The devastation from Hurricane Harvey in 2017 and the continuing losses from hurricanes and tropical storms, such as Florence in 2018, challenge risk analysts to consider how to use risk analysis methods to help societies better anticipate, prepare for, respond to, and recover from natural disasters.  Such disasters cause massive destruction to lives and livelihoods, critical infrastructures, real estate, and regional economies.  A primary goal for risk analysis of natural disasters is to help reduce this toll by showing how to apply risk assessment, risk management, and risk communication to allocate resources as wisely as we can to reduce losses from these low-frequency, high-consequence events before, during, and after their occurrence.

This special virtual issue of Risk Analysis, prepared a year after Harvey and in the wake of Florence, offers insights on how best to anticipate, communicate, manage, mitigate, respond to, recover from, and learn from such disasters.  The issue consists of papers drawn from past issues of Risk Analysis, with a deliberate bias toward the most recent and up-to-date papers.  It is an update of a virtual issue originally prepared following Hurricane Harvey, updated in September of 2018 to provide the most recent advances to readers and researchers.  The organization and contents of the issue are as follows.

Part 1. Predicting hurricane and flood occurrence and damage

The issue begins with a paper by Highfield et al. (2012) that highlights the limitations of the traditional hundred-year floodplain for informing property owners and local decision-makers about where heavy losses to life and property are most likely to occur and where investments in prevention and resilience are most likely to generate high returns.  They illustrate the limitations of floodplain information with data from Harris County, Texas and recommend steps to more accurately account for the high proportion of loss that occurs outside official floodplains.  Similarly, for two other counties in Texas, Travis and Galveston counties, Czajkowski, Kunreuther, and Michel-Kerjan find that residences in the same FEMA flood zones have flood risks that differ substantially (more than two-fold), while residences classified as belonging to different FEMA flood zones may have almost the same risks.  Storm surge risks are also substantial outside FEMA designated storm surge zones, suggesting that zone-based risk classification systems have strong limitations for informing flood insurance purchasing and risk mitigation investment decisions. 

Schneeberger et al. (2017) present a flood risk analysis model that takes into account spatial heterogeneity of flood events and that quantifies both average annual damages and impacts of low-probability, high-impact floods.  Wu et al. (2014) provide evidence from experiments with students on how people form judgments of hurricane strike probabilities at different locations from visually presented information on current tracks and uncertainty cones.  McRoberts et al. (2016) focus specifically on predicting locations, number, and magnitudes of power outages, adding to a series of increasingly powerful predictive models developed by Guikema and co-workers.   Iman et al. (2005) present uncertainty and sensitivity analysis for loss projections from a hurricane wind field model.  Such loss prediction models can be and have been improved via advances in machine learning and predictive analytics such as those illustrated in the paper of McRoberts et al. Finally, Terti et al. (2017) apply machine learning algorithms (a random forest classifier) to forecast vehicle-related human losses from flash floods, taking into account both physical and social dynamics factors and characteristics of the built environment.  The predictive performance of the loss forecasting method is demonstrated using data from the catastrophic flash floods of May 2015 in Texas and Oklahoma as a case study.

Part 2. Risk perceptions and preparation for hurricane and flood risks 

How people perceive and respond to risks of natural disasters in light of media reports, official advisories and warnings, past personal experience, and discussions with friends and in social media has been much studied in Risk Analysis in recent years.  We have selected the following handful of papers from a much larger number of insightful contributions that shed light on the dynamics of risk perception and behavior as disasters unfold.  Terpstra (2011) develops a path model for the emotional and cognitive drivers of citizen intentions to prepare for floods and points out the challenge of communicating flood risk effectively at the emotional level to people who have not experienced floods.Trumbo et al. (2014) use survey data from Gulf Coast residents interviewed shortly after Hurricanes Katrina and Rita in 2005 and again two years later to quantify how perceived risk fades and optimism bias for hurricane evacuation increases in the years following a hurricane.  Ge et al. (2011) analyze data from households in Florida to identify psychological drivers of expected participation in hurricane hazard mitigation incentive programs, finding that repeated reminders of the likelihood of severe losses from hurricanes may be necessary to increase participation.  Haer et al (2017) apply agent-based models (ABMs) of household decision-making about insurance purchases and protective investments under uncertainty about flood risks.  They find that expected utility (rational) decision models leading to less investment and higher risks, but also greater responsiveness to insurance premium discounts, than more psychologically realistic decision models in which low-probability/high-consequence events are over-weighted.  Predicted long-term flood risk depends on how social interaction, media influence, and experiences with floods affect household risk perceptions. 

