined). The fol- lowing is one example of a qualitative measurement system for injuries and deaths:
● Insignificant. No injuries or fatalities ● Minor. Small number of injuries but no fatalities;
first aid treatment required ● Moderate. Medical treatment needed but no
fatalities; some hospitalization ● Major. Extensive injuries, significant hospital-
ization; fatalities ● Catastrophic. Large number of fatalities and
severe injuries requiring hospitalization
Additional measures of consequence are possible, depending on the depth of analysis. These additional measures tend to require a great amount of resources, and are often not reported or cannot be derived from historical information. Examples include:
● Emergency operations. Can be measured as a ratio of responders to victims, examining the number of people who will be able to participate in disaster response (can include both official and unofficial responders) as a ratio of the number of people who will require assistance. This ratio will differ significantly depending on the hazard. For example, following a single tornado touch- down, there are usually many more responders than victims, but following a hurricane, there are almost always many more victims than respon- ders. This measure could include the first re- sponders from the community as well as the responders from the surrounding communities with which mutual aid agreements have been made. Emergency operations also can measure the mobilization costs and investment in pre- paredness capabilities. It can be difficult to measure the stress and overwork of the first
responders and their inability to carry out regular operations (fire suppression, regular police work, regular medical work).
● Social disruption (people made homeless/dis- placed). This can be a difficult measure because, unlike injuries or fatalities, people do not always report their status to municipal authorities (injuries and deaths are reported by the hospi- tals), and baseline figures do not always exist. It is also difficult to measure how many of those who are injured or displaced have alternative options for shelter or care. Measuring damage to community morale, social contacts and cohesion, and psychological distress can be very difficult, if not impossible.
● Disruption to economy. This can be measured in terms of the number of working days lost or the volume of production lost. The value of lost production is relatively easy to measure, while the lost opportunities, lost competitiveness, and damage to reputation can be much more difficult.
● Environmental impact. This can be measured in terms of the clean-up costs and the costs to repair and rehabilitate damaged areas. It is harder to measure in terms of the loss of aesthetics and public enjoyment, the consequences of a poorer environment, newly introduced health risks, and the risk of future disasters.
It does not matter what system is used for qualita- tive analysis, but the same qualitative analysis system must be used for all hazards being analyzed in order to compare risks. It may be necessary for disaster man- agers to create a qualitative system of measurement tailored to the country or community where they are working. Not all countries or communities are the same, and a small impact in one could be catastrophic to another, so the measurement system should accom- modate these differences. For example, a town of 500 people would be severely affected by a disaster that caused 10 deaths, while a city of 5 million may expe- rience that number of deaths in car accidents alone in a given week.
118 Introduction to International Disaster Management
Another benefit of creating an individualized sys- tem of qualitative analysis is the incorporation of the alternative measures of consequence (ratio of respon- ders to victims, people made homeless/displaced).
TRENDS
Both the likelihood and the consequences of cer- tain hazard risks can change considerably over time. Some hazards occur more or less frequently because of worldwide changes in climate patterns, while oth- ers change in frequency because of measures taken to prevent them or human movements into their path. These trends can be incremental or extreme and can occur suddenly or over centuries. Several short- term trends may even be part of a larger, long-term change.
Changes in Disaster Frequency
Changes in disaster frequency can be the result of both an increase in actual occurrences of a hazard and an increase in human activity where the hazard already exists. It is important to remember that a dis- aster is not the occurrence of a hazard, but the conse- quences of a hazard occurring. A tornado hitting an open field, for example, is not considered a disaster.
Changes in climate patterns, plate tectonics, or other natural systems can cause changes in the fre- quency of particular natural hazards, regardless of whether the cause of the changes are natural, like El Niño, or manmade, like global warming. Changes in frequency for technological or intentional hazards can be the result of many factors, such as increased or decreased regulation of industry and increases in international instability (terrorism).
Increases or decreases in human activity also can cause changes in disaster frequency. As populations move, they inevitably place themselves closer or far- ther from the range of effects from certain hazards. For instance, if a community begins to develop indus- trial facilities within a floodplain that was previously
unoccupied, or in an upstream watershed where the resultant runoff increases flood hazards downstream, it increases its risk to property from flooding.
Changes in Disaster Consequences
Like changes in disaster likelihoods, changes in consequences can be the result of changes in the attributes of the hazard itself or changes in human activity that place people and structures either at more or less risk.
Changes in the attributes of the hazard can occur as part of short- or long-term cycles, permanent changes in the natural processes if the hazard is natural, or changes in the nature of the technologies or tactics in the case of technological and intentional hazards. The consequences of natural hazards change only rarely independent of human activities. One example is El Niño events, with intense flooding increasing in some regions of the world and drought affecting others, pos- sibly for years. Technological and intentional hazards, however, change in terms of the severity of their consequences all the time. The attacks on the U.S. embassies in Kenya and Tanzania and the September 11 attacks on the World Trade Center and the Pentagon display an increase in the consequences of terrorist attacks aimed at Americans. A mutation of a certain viral or bacterial organism, resulting in a more deadly pathogen, can cause a drastic increase in consequences, as with the West Nile virus, mad cow disease, and SARS.
Changes in human activities are probably the most significant cause of increases in the consequences of disasters. These trends, unfortunately, are predomi- nantly increasing. While the effects of disasters world- wide are great, their consequences are the most devastating in developing countries. Smith (1992) lists six reasons for these changes (Smith, 1992):
1. Population growth. As populations rise, the number of people at risk increases. Population growth can be regional or local, if caused by movements of populations. As urban popula- tions grow, population density increases,
Chapter 3 Risk and Vulnerability 119
exposing more people to hazards than would have been affected previously.
2. Land pressure. Many industrial practices cause ecological degradation, which in turn can lead to an increase in the severity of hazards. Filling in wetlands can cause more severe floods. Lack of available land can lead people to develop areas that are susceptible to landslides, ava- lanches, floods, and erosion, or that is closer to industrial facilities, among other hazards.
3. Economic growth. As more buildings, technol- ogy, infrastructure components, and other struc- tures are built, a community’s vulnerability to hazards increases. More developed communi- ties with valuable real estate have much more economic risk than communities in which little development has taken place.
4. Technological innovation. Societies are becoming more dependent on technology. These systems, however, are susceptible to the effects of natural, technological, and intentional hazards. Technology ranges from communica- tions (the Internet, cell phones, cable lines, satellites) to transportation (larger planes, faster trains, larger ships, roads with greater capacity, raised highways) to utilities (nuclear power plants, large hydroelectric dams) to any number of other facilities and systems (high-rise build- ings, life support systems).
5. Social expectations. With increases in technol- ogy and the advancement of science, people’s expectations for public services, including availability of water, easy long-distance trans- portation, constant electrical energy, etc., also increase. When these systems do not function, the economic and social impacts can be immense.
6. Growing interdependence. Individuals, com- munities, and nations are increasing their inter- dependence on each other. The SARS epidemic showed how a pathogen could quickly impact dozens of countries on opposite sides of the world through international travel. In the late 1990s, the collapse of many Asian economies
sent ripple effects throughout all the world’s economies. The September 11th terrorist attacks in the United States caused the global tourism market to slump.
Disaster managers must investigate the validity of the trends they identify. It is not uncommon for a trend to exist that is based on incomplete records. The tech- nology used to detect many hazards has improved, allowing for detection where it formerly was much more difficult or impossible. Therefore, the lack of recorded instances of certain disasters may very possibly merely be based upon a lack of detection methods.
COMPUTING LIKELIHOOD AND CONSEQUENCE VALUES
Because there is rarely sufficient information to determine the exact statistical likelihood of a disaster occurring, or to determine the exact number of lives and property that would be lost should a disaster occur, using a combination of quantitative and quali- tative measurements can be useful. By combining these two methods, the hazards risk management team can achieve a standardized measurement of risk that accommodates less precise measurements of both risk components (likelihood and consequence) in deter- mining the comparative risk between hazards.
The process of determining the likelihood and con- sequence of each hazard begins with both quantitative and qualitative data and converts it all into a qualita- tive system of measurement that accommodates all possibilities that hazards present (from the rarest to the most common, and from the least damaging to the most destructive).
DEPTH OF ANALYSIS
The depth of analysis to be undertaken by disaster managers depends on three factors: the amount of time and money available, the risk’s seriousness, and
120 Introduction to International Disaster Management
its complexity. According to the information they gather during the identification and characterization of the hazards, disaster managers must decide what level of effort and resources each individual hazard requires.
Each hazard that is analyzed can be considered according to the range of possible intensities it could exhibit. Depending on its characteristics, the hazard may be broken down according to intensity, with separate analyses performed for each possible inten- sity. The likelihood and consequences for each cate- gory of intensity will be different, which in turn results in different treatment (mitigation) options (see Exhibit 3-2).
For instance, the general hazard of “earthquake” could be divided into events of magnitude 4, 5, 6, or 7, and so on. Generally, the lower the intensity of an event, the greater the likelihood of that event occur- ring, while its consequences tend to decrease. Several thousand earthquakes of very low intensity and mag- nitude occur daily with little or no consequences at all. However, the rarer large earthquakes must be treated differently because of their potential to inflict massive casualties and damages.
The degree of subdivision of hazards into specific intensities also depends upon the available time and
resources. More divisions will give disaster managers a more comprehensive assessment, but a point will come when the added time and resources spent no longer provide enough added value.
