IRCD Researchers Meeting Abstracts

Sudha Arlikatti, University of North Texas
Simon Andrew, University of North Texas
Laura Siebeneck, University of North Texas

Measuring and Comparing Organizations' Resilience in Responding to the 2011 Mega Floods in Thailand

Objectives: The study of organizations during disasters has placed little emphasis on understanding factors contributing to organizations' resilience. This paper provides an overview of research conducted in Thailand after the megafloods of 2011. The objectives are: to map predisaster resiliency at the provincial level; conduct 45 interviews with organizations' responding to the floods within the provinces of Bangkok, Pathumthani, and Ayutthaya; and determine the level of organizations' resilience in three provinces using social networks analysis.

Methods: Guided primarily by Cutter et al.'s Disaster Resilience of Place model, indicators were selected to allow for the measurement and mapping of social, economic, institutional, infrastructural, and community capital of 76 Thai provinces to create a baseline resiliency matrix. Using the Multidisciplinary Center for Earthquake Engineering Research framework, characteristics of robustness, redundancy, resourcefulness, and rapidity of 45 public, private and nonprofit organizations serving a rural, suburban, and urban province were captured through interviews. These were analyzed for recurrent themes as well as using social network analysis to capture resiliency.

Findings: Overall, the baseline resiliency model suggests that provinces with a higher percentage of residents living in urban provinces have higher resilience scores when compared to more rural areas. However qualitative analyses suggest that there are marked differences in resiliency capacities of organizations' at the local level. While all three provinces scored equally high in their self-assessment of resourcefulness, the urban province of Bangkok scored highest in all dimensions. Surprisingly the organizations' serving the rural province of Ayutthaya followed closely. These results are consistent with findings from the Social Network Analysis, which show that organizations' in the rural province were more resilient than the other two.

Conclusions: This paper highlights key findings related to organizations' resilience at the macro and micro-levels in Thai provinces. The difficulties in measuring resilience are influenced not only by scale and geographic regions but also disaster subcultures. Researchers must recognize these differences before making generalizations.

Return to Index

Kevin Ash, University of South Carolina
Ronald Schumann,University of South Carolina
Gregg Bowser, University of South Carolina

Am I Really in Danger? How Varying the Visual Representation of a Tornado Warning Can Alter Threat Perception

Tornado warnings are becoming more precise as remote sensing and communications technologies continue to improve. In 2007, the National Weather Service transitioned away from county-based warnings and now issues storm-based warnings. In the near future, the warning system will allow incorporation of probabilistic information regarding the presence of a tornado within a warned area.

In addition, smartphone and tablet applications increasingly support instantaneous dissemination of various types of graphical weather warning information. Private weather companies have also developed their own probabilistic graphics, already used by some broadcast meteorologists. However, the implementation of these new graphics has occurred without substantial research guidance on how the American public might interpret or act upon various visual representations of tornado risk.

This study examines two basic components of a tornado warning visualization: the color scheme and the position of a threatened person on the graphic. A questionnaire gauging the level of perceived fear and likelihood of protective action for a series of hypothetical warning scenarios was distributed to about 500 undergraduate students in Columbia, South Carolina. Three warning designs are included in the study: one similar to the current style of storm-based warnings, and two different versions of probabilistic-style storm-based warnings.

Using statistical and geostatistical methods, the three different warning designs are analyzed to determine whether high and low threat perception responses differ according to the graphic designs. Several attributes of the warning designs are found to influence threat perception, including the color scheme, position inside or outside the warned area, proximity to the highest threat area, and proximity to the edge of the warned area. Results suggest that respondents perceive fundamentally different areas of danger when viewing the probabilistic-style warnings than when viewing the design emulating current tornado warnings.

Return to Index

Natalie Baker, University of California, Irvine

Redefining Disaster Preparedness as a Situated Phenomenon

I examine and critique the broad concept of disaster preparedness, as it is currently understood, in this paper. I do so through empirical research conducted at a large university in Southern California, an area at risk for a catastrophic earthquake.

Using in-depth interviews, archival analysis of disaster planning materials, and field observation under a grounded theory framework, I answer three research questions. What does disaster preparedness mean for participants? How do participants prepare for a large earthquake? How do they imagine responding to both short- and long-term disaster scenarios?

In addressing question one, I observe how interpretations shape action or inaction in the face of a potential catastrophe and develop the notion of traditional preparedness as composed of two aspects: the collection and storage of material objects, resources, knowledge, and plans; and imagined successful response to future scenarios. The two aspects are highly moderated by major dimensions of trust—faith in infrastructure continuity and trust in reproducible realities.

For the second question I distinguish situated preparedness from traditional preparedness, and examine how it is enacted through two action categories. Explicit practices are the traditional planning actions people engaged in accordance with official recommendations. In contrast, implicit practices are taken for granted activities and resources people use in everyday life with potential to help people respond in disaster. I categorize these practices and develop a new conceptualization of preparedness as carried out through action blending between the sets of practices, modified through dynamic aspects of context.

The findings related to the final question identify important sources of disconnection between participants rooted in an expectation of personal responsibility and compliance when response is largely collective and situational. I use all three sets of findings to reconceptualize preparedness as a situated, collaborative effort.

