Risk Assessment of Extreme Events

1
Risk Assessment of Extreme Events

• Rae Zimmerman
• (New York University)
• Vicki M. Bier
• (University of Wisconsin-Madison)

2
I. Introduction and Scope

• Risk assessment is a means to characterize and
  reduce uncertainty to support our ability to deal
  with catastrophe
• Scope of this paper
• Application of risk assessment to both the built
  and natural environments under extreme events
• Understanding and management of human health,
  safety, and security

3
I. Introduction and Scope (cont.)

• Modern risk assessment for engineering began with
  Reactor Safety Study (1975)
• Applications to engineered systems and
  infrastructure are common
• Applications to chemical risks under dozens of
  federal environmental statutes
• E.g., drinking water, ambient water quality, and
  air quality standards
• Review and renewal of pesticide applications
• Levels of site cleanup under Superfund

4
II. What is Risk Assessment?

• Definition of risk assessment
• A systematic approach to organizing and
  analyzing scientific knowledge and information
  for potentially hazardous activities or for
  substances that might pose risks under specified
  circumstances
• National Research Council (NRC), 1994

5
II.A Definitions of Risk

• Both uncertainty and some kind of loss or
  damage (Kaplan and Garrick 1981)
• The potential for realization of unwanted,
  negative consequences of an event (Rowe 1976)
• The probability per unit time of the occurrence
  of a unit cost burden (Sage and White 1980)
• The likelihood that a vulnerability will be
  exploited (NRC 2002)

6
II.A Definitions of Risk (cont.)

• Terms to characterize acceptable risk in health
  and safety legislation
• Adequate
• Imminent
• Substantial
• Reasonable (vs. unreasonable)
• Posing grave danger
• At a zero level
• Significant (vs. de minimus)
• An ample or adequate margin of safety

7
II.B Relationship of Risk to Other Concepts

• Merriam-Websters Collegiate Dictionary 2002
• Hazard (a source of danger)  
• Catastrophe (a momentous tragic event)
• Chronic (long duration or frequent recurrence)
  
• NRC 2002 Threat (an adversary)
• Vulnerability (an error or a weakness)
• Extreme events (low frequency and high severity)
• Counter-expected events (believed to be unlikely)
• Unexpected events (not even anticipated)
• Uncertainty (lack of knowledge)
• Variability (differences among a population)

8
II.C Paradigms for Risk Assessment

• A form of systems analysis
• Answers three questions (Kaplan and Garrick
  1981)
• What can go wrong?
• How likely is it that that will happen?
• If it does happen, what are the consequences?
• Several integrated risk assessment/risk
  management frameworks have been proposed

9
II.C Paradigms for Risk Assessment (cont.)

• Deliberation frames analysis and analysis
  informs deliberation (Stern and Fineberg 1996)
• The combination of these two steps is termed the
  analytic-deliberative process
• An iterative process
• Deliberation and analysis are viewed as
  complementary

10
III.A Health Risk Assessment

• Hazard identification
• Risk estimation
• Exposure assessment
• Dose/response relationships (toxicity assessment)
• Risk characterization or risk calculation

11
III.A Health Risk Assessment (cont.)

• Hazard identification
• Structure activity relationships (structural
  toxicology)
• Case clusters
• Epidemiological studies
• Experimental chemical tests on lower order
  organisms (rapid screening)
• Animal tests

12
III.A Health Risk Assessment (cont.)

• Exposure assessment
• Sources, pathways, and sinks (or receptors)
• Health effects assessment

13
III.A Health Risk Assessment (cont.)

• Sources, pathways, and sinks (receptors)
• Source characterization (substances released,
  rates of release, temporal variations, location)
• Fate and transport
• Routes or pathways of exposure from environmental
  end points to human organisms
• Size, type, and sensitivity of population at risk 

14
III.A Health Risk Assessment (cont.)

• Health effects assessment
• Dose estimates or intake levels
• Absorption by the body
• General toxicity of the risk agent in the body
  (e.g., target organs, types of effects)
• State of health of the organism