In summary, effective qualitative risk analysis is performed using four steps:
1. Calculate the (quantitative) likelihood of each identified hazard (broken down by magnitude or intensity if appropriate)
2. Calculate the (quantitative) consequences that are expected to occur for each hazard (broken down by magnitude or intensity if appropriate), in terms of human impacts and economic/finan- cial impacts
3. Develop a locally tailored qualitative system for measuring the likelihood and consequence of each hazard identified as threatening the community
4. Translate all quantitative data into qualitative measures for each hazard’s likelihood and consequence
Disaster managers begin their hazard analysis by calculating (to the best of their ability and resources) the quantitative likelihoods and consequences of each identified hazard risk. It does not matter whether the likelihood or the consequence is analyzed first, or if they are done concurrently, as neither depends upon the other for information. It is important, however, that the quantitative analyses be completed before the qualitative ones, as the qualitative rankings will be based upon the findings of the quantitative analyses.
The following section describes the methods by which the hazards risk management team can perform the quantitative analyses of hazard risks.
QUANTITATIVE ANALYSIS OF DISASTER LIKELIHOOD
Quantitative analysis of the likelihood component of risk seeks to find the statistical probability of the occurrence of a hazard causing a disaster. These
Chapter 3 Risk and Vulnerability 121
EXHIBIT 3-2 f:N Curves
f:N curves, which plot historical hazard inten- sities and likelihoods against the amount of dam- age inflicted, can provide an estimation of both the likelihood of events of specific magnitude and the consequences should those events occur. Exam- ples of worldwide hazard f:N curves are shown in Figure 3-1.
Individual communities would plot f:N curves for their locality using local historical data. This graphical representation illustrates the justifica- tion for dividing hazards according to possible intensities.
122 Introduction to International Disaster Management
1000 10
1
0.1
0.01
Annual Probability
Annual Probability
100
10
1
Natural Disasters Events with death toll greater than or equal to x recorded worldwide in the period of 1900–1975
Transportation Disasters Events with death toll greater than or equal to x recorded worldwide in the period 1946–1975
Industrial / Fire Disasters
Events with death toll greater than or equal to x recorded worldwide in the period 1900–1975
Earthquake
Ficods
Storms
Landslides and Avalanches
10 100 1,000 10,000
Number of People Killed in One Event
Aircraft Accidents
Railway Accidents
Marine Accidents
Mining Accidents
Fires and Explosions
100,000 1,000,000
10 100 1,000 10,000
Number of People Killed in One Event
100,000 1,000,000
10 100 1,000 10,000
Number of People Killed in One Event
100,000 1,000,000
Volcano
Number of Events Recorded 1900 to 1975
10
100
1
Number of Events Recorded 1946 to 1975
10
100
1
Number of Events Recorded 1900 to 1975
10
1
0.1
0.01
Annual Probability
10
1
0.1
0.01
FIGURE 3-1 Examples of worldwide hazard f:N curves. (Source: United Nations Development Programme, 1994.)
analyses tend to be based upon historical data gath- ered in the process of describing identified hazard risks (often called a risk statement). The disaster man- agers performing a quantitative analysis of disaster likelihood must first establish a standard numerical measurement by which the results of all analyzed haz- ards will be reported.
One of the most commonly used quantitative mea- sures of likelihood, and the measure that will be used in this example, is the number of times a particular hazard causes a disaster per year. For example, “In country X, it is predicted that there will be 3 major snowstorms per year.” (For major events that occur less frequently, like a major flood, this number may be less than 1. A 20-year flood has a 5% chance of occur- ring in any given year, or would be expected to occur .05 times per year.) The hazard can now be analyzed according to the chosen standard. If the hazard is one that has been divided into individual intensities and magnitudes, a separate figure will be required for each magnitude or intensity.
If records have been maintained for disasters that occur regularly, such as flash floods or snowstorms, it will be fairly easy to calculate the number of occur- rences that would be expected to happen in a coming year or years. More often than not, however, sufficient information does not exist to accurately quantify the likelihood of a disaster’s future occurrence to a high degree of confidence. This is especially true for haz- ards that occur infrequently and/or with no apparent pattern of behavior, such as earthquakes, terrorism, or nuclear accidents. This inability to achieve precision is a fundamental reason that qualitative measures are used in the final determination of a hazard’s likelihood.
Rare and extremely rare hazards, such as terrorist attacks, nuclear accidents, or airplane crashes (outside of communities where airports exist) may have few if any data points to base an analysis upon. However, this does not mean that there is a 0% probability of the disaster occurring, even if there has been no previous occurrence. For these incidences, consulting with a subject matter expert (SME) is necessary to determine the likelihood of a disaster resulting from the hazard over the course of a given year and to gather any
information on the existence of a rising or falling trend for that particular hazard. Organizations, profes- sional associations, and other bodies, such as the United Nations, national governments, and research facilities, maintain risk data on particular rare hazards. Modeling techniques also can be used to estimate the likelihood of infrequent events.
The more often that a disaster occurs, the more data points those performing the quantitative likelihood assessment will have, and therefore, the more accurate the historical analysis will be (given that the collected data is, in fact, accurate). However, more information must be examined than simply the number of events per year.
The concept of increasing and decreasing trends in hazard likelihoods and consequences was introduced above. Both infrequent and frequently occurring disasters tend to exhibit either falling or rising trends over time, rather than having a steady rate of occurrence. These rising and falling trends must be accounted for if there is to be any accuracy attained in an analysis of likelihood.
For example, if a community has sustained approx- imately 35 wildfires per year for the past 40 years, it might easily be assumed that it is very likely there will be approximately 35 wildfires per year in the com- ing years. However, further inspection of historical records discovers that 40 years ago, there was 1 fire, and 39 years ago, there were 3 fires. The number of fires steadily increased until the historical record ended with 70 fires occurring in the past year. Over the 40-year period, the average number of wildfires is in fact 35 per year. However, the rate of wildfires has increased each year from 1 per year 40 years ago to 70 per year last year. Considering this trend, the expected number of wildfires next year cannot be expected to be 35, although the average per year is 35.
It must be assumed from this data that there is a ris- ing trend in the occurrence of wildfires, and that there is likely to be 70 or more fires in the coming year. Why this rising trend is occurring and what can be done to counteract it will need to be examined in the process of determining vulnerability and generating mitigation and preparedness options.
Chapter 3 Risk and Vulnerability 123
QUANTITATIVE ANALYSIS OF DISASTER CONSEQUENCES
The quantitative analysis of disaster consequences seeks to determine the number of injuries, the number of deaths, the cost of direct damages to property and infrastructure, and the indirect costs associated with the disaster. (Depending on the scope of the analysis, other factors such as homelessness or displacement may be considered as well.) A standard form of mea- surement must be established for deaths, injuries, and damages. It is most useful if the measurement is per occurrence, as opposed to per year or other time frame.
It will be necessary to analyze the expected conse- quences of each magnitude or intensity of a hazard if it has been broken down into subcategories.
HISTORICAL DATA
As with the likelihood component of risk, the calculation of hazard consequences should begin by examining the historical data on injuries, fatalities, and property/infrastructure damage and destruction that was gathered during the identification of hazards. However, as previously described, human behavior and/or changes in hazard characteristics often result in either increasing or decreasing trends in disaster con- sequences over time. Changes in settlement or new development, for example, can significantly increase community vulnerability for two different occurrences of a hazard.
Historical information does have its uses, however, especially with more common hazards for which data has been collected methodically and accurately for many years. Consequence data based upon historical information can act either as a benchmark to validate the findings of more in-depth analyses (described below) or as the actual estimation of consequences, should disaster managers decide to perform a lower level of analysis.
In the section addressing vulnerability, we will explain the process of describing the community and the environment. In this process, information is gath- ered on the physical community, the built environ- ment, and the social environment, as well as on the critical infrastructure and the interdependence of the community on surrounding and other external communities.
Using hazard maps created or obtained during the process of hazard identification, combined with the description of the community environment, disaster managers can develop numerical figures for the expected number of lives that will be lost, people who will be injured, and the dollar amount of the direct and indirect damages that may occur. (However, it is always important to keep in mind that even the most extensive analyses of consequences are imperfect, as they are heavily based upon assumptions and upon historical data that may or may not indicate future behavior of hazards.)
Consequence analyses must look not only at the location of structures in relation to the hazard but also at the vulnerability of each structure. For instance, imagine that a school is located in a floodplain. Disas- ter managers have obtained information indicating that the school has been raised to an elevation where it will only be affected by floods of magnitude greater than the 50-year (2% chance/year) flood. Using this information, disaster managers can deduce that such a structure will likely sustain no damage during the course of a 20-year (5% chance/year) flood event.
While disaster managers will likely not have the value of all structures within the community or be able to determine complete data pertaining to lost revenue and inventory, such data deficiencies probably will be consistent across all hazard consequence analyses and will therefore not necessarily cause the results of the analyses to be unreliable. Obviously, more data gener- ally results in more accurate assessments. However, the amount of data that can be collected will always be a factor of available time and resources. Moreover, the process of translating the quantitative data resulting from these analyses into the qualitative determination
124 Introduction to International Disaster Management
of likelihood and consequence can be tailored to accommodate for almost any lack of accuracy.
DEATHS/FATALITIES AND INJURIES
Disaster managers can estimate the number of peo- ple who will be hurt or killed by using two methods: estimation based upon historical data and changes in population, or modeling techniques.
To estimate the number of deaths and injuries using historical data, disaster managers must first assemble the data on historical incidences of disasters caused by the particular hazard. Then, using current on the community, a conversion to current conditions can be made. For example, imagine that a Category IV hurri- cane struck a community in 1955, causing 4 deaths and 35 injuries. The population of the community at the time was approximately 10,000. Today, the popu- lation is estimated to be 15,000; in other words, it has increased by a factor of 1.5. By multiplying the his- torical consequence data by this conversion factor, disaster managers could surmise that there would be approximately 6 deaths and 52 injuries if a Category IV hurricane struck today.