Part 3. Flood insurance policy and behaviors

A key societal defense against the financial havoc wrought by hurricanes and floods is flood insurance, which can help property owners recover in the aftermath of a flood.  Yet, as shown by Hurricane Harvey and other major hurricanes, many affected homeowners and businesses may be uninsured or underinsured.  Epple and Lave (1988) proposed a model three decades ago in Risk Analysis that analyzes the incentives leading rational economic agents to underinsurance and individual decisions that produce collectively economically inefficient outcomes.  They also considered conditions under which risk-informed individual decisions could lead to collectively efficient outcomes, and noted that zoning floodplains would typically not suffice for this purpose.  Fan and Davlasheridze estimate how the National Flood Insurance Program's Community Rating System (CRS) affects individual choices of residential location choices, finding that age, ethnicity, education, and prior exposure to risk explain risk perception and that water amenities dominate flood risk, leading people to move into areas with high flood risks.  Gao et al. (2014) introduce a game-theoretic model of insurer decisions that interacts with a homeowner decision model and a regional catastrophic loss model and use it to study how hurricane risks to residential buildings in North Carolina are affected by the decisions of primary insurers, homeowners with and without insurance, and reinsurers.  Finally, Kousky (2016) examines flood insurance policies for Atlantic and Gulf coast states of the United States between 2001 and 2010 and notes that, historically, being hit by at least one hurricane in the previous year increases net flood insurance purchases by 7.2%, but that this effect fades away within three years.  A plausible explanation is that insurance policies must be purchased to receive disaster aid; when this requirement is adjusted for, it appears that hurricanes increase voluntary insurance purchases by only 1.5%.

Part 4. Evacuation behavior and adaptive decision-making around floods

How do people respond to the prospect of a hurricane or flood that may or may not affect them?  Meyer et al. (2012) use web-based simulations of news and neighbor opinions to study how stated intentions to take protective action evolve as a simulated hurricane approaches.  They find that displays that show most likely tracks stimulate preparation more effectively than showing uncertainty cones and that the patterns of information use in their lab simulations correlate well with observations from field studies.  For the Houston-Galveston and Miami-Dade areas, Lazo et al. (2015) study how stated intentions to evacuate when presented with either (a) a forecast that a hurricane will strike where one lives or (b) an evacuation order vary with factors such as age, world view, and sources of information.  Among their findings are that older people and people who rely on personal sources rather than public sources of information are less likely to heed evacuation orders but more likely to respond to forecasts that their areas will be hit.  Thompson et al. (2017) provide a systematic review of the literature on evacuation from natural disasters, highlighting the importance of risk perception, demographic factors, previous experience, and having an evacuation plan.  For longer time scales, Koerth et al. (2017) review the literature on coastal household-level adaptations to flood risks and how adaptation activities and intentions depend on socioeconomic variables, cognitive variables, experience, and perceptions of who is or should be responsible for risk management adaptation activities, e.g., government vs. individuals.   Davidson et al. (2018) introduce and illustrate the practical application of an innovative integrated scenario‐based evacuation framework for supporting hurricane evacuation decisions.  The framework models the uncertain dynamics of human responses to unfolding events and represents major uncertainties via an ensemble of probabilistic scenarios.  Decision recommendations are optimized via a multistage stochastic programming model that minimizes risk and travel times he framework is illustrated with a case study for Hurricane Isabel (2003) in eastern North Carolina.  Blanon et al. (2018) further develop the ensemble-based approach to uncertainty analysis, combining meteorological, hydrologic model, and storm surge, tide, and wind-wave models to bracket inundation, river runoff, and wind hazards and to provide plausible best- and worst-case scenarios for risk-based evacuation decisions.