15
III.A Health Risk Assessment (cont.)

• Dose/response relationships (toxicity
  assessment)
• Dose/response models
• Empirical relationships between levels of
  exposure and effects

16
III.A Health Risk Assessment (cont.)

• Risk characterization or calculation
• Risk estimate
• Characterization of uncertainties, assumptions,
  and data quality

17
IIIB Engineering Risk Assessment

• Hazard identification 
• Assessment of accident occurrence frequencies  
• Consequence analysis 
• Risk characterization
• Uncertainty analysis

18
III.B Engineering Risk Assessment (cont.)

• Hazard identification
• System familiarization
• Hazard and operability studies
• Failure modes and effects analysis

19
III.B Engineering Risk Assessment (cont.)

• Assessment of accident occurrence frequencies

20
III.B Engineering Risk Assessment (cont.)

• Consequence analysis has two stages
• Migration of hazardous materials from sources to
  sinks
• Consequences of those materials for public health
  and safety
• Relevant consequence measures include
• Structural response of a building
• Costs of property damage, loss of use, repair
• Amount of hazardous material released
• Numbers of fatalities or other health effects

21
III.B Engineering Risk Assessment (cont.)

• Risk characterization
• Results presented graphically
• Probability distribution, complementary
  cumulative

22
III.C Spatial Dimensions

• Proximity is a key factor in the exposure portion
  of the risk equation
• Proximity can also affect
• Perceived severity of particular scenarios
• Conditional failure probabilities

23
III.C Spatial Dimensions

• Despite this, risk analyses rarely use
  sophisticated spatial concepts or models
• Methodology for doing so tends to be ad hoc
• Takes little advantage of GIS systems

24
IV. Understanding Uncertainty

• Sources of uncertainty
• Statistical variation
• Systematic error
• Subjective judgment
• Linguistic imprecision
• Variability
• Inherent randomness or unpredictability
• Disagreement
• Approximation

25
IV. Understanding Uncertainty (cont.)

• Uncertainty and variability have different
  implications for decision-making (NRC 1994)
• Uncertainty forces decision makers to judge how
  probable it is that risks will be overestimated
  or underestimated
• Variability forces them to cope with the
  certainty that different individuals will be
  subjected to different risks
• Large uncertainty suggests that further research
  may be desirable

26
V. Human Perceptions, Behavior, and Performance

• Evacuation responses in emergencies differ
  substantially from performance in tests and
  simulations
• Behavioral assumptions underlying many building
  codes and strategies are flawed
• Human behavior is extremely variable
• Healthy versus elderly, ill, or disabled
• Familiarity with a particular environment  
• Predicting the behavior of the public is a
  difficult challenge

27
V. Human Perceptions, Behavior, Performance
(cont.)

• Intentional hazards
• Estimating the likelihood and nature of
  intentional attacks is needed for intelligent
  benefit-cost analysis (Woo 2002)
• Protection from an adversary is different than
  protection against accidents
• Adversaries can choose to attack targets that
  have not been hardened
• Defensive measures may be less effective if they
  are known
• Optimal strategy depends on attacker behavior

28
VI. World Trade Center Disaster

• Unexpected or counter-expected
• Past experiences could have helped to identify
  risk of an attack (Barnett 2001)
• Lots of eventscould be interpreted as
  precursors of the calamity
• All the elements of the Sept. 11 catastrophe
  had historical precedent
• This points out the need for
• Methods of learning from past experience
• Vigilance to signs of problems

29
VII. Conclusions

• Risk assessment is a vital tool for dealing with
  extreme events
• Capabilities of risk assessment are challenged
  when we attempt to apply it to extreme and
  unanticipated events
• Need for methodological improvements to more
  fully incorporate
• Spatial dimensions
• Human values, attitudes, beliefs, and behavior
• Past experience

30
Acknowledgments

• This material is based upon work supported in
  part by
• The U.S. Army Research Laboratory and the U.S.
  Army Research Office under grant number
  DAAD19-01-1-0502
• The National Science Foundation under Cooperative
  Agreement No. CMS-9728805
• Any opinions, findings, conclusions, or
  recommendations expressed in this document are
  those of the authors