It must be kept in mind that these estimates do not account for mitigation measures taken or new devel- opment in the period between disasters. The more recently a comparable disaster has occurred, the more accurate the conversion will be. The use of modern modeling techniques, such as HAZUS and HAZUS- MH, a nationally standardized, GIS-based risk assess- ment and loss estimation tool developed by the Federal Emergency Management Agency (FEMA), can increase the accuracy of injury and death estimations.
MODELING TECHNIQUES
Various computer-modeling techniques are avail- able to assist disaster managers in estimating the injuries and deaths that would occur should a disaster
strike. For instance, HAZUS can be used to estimate the number of injuries and fatalities that would result from earthquakes of varying magnitudes, and it is being updated to include other hazards such as wind and floods. Other models give estimates for hazards such as chemical releases and floods.
The data collected on base maps and the hazard- specific maps created during the hazard identification and description process also can be used to estimate the population affected by the hazard.
Regardless of the method used, a high degree of accuracy is very difficult to attain when estimating the number of injuries and deaths that would occur in future disasters. Many confounding variables affect human behavior and the ability to react to hazard events, including warning times and warning accura- cies, the nature of the hazard, and the numbers, resources, and abilities of the emergency responders. These estimations should always be taken to be just that—estimations. The experience of the disaster management team and of other community experts such as first responders and the medical community can be just as valuable in making these estimates.
ABBREVIATED DAMAGE CONSEQUENCE ANALYSIS
If disaster managers choose to perform a lower level of analysis on the consequences of the commu- nity’s hazards, two pieces of information are needed. The first is the historical incidence of hazard damage for each disaster. The second is data on the popula- tion/structural changes in the community since the date of each historical disaster in order to compare to present-day data. Once that data is assembled, the team can calculate damages as they would be expected to affect the community as a comparison between the dates. For instance, imagine that a flood (of a specific magnitude) in 1955 caused $1 million in damages in a community. The community is found to have grown approximately 50% in the floodplain in the intervening years. Using this information, the hazards risk management team can estimate the
Chapter 3 Risk and Vulnerability 125
consequences of a future event of similar magnitude to be approximately $1.5 million in 1955 dollars, or $10,305,622 in 2005 dollars. Currency inflation con- verters are widely available on the Internet—one site is www.westegg.com/inflation/
If a certain hazard has not affected the community over a significantly extended period of time, or if it has never affected the community, the team may want to either use data from an example of the hazard affecting a community of comparable structure and size or avoid performing a quantitative analysis for the rare hazard.
FULL DAMAGE CONSEQUENCE ANALYSIS
A full damage consequence analysis requires that disaster managers consider the current estimated cost of all physical assets within the country. These include:
● Losses to structures. Estimated as a percentage of the total replacement value. This figure is obtained by multiplying the replacement value of the structure by the expected percent damage to the structure.
● Losses to contents. Estimated as a percentage of the total replacement value. This figure is obtained by multiplying the replacement value of the contents by the expected percent damage.
● Losses to structure use and function and cost of displacement. The losses to structure use is a function of the number of days the structure is expected to be out of use multiplied by the aver- age daily operating budget or sales (annual rev- enue or budget divided by 365 days). The cost of displacement is the product of the costs incurred as result of the business/service being displaced and the number of days that displacement is necessary. These calculations can apply to businesses, bridges, utilities, public services (libraries), and any other community asset.
To track calculated figures, a standardized work- sheet is often created. One example of a standardized worksheet provided by FEMA is shown in Figure 3-2.
Each hazard will affect structures and their con- tents differently. Many organizations and institutions have made available tables to determine this informa- tion for specific hazards. To perform a full damage consequence analysis, disaster managers will need to have the following information (which is often gath- ered during the process of describing the community and environment and determining the vulnerability of the community):
● Replacement value of all community assets (homes, businesses, and infrastructure)
● Replacement value of inventory (business inven- tory, personal property in homes, contents of government offices and other buildings)
● Operating budgets/annual revenues of businesses and government assets
● Costs of relocation of operations/services
Once quantitative figures have been calculated for both the likelihood and consequence components of risk, the disaster managers can begin the process of determining the qualitative values assigned to the like- lihood and consequence for each hazard (and hazard intensity or magnitude, if the hazard is subdivided into such). They should begin by selecting a system of qualitative measurement or by designing one that suits the needs of both the format of results in the quantita- tive analysis and the characteristics of the particular country or community.
A disaster, as defined in Chapter 1, is “a serious disruption of the functioning of society, causing widespread human, material, or environmental losses which exceed the ability of the affected society to cope using only its own resources” (UNDP, 1994). Therefore, a specific set of hazard consequences may constitute a disaster in one community but not in another. For instance, 10 injuries may exceed the capacity of the local clinic in a community of 500, but in a large city, 10 injuries could be easily managed.
Whether designing a new system of measurement or using an existing one, it is necessary for the disas- ter management team to be aware of the local capac- ity in order to know how many deaths and injuries and how much damage can be sustained before the local
126 Introduction to International Disaster Management
Chapter 3 Risk and Vulnerability 127
FIGURE 3-2 FEMA standardized loss estimation worksheet. (Source: FEMA, 2001.)
capacity is either stressed or exceeded. They will have the data collected in the hazard identification process and in the description of the community and the envi- ronment (described later in this chapter) upon which to base their new or acquired system of measurement.
Creating two measures of consequence can be ben- eficial: one measuring the tangible physical/material losses associated with cost, and another measuring the intangible losses of deaths/fatalities and injuries. Each qualitative term should have two measures associated with it, corresponding to deaths/injuries and costs. In many instances, the tangible and intangible rankings will not be the same. For instance, there may be no physical damages to structures in a chemical spill, but many people may be injured or die. Other events may cause no immediate deaths or injuries, but cause a great amount of physical loss, such as a large-scale power outage. In either case, the factor that achieves the qualitative measure of greater (higher) conse- quence is used to determine the consequence of the hazard.
Tables 3-1 to 3-4 provide multiple examples of qualitative measures of likelihood and consequence.
Once a measurement system has been chosen, the disaster managers can assess each hazard according to its qualitative likelihood and consequences, using the quantitative data obtained in the previous steps of the hazard analysis process. These qualitative rankings are then recorded and assessed according to a risk assessment matrix (described below.)
When assessing the qualitative ranking for a hazard consequence, two different types of consequences are usually examined—human impacts (injuries and deaths/fatalities) and material/physical losses. In determining the qualitative consequence ranking, the
128 Introduction to International Disaster Management
TABLE 3-1 An example of a qualitative likelihood measurement system
Rating Description and indicative probability
Almost certain Expected to occur; many recorded incidents; may occur or be exceeded once every 1 to 5 years
Likely Will probably occur; may occur or to be exceeded once every 20 years
Possible Might occur; may occur or be exceeded once every 100 years; will generally be close to or exceed past records of severity
Unlikely Is not expected to occur; little opportunity, reason, or means to occur; may occur or be exceeded once every 250 years
Rare May only occur in exceptional circumstances; may only occur or be exceeded once every 500 years or more
Source: Cameron, 2002.
TABLE 3-2 An example of a qualitative likelihood measurement system
Descriptor Description
Almost certain Is expected to occur in most circumstances; and/or high level of recorded incidents and/or strong anecdotal evidence; and/or a strong likelihood the event will recur; and/or great opportunity, reason, or means to occur; may occur once every year or more
Likely Will probably occur in most circumstances; and/or regular recorded incidents and strong anecdotal evidence; and/or considerable opportunity, reason or means to occur; may occur once every five years
Possible Might occur at some time; and/or few, infrequent, random recorded incidents or little anecdotal evidence; and/or very few incidents in associated or comparable organizations, facilities or communities; and/or some opportunity, reason or means to occur; may occur once every twenty years
Unlikely Is not expected to occur; and/or no recorded incidents or anecdotal evidence; and/or no recent incidents in associated organisations, facilities or communities; and/or little opportunity, reason or means to occur; may occur once every one hundred years
Rare May occur only in exceptional circumstances; may occur once every five hundred or more years
Source: Emergency Management Australia, 2000.
hazards risk management team will choose whichever ranking is greater. (Differences between the severity of human and material losses often exist. A poisonous gas leak is a good example of a hazard where few material or physical damages are likely, but many
deaths and injuries could occur. In that case, the hazards risk management team would probably base their assessment on the human consequences of the hazard rather than the material/physical consequences.)
Chapter 3 Risk and Vulnerability 129
TABLE 3-3 An example of a qualitative consequence measurement system
Descriptor Human life and health Property, financial, environmental
Insignificant No injuries or fatalities Inconsequential or no damage Small number or no people are Little or no disruption to community
displaced and only for a short No measurable impact on duration environment
Little or no personal support Little or no financial loss required
Minor Small number of injuries but no Some damage fatalities. First aid treatment Small impact on environment with no required last effects
Some displacement of people Some financial loses (<24 hrs)
Some personal support required Some disruption (<24 hrs)
Moderate Medical treatment required but no Localized damage that is rectified by fatalities. Some hospitalization routine arrangements. Normal
Localized displacement of people community functioning with some who return within 24 hours inconvenience
Some impact on environment with long-term effect
Significant financial loss
Major Fatalities Significant damage that requires Extensive injuries, significant external resources. Community only
hospitalization partially functioning, some services Large number displaced (>24 hrs unavailable
duration) Some impact on environment with External resources required for long-term effects
personal support Significant financial loss—some financial assistance required
Catastrophic Significant fatalities Extensive damage Large number of severe injuries Extensive personal support Extended and large numbers Community unable to function
requiring hospitalization without significant support General and widespread Significant impact on environment
displacement for extended and/or permanent damage duration
Source: Cameron, 2002.