Part 5. Economic and societal consequences of floods and hurricanes

Hurricanes and floods have widespread impacts on life, health, wellbeing, and property that unfold on multiple time scales.  To help understand the magnitudes of different consequences over time, Part 5 of this issue examines impacts of hurricane-related floods on deaths, hazardous material releases, mental health, and economic losses.  Part 6 further examines the dynamics of economic impacts and recovery.

Jonkman et al. (2009) examine the death toll from flooding in New Orleans after Hurricane Katrina, reporting that nearly 60% of the roughly 1100 fatalities occurred among people over age 65.  Santella et al. (2010) describe the causes, effects, and frequency of over 200 onshore releases of hazardous chemicals, petroleum, and natural gas during Hurricane Katrina and consider risk-based facility designs and improvements that could reduce the risks of such releases in future.  As seen from chemical plant fires and toxic releases during Hurricane Harvey, there is still substantial potential for these risks during large hurricanes and floods.  Zahran et al. (2011) examine mental health resilience following Hurricanes Katrina and Rita and conclude that single mothers had productivity loss from post-disaster stress and disability estimated at over $130 million. Koks et al. model the dynamics of economic recovery following different hypothetical flood events in the port area or Rotterdam, using an input-output model of interactions among economic sectors, and conclude that high-probability and low-probability flood events have qualitatively different effects on scale of damage and time to full recovery.   

Part 6. Recovery, resilience, adaptation, and learning from natural disasters

Li et al. (2013) also apply input-output modeling to predict how the regional economy around London would recover from a disastrous flood under different management plans.  They estimate that full recovery could take close to six years, but that proportional rationing designed to facilitate reconstruction, e.g., by promoting transportation recovery and healthcare as preconditions for recovery in other sectors, can speed the path to full recovery.  Woodward et al. (2014) also consider flooding risks in London, but from the perspective of optimizing the selection and timing of investments in risk-reducing measures before a flood occurs to reduce the likelihoods and sizes of potential losses.  They apply key concepts from real options theory and multiobjective optimization to optimize investment strategies, taking into account the value of future flexibility and adaptive decision-making as more information becomes available.   

This issue closes with a paper by Berner et al. (2017) that describes how to use verified and validated simulation models to identify potential “black swan” events – previously unthought-of combinations of events that could lead to much larger losses than any previously experienced.  They show how to use this simulation-based approach to identify conditions leading to potentially very large losses in electric power networks during hurricanes.  Identifying such potential black-swan events before they occur reduces the potential for a negative surprise and can help planners to identify defensive actions to prevent or reduce extremely large losses and to optimize allocations of resources to facilitate infrastructure resilience and recovery following a disaster.  

Further Readings

The papers in this special virtual issue are a small sample of a large body of useful papers that have been published in Risk Analysis on hurricane and flood risks and how best to assess, communicate, and manage them.For further reading, we recommend starting with the Special Issue on the Risk of Extreme and Catastrophic Events, edited by Yacov Haimes and Jim Lambert, published in November of 2012.  The references in the articles in this issue and in that collection provide excellent points of entry into the relevant literatures in each of the six areas just described.  

The Editors of Risk Analysis join the publisher, Wiley, in hoping that this free virtual issue will be useful to risk analysts and other researchers and practitioners in areas and disciplines that overlap with risk analysis of natural disasters in thinking creatively and productively about what else to do to mitigate the risks from future hurricanes and floods, as well as other natural disasters.  The articles in this issue offer evidence that there is abundant need and opportunity for improvements.  Experience with Hurricane Harvey highlights the importance and urgency of making these advances to reduce the human toll from similar future events.  The Editors welcome future articles that substantially advance our ability to predict, communicate, and optimally reduce hurricane and flood risks and to recover as quickly and well as possible from their devastating consequences when they occur.