RISK EVALUATION
Risk evaluation is conducted in order to determine the relative seriousness of hazard risks for the country or community being assessed by the disaster manager. Using the processes listed above and in Chapter 2 to identify hazards that threaten the community, charac- terize them, and determine their likelihoods and con- sequences, the disaster managers will have gathered as much information.
By the time the risk evaluation process begins, each hazard will have been identified, described, mapped, and analyzed according to its likelihood of
occurrence and its consequences should a disaster occur. All countries and communities undoubtedly face a range of natural, technological, and intentional hazards, each of which requires a different degree of mitigation and risk reduction. Unfortunately, commu- nities rarely are able to dedicate sufficient resources to mitigation to lower all of the community’s risks to the lowest possible levels.
As will be shown in Chapters 4 and 5, there are hazards for which the technology exists for mitigation but are cost prohibitive. An example of a risk mitiga- tion measure that is very expensive is the conversion (retrofit) at wastewater treatment plants to less dan-
130 Introduction to International Disaster Management
TABLE 3-4 An example of a qualitative consequence measurement system
Descriptor Description
Insignificant No injuries or fatalities. Small number or nil people are displaced and only for short duration. Little or no personal support required (support not monetary or material). Inconsequential or no damage. Little or no disruption to community. No measurable impact on environment. Little or no financial loss
Minor Small number of injuries but no fatalities. First aid treatment required. Some displacement of people (less than 24 hours). Some personal support required. Some damage. Some disruption (less than 24 hours). Small impact on environment with no lasting effects. Some financial loss
Moderate Medical treatment required but no fatalities. Some hospitalization. Localized displacement of people who return within 24 hours. Personal support satisfied through local arrangements. Localized damage that is rectified by routine arrangements. Normal community functioning with some inconvenience. Some impact on environment with no long-term effect or small impact on environment with long-term-effect. Significant financial loss
Major Extensive injuries, significant hospitalization, large number displaced (more than 24 hours duration). Fatalities. External resources required for personal support. Significant damage that requires external resources. Community only partially functioning, some services unavailable. Some impact on environment with long-term effects. Significant financial loss— some financial assistance required
Catastrophic Large number of severe injuries. Extended and large numbers requiring hospitalization. General and widespread displacement for an extended duration. Significant fatalities. Extensive personal support. Extensive damage. Community unable to function without significant support. Significant impact on environment and/or permanent damage
Source: Emergency Management Australia, 2000.
gerous chemicals, such as using liquid chlorine bleach or other disinfection technologies instead of the more volatile chlorine gas. Exhibit 3-3 illustrates the danger posed by chlorine gas, which is still widely used despite its known dangers.
Other risks may have many options available, each with an associated cost and benefit. Some have direct risk reductions with each incremental increase in cost. A classic example is the practice of increasing the number of firefighters or police officers in a commu- nity, which, until reaching a threshold, results in decreased fire hazard risk and decreased crime risk.
Fortunately, however, not all risks require immedi- ate action, and some do not require any action at all. These include those risks for which both the likeli- hood and the consequences of the risk are extremely low, such as a small meteor strike. While some risks can be reduced easily, others may require exorbitant cash resources, time, and a committed effort to achieve even slight reductions. These possibly limit-
ing factors must also be considered by disaster managers.
In addition to actual reductions in risk related to the likelihood and consequences of a hazard, several risk factors must be considered that will weigh heavily on the perceived “seriousness” of the risk and therefore affect mitigation priorities. For instance, a manmade risk is likely to be considered much less “acceptable” than one that is natural in origin. The degree to which these manmade risks are perceived to be unacceptable can be an important determining factor in assigning mitigation funding. Smith (1992) discusses voluntary and involuntary risks and states, “[T]here is a major difference between voluntary and involuntary risk perception with the public being willing to accept vol- untary risks approximately 1000 times greater than involuntary risks.”
Risk perception issues also weigh heavily upon such decisions. For instance, consider a rural commu- nity in which one person dies per year as result of
Chapter 3 Risk and Vulnerability 131
EXHIBIT 3-3 Description of the Dangers of Using Chlorine Gas to Purify Water
Chlorine is often used as a disinfectant in most of the world’s water systems because of its cost- effectiveness. The chemical is usually stored in a pressurized, liquid state. When released, chlorine vaporizes into a highly toxic, invisible gas that con- centrates at ground levels. Germany used chlorine gas during World War I for this reason, because it would settle into the trenches where British troops were hiding.
It has been estimated that anyone located within two or three miles from a ruptured 90-ton chlorine railcar would be killed if directly exposed to the ensuing cloud. Injuries, including fluid in the lungs and a permanently reduced breathing capacity, could result at distances as great as 10 miles.
Because of the increasing risk of terrorism and other criminal attacks on storage facilities, the [US] Environmental Protection Agency (EPA) has dis-
tributed guidelines that encourage US chemical industry businesses to employ safer technologies. One such facility, the Washington, DC-based Blue Plains wastewater treatment plant, heeded this advice and fully converted from the use of chlorine- gas disinfectant to the safer liquid chlorine bleach. The plant’s close proximity to the nation’s capital placed it at high perceived risk of terrorist attack, but only as long as the highly-volatile chlorine gas was stored on the site. In switching to liquid chlorine bleach, the threat has essentially been eliminated.
Many other drinking and wastewater treatment plants have also switched to safer technologies. In addition to liquid chlorine bleach, ultraviolet light and ozone may be used to purify the water.
Source: Davis, 2002.
cave-ins of abandoned mine shafts and approximately four people per year are drowned in a river that regularly experiences swift currents following storms. There is likely to be considerable public outcry over the yearly incidence of fatal accidents from the aban- doned mines, while the river drowning is viewed as a controllable, easily reduced, voluntary, preventable, observable hazard whose effects are known to those exposed (risk perception concepts are described in greater detail later in this chapter).
There are also risks that societies are able to elimi- nate altogether but choose not to because the benefits that result from such risks would also disappear (see Exhibit 3-4). This essentially implies that, when eval- uating risks, disaster managers must also consider the negative consequences of mitigation or elimination. Eliminating certain beneficial risks can result in adverse effects on the community or society. Exam- ples of situations where the benefits are believed to outweigh the risks include the aesthetic value to
homeowners and collected property taxes for the com- munity from beachfront property construction; col- lected taxes and created jobs for a community that result from the existence of a factory that produces, stores on-site, or emits hazardous materials; and the reduced reliance on fossil fuels and cheaper power generation costs that exists as result of a nuclear power plant.
One of the primary goals of disaster managers is to formulate a prioritized list of hazard risks to be miti- gated. This list should be based upon a combination of factors that includes the hazard’s likelihood and con- sequences, the county’s or community’s priorities and criteria (in regards to their views on the acceptability of different risks), the benefit-to-cost ratios of mitigat- ing different risks, and the political and social ramifi- cations of certain mitigation decisions.
Hazards were examined individually in each previ- ous step of this process. During the risk evaluation step of the process, risks are compared to each other,
132 Introduction to International Disaster Management
EXHIBIT 3-4 Acceptability of Risk
Almost everything that provides a benefit also creates some level of risk for either the benefac- tor(s) or for others who do not necessarily enjoy those benefits. This risk ranges from barely measur- able to severe. The side effects of certain prescrip- tion drugs, negative health effects from “fast food,” or skin cancer from the sun are a few examples at the personal level. On a larger scale, more specifi- cally related to disaster management, is the inunda- tion danger associated with the construction of a power-generating dam. As a society, citizens have come to accept most of these risks without ques- tion, although many present much greater risks than some others people refuse to accept.
For instance, tens of thousands of people are killed and over tens of millions suffer disabling injuries each year from falls while using stairs in their homes and elsewhere (Roderick, 1998). It is
unlikely that stairways will be eliminated, despite the fact that they injure and kill many more people than hazards like saccharin, fluoroscopes (shoe- fitting X-ray machines), and extralong tandem trailer trucks, for instance. Why are people willing to accept one risk and not another? The answer can be found in the perceived benefits of each risk. Peo- ple perceive that the benefit of having multiple sto- ries in a house or other building is worth the risk of injury or death from using stairways. Society does not perceive the risk of injury, illness, or death resulting from saccharin, fluoroscopes, or tandem trucks to be worth the benefits gained from each (low-calorie sweetener, an X-ray look at your foot inside a shoe, and the truck’s greater carrying capacity), even though each of these three examples poses less of an absolute population risk than stair- ways do.
and questions of priority begin to be answered. Prior- itization can take place by many methods, and while there is no single correct method, there are many that have been used with success in the past.
The following may be used to determine the prior- itization of risk treatment:
1. Creating a risk matrix 2. Comparing hazard risks against levels of risk
estimated during the analysis process with pre- viously established risk evaluation criteria
3. Evaluating risks according to the SMAUG methodology (seriousness, manageability, ac- ceptability, urgency, growth)
The final output of risk evaluation should be a pri- oritized list of risks, which will be used to decide treatment (mitigation) options.
Hazard analysis determined qualitative values describing the likelihood and consequence of each hazard. For those hazards known to exhibit a range of magnitudes or intensities, the likelihood and conse- quence values were determined for several magni- tudes or intensities across the range of possibilities.
Assigning these qualitative values was the first step in a process that allows for a direct comparison of the risks faced by a community. Armed with both the like- lihood and consequence values, disaster managers can now begin comparing and ranking the identified risks.
To compare hazards according to their likelihood and consequences, the team must select or create a risk matrix to suit the needs of the country or commu- nity. A risk matrix is a direct comparison of the two components of a hazard’s risks. In other words, it plots the likelihood and consequence of hazards together in various combinations, with one risk component falling on the X-axis and the other on the Y-axis.
While it does not matter which of these two risk components goes on which axis, the values used must exactly match the values used in the risk analysis qualitative assessments. This is because the terminol- ogy must be consistent throughout the process of “cal- culating” risk from likelihood and consequence, much as if quantitative (numerical) values were being used. For instance, if the possible range of values for the
likelihood of a risk included the values “Certain,” “Likely,” “Possible,” “Unlikely,” “Rare,” and “Extremely Rare,” then the risk matrix must include all of those values (on the appropriate axis), in logical consecutive order.
Plotting these values on the matrix results in indi- vidual boxes representing unique combinations of likelihood and consequence. The likelihood and con- sequence values upon which the individual boxes are based can be determined by tracing from that box back to the values indicated on each axis. The number of possible combinations will be the product of the number of likelihood values times the number of consequence values (i.e., if there are 5 values for likelihood and 6 for consequence, the matrix will have 30 possible combinations required to evaluate risk).
Disaster managers must decide whether to use a pre-existing risk matrix or to make a custom risk matrix that suits their specific needs. If they chose to create their own systems of qualitative measure- ment in the risk analysis process, they must make their own risk matrix. However, even if they used an existing set of qualitative measurements in the risk analysis process, a risk matrix to evaluate each risk may not exist, in which case they would need to make one.
To create a risk matrix, disaster managers must first establish levels, or “classes,” of risk representing increasing severity. The levels should range from those that are so low that mitigation is not necessarily needed to risks that are so high that efforts to mitigate them are of highest priority.
One example of such a system is described in the FEMA’s “MultiHazard Identification and Risk Assess- ment” publication (1997). Their risk matrix values are:
1. Class A. High-risk condition with highest prior- ity for mitigation and contingency planning (immediate action)
2. Class B. Moderate to high risk condition with risk addressed by mitigation and contingency planning (prompt action)
Chapter 3 Risk and Vulnerability 133
3. Class C. Risk condition sufficiently high to give consideration for further mitigation and plan- ning (planned action)
4. Class D. Low-risk condition with additional mitigation contingency planning (advisory in nature)
Emergency Management Australia (EMA) (2000) describes risks according the following breakdown:
1. Extreme risk 2. High risk 3. Moderate risk 4. Low risk
Other systems include “Intolerable, Undesirable, Tolerable, Negligible” and “Severe, High, Major, Sig- nificant, Moderate, Low, Trivial.”
Once these values have been determined and defined as they apply to the disaster manager’s priori- ties, they should be assigned to each combination of likelihood and consequence shown on the matrix. How they are assigned must be determined by per- sonal judgment, expert knowledge, and previously established risk management criteria. An example of a risk matrix from FEMA is shown in Figure 3-3.
Once the values have been assigned to each box on the matrix, each hazard can be evaluated accordingly and the derived values recorded. Because each “risk level” will likely be assigned to more than one matrix box, and because several risks could elicit the same combination of likelihood and risk, the hazards risk management team will not be creating an ordered list of risk priorities, but rather several categories of risk with several hazards falling within each category group. In other words, the disaster manager will have several “classes” of risk, each containing several risks for which no intraclass priorities have been determined.
For instance, if a 50-year flood was determined to be a Class C risk, and an accident involving a truck carrying hazardous materials was determined to be a Class C risk, they would be considered equal risks according to the risk matrix. The results of the risk matrix allow disaster managers to further classify the
hazards threatening their country or community but do not provide a definitive list of priorities for mitigation. Such a list requires further evaluation, as will be described.
It is helpful for disaster managers to begin record- ing the results of their evaluations on a concise form that allows fast and easy reference to risk evaluation output data so this data can be more easily compared in the prioritization step. Risk registers, as they are called, provide a useful tool, and should include the following information:
● Name of the risk (including specific magnitude and/or intensity if the risk has been broken down into these categories)
● Qualitative likelihood value ● Qualitative consequences value ● Level of risk as determined by evaluation on the
risk matrix
134 Introduction to International Disaster Management
FIGURE 3-3 FEMA “MultiHazard Identification and Risk Assessment” risk matrix. Class A: High-risk condition with highest priority for mitigation and contingency planning (immediate action) Class B: Moderate to high-risk condition with risk addressed by mitigation and contingency planning (prompt action) Class C: Risk condition sufficiently high to give consideration for further mitigation and planning (planned action) Class D: Low-risk condition with additional mitigation contingency planning (advisory in nature)
● Priority rating ● Additional information, including any of the
following � Description of possible consequences � Adequacy of existing mitigation measures or
controls � Known mitigation options and alternatives � Acceptability of risk
Because people have different risk perceptions, and because there may be more risks than there are resources to mitigate them, disaster managers must develop risk evaluation criteria before any risk identi- fication or analysis takes place. Risk evaluation criteria help disaster managers and citizens make judgments about what they consider to be the most serious risks and set forth performance measures to judge progress in mitigating the community’s risks.
In establishing these contextual criteria, disaster managers will also define the political, social, eco- nomic, legal, and physical environment within which all of the hazards can occur. Some of criteria include:
● Population issues � Death and injuries � Displacement � Loss of homes and property � Loss of jobs and income � Loss of sense of security � Loss of sense of community
● Business sector issues � Damage to facilities � Loss of income � Business disruption costs � Insurance losses � Loss of market share � Loss of trained employees � Bankruptcy
● Community issues � Damage or destruction of community infra-
structure (i.e., roads, bridges, hospitals, jails, city halls, community service centers, etc.)
� Loss of tax revenues � Disaster response and recovery costs
� Reduced funding for other community priori- ties (i.e., education, social services, etc.)
� Loss of population base � Increased community debt and borrowing � Economic repercussions � Environmental harm � Loss of culture/heritage
Disaster managers would also define their analysis as it relates to mitigating the country or community’s hazards. This could include several or all of the following:
● Legal requirements ● Cost and equity ● Risks that are clearly unacceptable ● Risks that should be kept as low as reasonably
practicable
Additionally, risks that have been evaluated according to the risk matrix will need to be verified for accuracy. It is possible that a risk may have been placed in a category that defines it as being either too great or not great enough—only further analysis can correct such errors.
The Purpose of Evaluating Risk
Gaye Cameron of the University of New South Wales (2002) writes, “The purpose of evaluating risks is to determine that risk levels resulting from the risk analysis step [including the results of the risk matrix] reflect the relative seriousness of each risk.” She men- tions three tasks that are important to perform at this point in the hazards risk management process:
● Identify which risks require referral to other agencies (i.e., is the risk one that is better miti- gated by another local, regional, or national agency rather than one that needs to be con- sidered for mitigation options by the disaster managers?)
● Identify which risks require treatment by the disaster managers
● Further evaluate risks using judgment based upon available data and anecdotal evidence to
Chapter 3 Risk and Vulnerability 135
further determine the accuracy of the final risk value recorded.
A risk that might be better mitigated by another local, regional, or national agency is hazardous mate- rial exposure and other accidents that might occur at or from an extra-jurisdictional utility (like a nuclear power plant) that is adjacent to a second country or community. Hazards that are created in one jurisdic- tion but whose consequences affect another have caused many cantankerous debates throughout his- tory. These types of cross-jurisdictional problems are most severe on rivers and streams. Pollution content, increased flooding potential, and even decreased quantities of water can all occur in one jurisdiction but be caused by the actions of another. An illustrative example is changes in a river’s hydrology brought about by the construction of manmade levees (water- retention walls built along the banks of rivers that allow for higher water levels before flooding occurs). Dams and levees are river structures that often cause these problems. They can cause flooding, both up- stream from rising water levels in reservoirs behind the dam and downstream from forced release or fail- ure of the dam.
Cameron (2002) writes that there are two over- arching issues that need to be addressed in the risk evaluation process. First, risk levels must be con- firmed. Through a process of stakeholder consultation, these levels are reviewed to ensure:
1. They reflect the relative seriousness of each risk.
2. The likelihood and consequence descriptions utilized for risk analysis are appropriate.
3. Local issues have been considered.
Cameron adds, “If, following stakeholder consulta- tion, the risk level is considered inappropriate the risk should be subjected to further analysis using new information or data.”
Second, risk acceptability must be addressed:
In almost all circumstances risk acceptability and treat- ment will be determined and/or carried out by the agency or agencies responsible for managing the treatment of risks.
For those risks where no agency is responsible, the [disaster managers] will prepare treatment options for the manage- ment of the identified risks. (Cameron, 2002)
For each risk, the levels of risk acceptability (by both the public and by the disaster managers) must be determined in order for the level of mitigation effort required to be determined. Risk acceptability will be discussed in greater detail later.
Once the risk levels of each hazard have been com- pared to the previously established risk evaluation cri- teria, the risks must be prioritized, or ranked in the order that the disaster managers feel they should be addressed.
This prioritization can be accomplished in many ways, most of which rely upon the information gath- ered in the previous steps of the process and build upon the results of the risk matrix. Risk prioritization takes the evaluation of a country or community’s haz- ards beyond merely comparing risks as factors of like- lihood and consequence, and uses the expert judgment of the hazards risk management team to add experi- ence, knowledge, and contextual influence to the final determination of mitigation priority.