Papers Cited, By Section

1. PREDICTING HURRICANE AND FLOOD OCCURRENCE AND DAMAGE

Examining the 100-Year Floodplain as a Metric of Risk, Loss, and Household Adjustment
RISK ANALYSIS
Volume 33, Issue 2, February 2013, Pages: 186–191, Wesley E Highfield, Sarah A Norman and Samuel D Brody

Quantifying Riverine and Storm-Surge Flood Risk by Single-Family Residence: Application to Texas
RISK ANALYSIS
Volume 33, Issue 12, December 2013, Pages: 2092–2110, Jeffrey Czajkowski, Howard Kunreuther and Erwann Michel-Kerjan

A Probabilistic Framework for Risk Analysis of Widespread Flood Events: A Proof-of-Concept Study
RISK ANALYSIS
Klaus Schneeberger, Matthias Huttenlau, Benjamin Winter, Thomas Steinberger, Stefan Achleitner and Johann Stötter

Effects of Track and Threat Information on Judgments of Hurricane Strike Probability
RISK ANALYSIS
Volume 34, Issue 6, June 2014, Pages: 1025–1039, Hao-Che Wu, Michael K. Lindell, Carla S. Prater and Charles D. Samuelson

Improving Hurricane Power Outage Prediction Models Through the Inclusion of Local Environmental Factors
RISK ANALYSIS
D. Brent McRoberts, Steven M. Quiring and Seth D. Guikema

Uncertainty Analysis for Computer Model Projections of Hurricane Losses
RISK ANALYSIS
Volume 25, Issue 5, October 2005, Pages: 1299–1312, Ronald L. Iman, Mark E. Johnson and Charles C. Watson Jr.

Toward Probabilistic Prediction of Flash Flood Human Impacts
RISK ANALYSIS
October 2017, Galateia Terti, Isabelle Ruin, Jonathan J. Gourley, Pierre Kirstetter, Zachary Flamig, Juliette Blanchet, Ami Arthur and Sandrine Anquetin

2. RISK PERCEPTIONS AND PREPARATION FOR HURRICANE AND FLOOD RISKS

Emotions, Trust, and Perceived Risk: Affective and Cognitive Routes to Flood Preparedness Behavior
RISK ANALYSIS
Volume 31, Issue 10, October 2011, Pages: 1658–1675, Teun Terpstra

An Assessment of Change in Risk Perception and Optimistic Bias for Hurricanes Among Gulf Coast Residents
RISK ANALYSIS
Volume 34, Issue 6, June 2014, Pages: 1013–1024, Craig Trumbo, Michelle A. Meyer, Holly Marlatt, Lori Peek and Bridget Morrissey

Florida Households’ Expected Responses to Hurricane Hazard Mitigation Incentives
RISK ANALYSIS
Volume 31, Issue 10, October 2011, Pages: 1676–1691, Yue Ge, Walter Gillis Peacock and Michael K. Lindell

Integrating Household Risk Mitigation Behavior in Flood Risk Analysis: An Agent-Based Model Approach
RISK ANALYSIS
Toon Haer, W. J. Wouter Botzen, Hans de Moel and Jeroen C. J. H. Aerts

3. FLOOD INSURANCE POLICY AND BEHAVIORS

The Role of Insurance in Managing Natural Hazard Risks: Private Versus Social Decisions
RISK ANALYSIS
Volume 8, Issue 3, September 1988, Pages: 421–433, Dennis Epple and Lester B. Lave

Flood Risk, Flood Mitigation, and Location Choice: Evaluating the National Flood Insurance Program's Community Rating System
RISK ANALYSIS
Volume 36, Issue 6, June 2016, Pages: 1125–1147, Qin Fan and Meri Davlasheridze

Modeling Insurer-Homeowner Interactions in Managing Natural Disaster Risk
RISK ANALYSIS
Volume 34, Issue 6, June 2014, Pages: 1040–1055, Yohannes Kesete, Jiazhen Peng, Yang Gao, Xiaojun Shan, Rachel A. Davidson, Linda K. Nozick and Jamie Kruse

Disasters as Learning Experiences or Disasters as Policy Opportunities? Examining Flood Insurance Purchases after Hurricanes
RISK ANALYSIS
Volume 37, Issue 3, March 2017, Pages: 517–530, Carolyn Kousky

4. EVACUATION BEHAVIOR AND DECISION-MAKING AROUND FLOODS

Dynamic Simulation as an Approach to Understanding Hurricane Risk Response: Insights from the Stormview Lab
RISK ANALYSIS
Volume 33, Issue 8, August 2013, Pages: 1532–1552, Robert Meyer, Kenneth Broad, Ben Orlove and Nada Petrovic