In risk prioritization, disaster managers must consider the degree of control over each risk and the cost, benefits, and opportunities presented by each risk, and decide which risks are unacceptable at any cost.
One such method for the evaluation of risk, the so-called “SMAUG” (Seriousness, Manageability, Acceptability, Urgency, Growth) approach, designed by Benjamin Tregoe and Charles Kepner, has gained wide acceptance by emergency managers in Australia and New Zealand.
According to this methodology, disaster managers consider five individual factors in determining how a list of risks can be generated that reflects the estab- lished priorities of the community. This list includes (each factor is accompanied by the upper and lower extremes by which each risk could be evaluated):
1. Seriousness a. The risk will affect many people and/or will
cost a lot of money (see Exhibit 3-5).
136 Introduction to International Disaster Management
b. The risk will affect few or no people or will cost little or nothing.
2. Manageability a. The risk could be affected by intervention. b. The risk cannot be affected by intervention.
3. Acceptability a. The risk is not acceptable in terms of politi-
cal, social, or economic impact. b. The risk will have little political, social, or
economic impact.
Chapter 3 Risk and Vulnerability 137
EXHIBIT 3-5 Considering Extreme Events
Rae Zimmerman and Vicki Bier, in their article “Risk Assessment of Extreme Events,” shed some light on the extra considerations that must be made when prioritizing hazard lists that include extreme event hazards that are manmade and intentional, such as terrorism.
They write, “Predicting human behavior in emergency situations is already difficult. However, in attempting to estimate and manage the risks of intentional attacks, further difficulties become apparent. First, as pointed out by Woo (2002), ‘some idea of event likelihood is needed for intelli- gent benefit-cost analysis.’ However, estimating the likelihood and nature of intentional attacks is an area with which most risk assessors are not yet familiar, although there has been some related work on this problem in other fields. For example, Dickey (1980) interviewed bank robbers to under- stand the criteria that they used in choosing banks to rob; he found that they preferred banks located near major highways and banks with a single point in the lobby from which they could see all of the employees at once. Similarly, Crowe (2000) and de Becker (1997) report that criminals choose targets based not only on the attractiveness of the target but also on the likelihood that they would be discovered and apprehended. Interviews with incarcerated ter- rorists could presumably be used to explore the cri- teria they use in selecting targets, which could be factored into quantitative risk assessments.
“More significantly, protection against a knowl- edgeable and adaptable adversary is a fundamen- tally different challenge than protection against accidents or acts of nature. For example, earth-
quakes do not get stronger or smarter just because we defend our buildings against them. However, if adversaries know or can easily learn about their tar- get’s defensive measures, then they can actively choose to either bypass or circumvent those defenses. Progress in and increased reliance upon detection technologies has made this more impor- tant to take into account. For example, metal- screening devices prior to September 11th increased the security and safety of air travel. A net- work news report early in 2002 suggested that the box cutters used by the terrorists on September 11th to gain control of the hijacked airplanes fell just below the detection settings of such screening devices.
“As noted by Dresher (1961), optimal allocation of defensive resources requires that ‘each of the defended targets yield the same payoff to the attacker.’ Thus, even if some components can be shored up quite inexpensively, focusing protective investments there can lead to wasted resources if adversaries choose to attack targets that cannot be shored up cost effectively. In other words, critical assets must be defended against all possible attacks, which is much more difficult than just shoring up a few ‘weak links.’ As a result, Ravid (2001) con- cludes that security improvements are generally more costly than safety improvements: ‘[I]nvest- ment in defensive measures, unlike investment in safety measures, saves a lower number of lives (or other sort of damages) than the apparent direct con- tribution to those measures.”
Source: Zimmerman and Bier, 2002.
4. Urgency a. The risk urgently needs to be fixed. b. The risk could be fixed at a later time with
little or no repercussions. 5. Growth
a. The risk will increase quickly. b. The risk will remain static. (Lunn, 2003)
Using the SMAUG criteria for evaluation, disaster managers can more precisely determine priorities for mitigating individual risks, beyond the characteriza- tions that resulted from the risk matrix. After the risk matrix evaluation, risks were grouped into categories of seriousness. Now they can be assigned a numerical order defining specific priorities.
It is important to note that the list of priorities will likely change as the risk mitigation options are con- sidered. Risk evaluation has given the hazards risk management team a better idea of those risks for which mitigation must be conducted at all costs, due to their absolute unacceptability. However, for risks with similar mitigation priority rankings, the factors of cost effectiveness of mitigation, technological availability of mitigation options, and other risk treat- ment factors will require revisiting this priority list and re-ranking risks using additional information.
RISK ACCEPTABILITY
In performing hazard risk assessments and analy- ses of risk, disaster managers must make decisions about what risks to treat, what risks to prevent at all costs, and what risks can be disregarded because of either low consequence, low frequency, or both. These decisions are based upon the acceptability of risk.
Unfortunately, no disaster manager will ever have complete information about all risks faced by the country or community in regards to the number of people and the area affected, the actual frequency of the hazard in the future, and the actual benefit to be attained through mitigation, among many other fac- tors. If the disaster manager did have all of this infor- mation, determining risk acceptability and making
mitigation decisions would be simple. However, in the absence of this perfect information, judgments must be made about the severity of risk for each hazard, and whether or not the community is willing to accept that risk in light of the known information.
Because disaster managers do not work in a vac- uum, many factors, be they political, social, or eco- nomic, influence the collective determination of what risks are acceptable and what risks are not. The mech- anisms by which they can begin to determine such categorization are explained below.
The disaster managers have thus far identified the risks affecting the country or community, analyzed them individually, and evaluated them collectively. They are now left with an ordered list of risks that they must consider for treatment. Ideally, they would treat all risks such that nobody would have to worry about them ever again, but that risk-free world scenario is inconceivable despite modern tech- nology and engineering. While most risks can be reduced by some amount, few can be completely eliminated, and rarely if ever do the funds exist to reduce all risks by an amount that is acceptable to everyone in the community. There will never be com- plete satisfaction with the ultimate decisions made by disaster managers, mostly because of differences in perception.
Two factors confounding the acceptability of risks are the benefits associated with certain risks, and the creation of new risks by eliminating existing ones. For instance, to completely eliminate the risk from nuclear power generation plants, they would need to be dis- mantled and taken out of service. The resulting short- age of power would require that fossil-fuel-burning plants increase their production, which in turn would create increased carbon-based pollution, which would likewise create increased health and environmental risks.
ALTERNATIVES
Derby and Keeney (1981), two risk management experts, write:
138 Introduction to International Disaster Management
The key aspect of acceptable risk problems is that the solution is found by a decision among alternatives. The generic problem involves choosing the best combination of advantages and disadvantages from among several alterna- tives. The risk associated with the best alternative is safe enough.
This is an important distinction—that risks deemed “acceptable” are not necessarily those with risk levels for which we are “happy.” They continue:
We all would prefer less risk to more risk if all other consequences were held fixed. However, this is never the case. In a situation with no alternatives, then the level of safety associated with the only course of action is by defini- tion acceptable, no matter how disagreeable the situation. Said another way, acceptable risk is the risk associated with the best of available alternatives, not with the best of the alternatives which we would hope to have available.
There are several factors that together influence the determination of risk acceptability. They include personal, political/social, and economic reasons. Although the three are interrelated, different processes drive them. These processes are described next.
PERSONAL
The personal factors that dictate whether a risk would be considered “acceptable” mirror the risk per- ception characteristics described below. For example, a risk whose consequences are “dreaded”, such as the radiation sickness that could result from a meltdown at a nuclear power plant, is likely to be found less acceptable to individual members of the public than the long-term effects of increased solar radiation (such as skin cancer), which may be caused by a decrease in the ozone layer from increased automobile emissions.
The United Nations Development Programme (UNDP) training program in Vulnerability and Risk Assessment (1994) describes the differences in indi- vidual acceptance between risks that are voluntary and involuntary:
Some risks are entered into voluntarily and a distinction is sometimes made between voluntary and involuntary risks. Many recreational activities and sports involve considerable
levels of personal risk entered into voluntarily. Indeed the thrill of the risk is part of the enjoyment of the recreation. The benefits of the risk outweigh the costs and so the per- ception of the risk is reduced; i.e., the threat level that is deemed acceptable is much higher than a risk that is imposed from outside or involuntary.
Other factors that have been shown to affect public acceptance of risk include personal values, gender, ethnicity, education level, and the treatment of the risk by the media.
POLITICAL/SOCIAL
The political/social acceptability of risk is the product of either democratic processes or other col- lective mechanisms of determination. In other words, political and social influences are representations of many personal determinations of acceptability. While it is almost certain that not every individual citizen will be happy with the final decisions made concern- ing a risk’s acceptability and treatment, the choices made will reflect the feelings of the majority if those choices are influenced by political and social acceptability.
Because of the differences in the makeup of differ- ent communities and populations, risk acceptance will not be universal. It is likely to change from place to place, from time to time, and from hazard to hazard (Alesch, 2001). Acceptability is likely to change even within individual communities over time as the makeup of that community changes. It is these differ- ences that make public participation in the disaster management process important.
ECONOMIC
Because countries or communities can rarely sup- port the level of funding required to mitigate all risks, the risk acceptability decision must be influenced by how much each mitigation alternative would cost and what other possible risk mitigation measures would be offset through funding of a specific mitigation effort.
Chapter 3 Risk and Vulnerability 139
In general, disaster managers will have to address the costs of reducing a risk in terms of the benefits (actual risk reduction) that would result. Some com- munities have chosen to simply live with a risk because the costs of mitigating its consequences are prohibitive, and eliminating the risk is unthinkable. For a simplified example, consider the use of the auto- mobile, which highlights the cost/benefit scenario. At present, over a million road traffic fatalities occur throughout the world each year. This obviously pre- sents a great risk. With increased cost, car manufac- turers could easily make their cars much safer, and these fatality rates could be reduced significantly. However, such a cost would make automobiles too expensive for the average consumer. Thus, we accept the loss of over a million lives per year for the benefit of having affordable cars. Even if manufacturers spent the money to make cars completely “safe” for occu- pants, however, there would still be an inherent risk associated, as indicated by the great number of fatali- ties that are caused by pedestrians who are struck by cars (shown in Figure 3-4). The cost of totally elimi-
nating this particular risk associated with automobiles is inconceivable.
W. Kip Viscusi, in the article “Economic Founda- tions of the Current Regulatory Reform Efforts” (1996), describes how the economics of an accept- ability decision can be influenced by the political and social aspects of that decision. To illustrate his point, he produced a list of risk-reducing regulations that fail a cost/benefit “test” (cost is greater than the benefit), and a list of risk-reducing regulations that pass a cost/benefit “test” (benefit is greater than the cost). His results are shown in Table 3-5 and Table 3-6.
“Injustices” are commonly seen in the disaster management decision-making process, especially concerning the treatment and acceptability of hazard risks (MPPP, 1999). The following are three criticisms of the processes by which risk acceptability is determined:
1. Those with money and vested interests can influ- ence the process of determining the acceptabil- ity of risk. Because the process of determining
140 Introduction to International Disaster Management
28.3 26.4
19 18.6 18.5 17.4
16.2 14.8
12 11
0
5
10
15
20
25
30
Af ric
a LM
I
Ea st er
n M
ed ite
rra ne
an L
M I
Ea st er
n M
ed ite
rra ne
an H
I
So ut
he as
t A sia
LM I
W es
te rn
P ac
ifi c LM
I
Eu ro
pe L
M I
Am er
ic as
LM I
Am er
ic as
HI
W es
te rn
P ac
ific HI
Eu ro
pe HI
Region HI = High Income Countries
LMI = Low/Middle Income Countries
F a
ta li
ti e
s p
e r
1 0
0 ,0
0 0
P e
o p
le
FIGURE 3-4 Worldwide road traffic fatalities. (Source: World Health Organization, 1999.)
risk acceptability (including mitigation spend- ing and regulatory practices) is influenced by politics and may be shaped by political ideol- ogy, it is possible for corporate or interest groups to lobby and influence those decisions. This can be seen with hazards such as handguns and assault rifles, environmental degradation, soil and water pollution, or construction in haz- ardous areas. Increased citizen participation in the process can decrease this type of injustice. By increasing the decision-making power of the
general public, a more democratic outcome is possible (though not guaranteed).
2. Setting a dollar figure (in cost-benefit analyses) on a human life is unethical and uncon- scionable. This is primarily a factor related to involuntary risks. To the individuals whose lives are being placed at risk, any dollar figure will seem low or inappropriate as a tradeoff for the acceptance of the risk. Many people would (understandably) feel that their life is too great a price to pay for the existence of any involuntary
Chapter 3 Risk and Vulnerability 141
TABLE 3-5 The cost of risk-reducing regulations that fail a benefit cost test per life saved
Annual Cost per life saved Regulation Initial annual risk lives saved (millions of $)
Grain dust 2.1 in 10,000 4 5.3
Radionuclides/uranium mines 1.4 in 10,000 1.1 6.9
Benzene 8.8 in 10,000 3.8 17.10
Arsenic/glass plant 8.0 in 10,000 0.110 19.20
Ethylene oxide 4.4 in 100,000 2.8 25.60
Arsenic/copper smelter 9.0 in 10,000 0.060 26.50
Uranium mill tailings (inactive) 4.3 in 10,000 2.1 27.60
Uranium mill tailings (active) 4.3 in 10,000 2.1 53.00
Asbestos (OSHA, 1986) 6.7 in 100,000 74.7 89.30
Asbestos (EPA, 1989) 2.9 in 100,000 10 104.20
Arsenic/glass manufacturing 3.8 in 100,000 0.25 142.00
Benzene/storage 6.0 in 10,000,000 0.043 202.00
Radionuclides/DOE facilities 4.3 in 1,000,000 0.001 210.00
Radionuclides/elemental 1.4 in 100,000 0.046 270.00 phosphorous
Benzene/ethylbenzenol styrene 2.0 in 1,000,000 0.006 483.00
Arsenic/low-arsenic copper 2.6 in 10,000 0.09 764.00
Benzene/maleic anhydride 1.1 in 1,000,000 0.029 820.00
Land disposal 2.3 in 100,000,000 2.52 3,500.00
EDB 2.5 in 10,000 0.002 15,600.00
Formaldehyde 6.8 in 10,000,000 0.010 72,000.00
Viscusi (1996) assumes that $2.8 million per life saved is an acceptable cost. Any cost greater than $2.8 million per life fails the cost/benefit test.
Source: Viscusi, 1996.
risk. The cognitive processes that dictate these “price of a human life” determinations are often different for voluntary risks. As the automobile safety example illustrates, people are willing to accept a certain increase in risk to their own lives for the benefit of more-affordable prod- ucts. How much more affordable differs by per- son. But, as shown by relatively recent lawsuits against tobacco companies by smokers who became ill people may be unwilling to accept some voluntary risks despite previous knowl- edge about those risks.
Because of the controversial nature of plac- ing a value on life, it is rare that a risk assess- ment study will actually quote a dollar figure for the amount of money that could be saved per human life loss accepted. Postevent studies have calculated the dollar figures spent per life during a crisis, but to speculate on how much a
company or government is willing to spend to save or risk a life would be extremely unpalat- able for most.
3. Risk management is usually an undemocratic process, as those who may be harmed are not identified or asked if the danger is acceptable to them. It is not difficult to call to memory a case in which a vulnerable or disadvantaged group of people was exposed to a risk whose benefits were enjoyed by others. Many toxic waste dumps are located in impoverished parts of towns, cities, and states, although the people in those communities had little say in deciding the location of such materials. Related to this injus- tice is the reality that the impoverished are usu- ally less able to avoid such risks, as the property or jobs available to them are often associated with these very same risks. It is often the poor who must live in the highest risk areas of a
142 Introduction to International Disaster Management
TABLE 3-6 The cost of risk-reducing regulations that pass a benefit cost test per life saved
Annual Cost per life saved Regulation Initial annual risk lives saved (millions of $)
Unvented space heaters 2.7 in 100,000 63 .1
Oil and gas well service 1.1 in 1,000 50 .1
Cabin fire protection 6.5 in 100,000,000 15 .2
Passive restraints/belts 9.1 in 100,000 1,850 .3
Underground construction 1.6 in 1,000 8.1 .3
Alcohol and drug control 1.8 in 1,000,000 4.2 .2
Servicing wheel rims 1.4 in 100,000 2.3 .2
Seat cushion flammability 1.6 in 10,000,000 37 .6
Floor emergency lighting 2.2 in 100,000,000 5 .7
Crane-suspended personnel 1.8 in 1,000 5 1.2 platform
Concrete and masonry 1.4 in 100,000 6.5 1.4 construction
Hazard communication 4 in 100,000 200 1.8
Benzene/fugitive emissions 2.1 in 100,000 .310 2.8
Source: Viscusi, 1996.
floodplain, or under high-tension power lines, or along highways. These people bear a larger share of the population risk, while many others enjoy much lower risk levels from those partic- ular hazards, even though they enjoy a dispro- portionate amount of the benefits. Thus, risk communication and public participation are important to counteract these injustices.
In determining the treatment of risks in a country or community, disaster managers must consider each hazard according to its current risk level, and deter- mine if the risk is too great to be left as is. If it is deter- mined to be too great, they must analyze what can be done to reduce the risk, and then make another determination as to the acceptability of the new risk level.
Several methods for determining the acceptability of risks have been developed in the past, and are used to varying degrees (dependent upon the needs of those performing the risk evaluation). They include:
● The “no go” alternative. This alternative, which is not always available, is the complete elimina- tion of the risk. Such action can be easier with technological hazards, especially those that are new. How easy depends on how dependent soci- ety has become on the technology in question. For example, when DDT was found to be bioac- cumulating in birds and mammals and was feared to eventually lead to a “silent spring” (a “silent spring,” as described by Rachel Carson, is what would result if DDT were used to the extent that all birds died as a result), the chemical was banned from use. There were alternatives to DDT, and while they may not have been as cost efficient or effective, they were not perceived as being as harmful. For some countries, the more expensive alternatives were acceptable, while in others DDT is still the preferred, cheap option.
However, with hazards that have established a unique niche in society, such as the automobile, elim- inating the risk is close to impossible. Eliminating risks is often only possible with the existence of viable
alternatives. The possibility of eliminating the risk must always be considered in the assessment. (Because the option is to eliminate the risk, and not the hazard, natural disasters can be considered for this option—if either the consequences or the frequency is lowered to zero, the risk becomes zero. However, this option is rarely possible given economic and techno- logical constraints.) The emergence of hybrid cars that rely on a combination of gasoline and electric power is a sign of movement toward a viable alternative in terms of fossil-fuel dependence.