Factors Affecting Hurricane Evacuation Intentions
RISK ANALYSIS
Volume 35, Issue 10, October 2015, Pages: 1837–1857, Jeffrey K. Lazo, Ann Bostrom, Rebecca E. Morss, Julie L. Demuth and Heather Lazrus

Evacuation from Natural Disasters: A Systematic Review of the Literature
RISK ANALYSIS
Volume 37, Issue 4, April 2017, Pages: 812–839, Rebecca R. Thompson, Dana Rose Garfin and Roxane Cohen Silver

Household-Level Coastal Adaptation and Its Drivers: A Systematic Case Study Review
RISK ANALYSIS
Volume 37, Issue 4, April 2017, Pages: 629–646, Jana Koerth, Athanasios T. Vafeidis and Jochen Hinkel

An Integrated Scenario Ensemble‐Based Framework for Hurricane Evacuation Modeling: Part 1—Decision Support System
RISK ANALYSIS
March, 2018, Rachel A. Davidson, Linda K. Nozick, Tricia Wachtendorf, Brian Blanton, Brian Colle, Randall L. Kolar, Sarah DeYoung, Kendra M. Dresback, Wenqi Yi, Kun Yang and Nicholas Leonardo

An Integrated Scenario Ensemble‐Based Framework for Hurricane Evacuation Modeling: Part 2—Hazard Modeling
RISK ANALYSIS
April 2018, Brian Blanto, Kendra Dresback, Brian Colle, Randy Kolar, Humberto Vergara, Yang Hong, Nicholas Leonardo, Rachel Davidson, Linda Nozick, Tricia Wachtendorf

5. ECONOMIC AND SOCIETAL CONSEQUENCES OF FLOODS AND HURRICANES

Loss of Life Caused by the Flooding of New Orleans After Hurricane Katrina: Analysis of the Relationship Between Flood Characteristics and Mortality
RISK ANALYSIS
Volume 29, Issue 5, May 2009, Pages: 676–698, Sebastiaan N. Jonkman, Bob Maaskant, Ezra Boyd and Marc Lloyd Levitan

Petroleum and Hazardous Material Releases from Industrial Facilities Associated with Hurricane Katrina
RISK ANALYSIS
Volume 30, Issue 4, April 2010, Pages: 635–649, Nicholas Santella, Laura J. Steinberg and Hatice Sengul

Economics of Disaster Risk, Social Vulnerability, and Mental Health Resilience
RISK ANALYSIS
Volume 31, Issue 7, July 2011, Pages: 1107–1119, Sammy Zahran, Lori Peek, Jeffrey G. Snodgrass, Stephan Weiler and Lynn Hempel

Integrated Direct and Indirect Flood Risk Modeling: Development and Sensitivity Analysis
RISK ANALYSIS
Volume 35, Issue 5, May 2015, Pages: 882–900, E. E. Koks, M. Bočkarjova, H. de Moel and J. C. J. H. Aerts

6. RECOVERY, RESILIENCE, ADAPTATION, AND LEARNING FROM NATURAL DISASTERS

Modeling Imbalanced Economic Recovery Following a Natural Disaster Using Input-Output Analysis
RISK ANALYSIS
Volume 33, Issue 10, October 2013, Pages: 1908–1923, Jun Li, Douglas Crawford-Brown, Mark Syddall and Dabo Guan

Disasters as Learning Experiences or Disasters as Policy Opportunities? Examining Flood Insurance Purchases after Hurricanes
RISK ANALYSIS
Volume 37, Issue 3, March 2017, Pages: 517–530, Carolyn Kousky

Adaptive Flood Risk Management Under Climate Change Uncertainty Using Real Options and Optimization
RISK ANALYSIS
Volume 34, Issue 1, January 2014, Pages: 75–92, Michelle Woodward, Zoran Kapelan and Ben Gouldby

The Use of Simulation to Reduce the Domain of “Black Swans” with Application to Hurricane Impacts to Power Systems
RISK ANALYSIS
Christine L. Berner, Andrea Staid, Roger Flage and Seth D. Guikema