● Accept the risk. A second option is to simply accept the risk as it is—to do nothing. Certain risks may be so low that the money spent to reduce them would be better spent to treat a more severe hazard. In risk matrices, the risks that fall within the lowest category of both consequence and likelihood are generally the risks that are considered acceptable. After all other risks have been treated to the satisfaction of the hazards risk management team, the low risks can be revisited.
● Establish a “de minimis risk” level. De minimis risk dictates that a level of statistical risk for haz- ards exists, below which people need not concern themselves. This level is often set at either 1 in 100,000 or 1 in 1,000,000, and is set either for a one-year period, or for a lifetime (70 years). The term de minimis is a shortened version of the Latin phrase de minimis non curat lex, which means “the law does not care about very small matters.” This concept is widely used throughout Europe to set guidelines for acceptable levels of risk exposure to the general population. An example of its use in the United States includes a regulation de minimis risk set by the Environ- mental Protection Agency for human lifetime risk from pesticides of 1 in 1,000,000 over a 70- year lifetime (PMEP, 1997).
De minimis does not seek to prohibit any risk above the levels set. The theory only states that, if a risk falls below that level, no resources need to be spent on its prevention. If a product poses less risk than the de minimis level, for example, then it should
Chapter 3 Risk and Vulnerability 143
be authorized for production and/or distribution. However, if the risk associated with a product does not fall below the de minimis level, then risk man- agers need to assess if anything can be done to reduce its risk and if the costs outweigh the benefits, among many other issues.
Proponents for de minimis feel that governments can avoid wasting their time trying to increase the safety of risks already satisfying de minimis require- ments, thus freeing them up to spend their resources on other risks of greater concern. Opponents are con- cerned that some risks exist for which even a 1 in 1,000,000 risk would be too high (Mumpower, 1986). One of their contentions is that risks that affect huge populations would result in a high number of deaths even though the risk is so “low.” The smallpox vac- cine, for example, has a 1 in 1,000,000 risk of death. However, if the entire world population were to be vaccinated, approximately 6000 fatalities would ensure. A third group feels that the de minimis strat- egy is effective only if there are two de minimis levels working in conjunction—one that measures absolute risk (1 in 1,000,000 for example), and another that sets the maximum number of allowable expected fatalities (X number of fatalities for country Y, for example).
● Establish a “de manifestis risk” level. Related to de minimis risk is the concept of de manifestis risk, or “obnoxious risk.” With de manifestis risk, there is a risk level above which mitigation is mandatory. In practice, this level is generally set at 1 in 10,000 per vulnerable individual. This practice is often cited in regards to secondhand smoke exposure in the workplace (Repace Asso- ciates, 1999).
● Perform cost-benefit analyses of risks. Cost- benefit, or benefit-cost, analyses are probably the most widely used and widely accepted method by which risks and alternatives are evaluated for acceptability. The Massachusetts Precautionary Principle Project (1999) writes:
[Cost-Benefit Analyses are] where the risks reduced by taking a protective action (like imposing a stricter
regulation on emissions) are equated to benefits (such as a life saved or reduced health costs.) The “benefit” is then compared to the estimated “costs” of imple- menting the protective action (cost to the industry to install better pollution controls). Often a determina- tion is made as to how much “cost” it is worth to save that life, usually 2 million dollars.
If the cost of controls greatly exceeds the cost of the life saved, regulatory actions may not be taken. Among other flaws, cost-benefit analysis fails to consider who reaps the benefits and who assumes the cost. It also perpetuates the myth that we must decide between economic growth and environmental protection. Cost benefit analysis is also heavily biased towards costs of regulation today, discounting less quantifiable costs such as health damage and benefits of prevention. Cost benefit analysis often overestimates the costs of regulation. It also tries to quantify the unquantifiable, or translate the non-economic, i.e., namely pain and suffering, ill- ness, and disease, into money. Many consider this unethical.
Following the September 11th terrorist attacks, in which hijacked commercial airplanes were used as weapons, considerable effort went into (and continues to go into) securing airways around the world. As security measures increase, so does the cost of ensur- ing that security, and most of this cost is passed along to the consumer. Questions that require people to con- sider the financial cost of their own safety are often used to determine individual risk-seeking or risk- averse behavior.
Related to cost-benefit decisions are cost-effective- ness decisions. In the case of cost-effectiveness deci- sions, the minimum “unit cost” to reduce maximum risk is favored in considering the alternatives for risk mitigation within and between risks.
● Acceptable risk as the best choice among alter- natives. Derby and Keeney (1981) write that “The answer to ‘How safe is safe enough?’ depends upon [five steps]. . . . Acceptable risk is determined by what alternatives are available, what objectives must be achieved, the possible consequences of the alternatives, and the values to be used.” The five steps they are referring to are:
144 Introduction to International Disaster Management
1. Define the alternatives 2. Specify the objectives and measures of effec-
tiveness to indicate the degree to which they are achieved
3. Identify the possible consequences of each alternative
4. Quantify the values for the various consequences
5. Analyze the alternatives to select the best choice.
● Disaster managers will have already completed most of these steps by the time they are deciding which risks to treat. Derby and Keeney provide graphical illustrations of four factors that influ- ence how risk alternatives are chosen and deter- mined to be acceptable. These examples are shown in Figures 3-5 to 3-8 and are discussed in the following.
In Example A, it is assumed that the benefits of all the alternatives are equal. The differences are only in their financial cost and the level of risk (with 0 being the optimal level for both cost and risk). If only alter- natives K and L are available, then the choice is
between high cost with low risk and low cost with high risk. The acceptable risk would be the level of risk associated with the particular alternative chosen, either K or L.
If another alternative, M, were introduced into the problem, then M with lower cost and lower risk would be preferred to either K or L. Consequently, accept- able risk is now the safety level of alternative M. This risk is different from the level associated with the other alternatives. Clearly, the appropriate level of risk depends on the alternatives available.
Example B shows how acceptable risk changes with what objectives are achieved. In this example, only alternatives K and L are (known to be) available. If the sole objective is to minimize the risk, alternative K would be chosen. The acceptable risk would then be the risk level associated with K. On the other hand, if the sole objective is to minimize the cost, the alterna- tive L would be chosen. Acceptable risk under this objective would be the risk level for L. Each objective leads to choosing different alternatives. In each case, the acceptable risk changes with the objective used to make the choice.
Example C shows how new information can change the determination of what is considered
Chapter 3 Risk and Vulnerability 145
COST
K
L
RISK
*M
0 0
FIGURE 3-5 Risk acceptability Example A. (Source: Derby and Keeney, 1981.)
COST
RISK
Minimum Risk
Minimum Cost
K
L
0 0
FIGURE 3-6 Risk acceptability Example B. (Source: Derby and Keeney, 1981.)
acceptable risk. In this example, we assume that alternative M determines the acceptable risk, as in Example A. However, additional information pro- vided by experience, research, development, or analy- sis reveals that the initial assessment of alternative M must be revised. Instead of confirming that M has lower cost and lower risk than both alternatives K and L, the new information shows that M has both the high cost of K and the high risk of L. The acceptable risk is now determined by the choice between K and L.
Example D illustrates the effect of values and pref- erences on the choice between alternatives. In this example, different preferences for trading off in- creased cost for lower risk are represented by the two curves. In Case 1, the trade-off curve reflects the willingness to incur large costs to reduce risk by small amounts. Alternative K is the most attractive choice with this preference. In Case 2, the trade-off curve reflects less of a willingness to increase costs in exchange for specific reductions in risk. This prefer- ence selects alternative L as the best choice. Since acceptable risk is determined by the choice between the two alternatives, these different preferences change what is considered acceptable.
146 Introduction to International Disaster Management
COST
K M
L
RISK0 0
FIGURE 3-7 Risk acceptability Example C. (Source: Derby and Keeney, 1981.)
Case 1
Case 2 L
K
COST
RISK0 0
- Intro-to-Int-Disaster-Manangement 139
- Intro-to-Int-Disaster-Manangement 140
- Intro-to-Int-Disaster-Manangement 141
- Intro-to-Int-Disaster-Manangement 142
- Intro-to-Int-Disaster-Manangement 143
- Intro-to-Int-Disaster-Manangement 144
- Intro-to-Int-Disaster-Manangement 145
- Intro-to-Int-Disaster-Manangement 146
- Intro-to-Int-Disaster-Manangement 147
- Intro-to-Int-Disaster-Manangement 148
- Intro-to-Int-Disaster-Manangement 149
- Intro-to-Int-Disaster-Manangement 150
- Intro-to-Int-Disaster-Manangement 151
- Intro-to-Int-Disaster-Manangement 152
- Intro-to-Int-Disaster-Manangement 153
- Intro-to-Int-Disaster-Manangement 154
- Intro-to-Int-Disaster-Manangement 155
- Intro-to-Int-Disaster-Manangement 156
- Intro-to-Int-Disaster-Manangement 157
- Intro-to-Int-Disaster-Manangement 158
- Intro-to-Int-Disaster-Manangement 159
- Intro-to-Int-Disaster-Manangement 160
- Intro-to-Int-Disaster-Manangement 161
- Intro-to-Int-Disaster-Manangement 162
- Intro-to-Int-Disaster-Manangement 163
- Intro-to-Int-Disaster-Manangement 164
- Intro-to-Int-Disaster-Manangement 165
- Intro-to-Int-Disaster-Manangement 166
- Intro-to-Int-Disaster-Manangement 167
- Intro-to-Int-Disaster-Manangement 168
- Intro-to-Int-Disaster-Manangement 169
- Intro-to-Int-Disaster-Manangement 170
- Intro-to-Int-Disaster-Manangement 171
- Intro-to-Int-Disaster-Manangement 172