Return to Index

Mara Benadusi, University of Catania,Italy

Pure Victims or Active Survivors? Degrees of Resilience and Vulnerability

By charting the branching paths taken by the vulnerability-resilience pairing in a Sri Lankan post-tsunami village, this paper deconstructs some of the assumptions connected to these concepts in current disaster risk reduction policies and practices. In this field, catastrophe survivors are commonly called on to engage with a dichotomous classification that represents them as either passive and vulnerable or resilient and active. This trend has its consequences. In order to be considered worthy of aid, those affected by natural disasters must find the most opportune balance between the two opposing identities: pure victim or active survivor. And yet both resilience and vulnerability require a great deal of agency, from those called on to personify them as well as those tasked with identifying, soliciting and in some way caring for them. Both lend themselves to varying definitions and stagings according to the circumstances of social life.

The paper will ethnographically explore how the local actors involved in post-tsunami reconstruction in Sri Lanka viewed the academic literature and the technical vocabulary of humanitarian action term resilience and vulnerability, and what kinds of "things" these might be for them. Attempting to go beyond the sphere of cultural intimacy, I will describe the diverse positionings of people encountered in the field: religious leaders, humanitarian aid workers, local and national politicians, government officials, and rural farmers.

A problematic portrait emerges. Still vulnerable yet already resilient, the survivors were encouraged to strategically employ one or the other facade of disaster management's "two-faced Janus," calibrating the features that rendered them desirable for gifting: a moving need for help and an untiring ability to cope with uncertainty.

Return to Index

Eric Best, University of Delaware

Practicing What We Publish: The Gap Between Building Evacuation Research and Building Evacuation Modeling

This paper presents an overview of the state-of-the-art in building evacuation theory in the field of emergency management and current best practices in the building evacuation modeling community. While there are significant new developments in the field, much of the academic work focusing on how people evacuate is ignored in models produced today. There is much to be done to modify these models to account for the new developments.

This paper presents SocEvac, a model developed at the Disaster Research Center that focuses on how social behavior interactions and extreme density situations and contraflows influence evacuations.

SocEvac allows for model users to test building evacuation theories from sociology and emergency management in simulation environments. The model focuses on the inefficiencies related to evacuations, including incomplete information, desire for groups to evacuate together, and asymmetric skills and information that vary between evacuation participants.

The intention of this model and paper is to offer tools to integrate emergency management theories and best practices in evacuation models, using SocEvac as an example of what is already possible.

Return to Index

Christine Bevc, University of North Carolina at Chapel Hill
Matt Simon, University of North Carolina at Chapel Hill
Tanya Montoya, University of North Carolina at Chapel Hill
Jennifer Horney, University of North Carolina at Chapel Hill

Institutional Impediments: An Evaluation of Local Planning Practices for Vulnerable and At-Risk Populations

In the wake of recent policies, the North Carolina Preparedness and Emergency Response Research Center introduced the Vulnerable and At-Risk Populations Resource Guide to aid local health departments planning for vulnerable and at-risk populations. Since January 2012, 73 percent of LHDs in North Carolina accessed the Guide (as of April 1, 2013), along with more than 60 departments representing 19 states outside North Carolina.

This paper focuses on identifying institutional facilitators and barriers to preparedness planning for vulnerable and at-risk populations at the local level. Researchers conducted a multi-level and mixed-methods evaluation study over an 18-month period, including surveys, focus groups, interviews, and secondary data, to examine planning practices across multiple local, regional, and state jurisdictions. The research team compared the respondent reported populations of interest to the corresponding census data to determine whether there were possible differences in planning priorities.

Supplemental survey data was used to identify the extent of evidence-based decision making. Data collected from evaluation surveys were used to triangulate key institutional facilitators and barriers associated with planning for at-risk populations, including challenges to conducting assessments and lack of resources. Results also identified barriers within institutional culture and disconnects between planning priorities and population densities.

The findings support a more in-depth understanding of the public institutions and lack of disaster preparedness that contribute to the unsafe conditions and dynamic pressures that contribute to the progression of vulnerability, under Blaikie's Pressure and Release Model.

These results discuss the role of local agencies in preparedness and the implications for the organizational and bureaucratic impediments to planning implementation. This research also contributes to practice-based efforts to improve response services and address public health preparedness planning as they relate to vulnerable and at-risk populations.

Return to Index

Gregg Bowser, University of South Carolina

Determining the Differences in Evacuation Influences and Perceptions within the Elderly Population

The United States is becoming a "grayer" nation. U.S. Administration of Aging projections indicate that by 2030 nearly 20 percent of the national population will be aged 65 or older, with a significant portion of this growth occurring along the hurricane-prone Atlantic and Gulf coasts. Such an unprecedented demographic shift creates new challenges for emergency management. Previous research shows that older adults do not perceive risks and warnings the same way as other groups, and as a result may not react to risks in the same way as the rest of the population. Disproportionately high fatality rates for the elderly in recent disasters indicate that these differences are a key determinant of survival in a disaster. Crucial information about how elderly perceive and respond to hazards is missing.

This research improves understanding of how the elderly population perceives and responds to the threat of hurricanes in one of the most rapidly graying states in the country: South Carolina. One goal of this research is to identify and explore key differences in risk perception and hurricane behavior between the younger population and the elderly. Another goal is to compare changes in behavior and intended action between two different segments of the elderly population: the "young-old" (those aged 65 to 74) and the "aged and oldest-old" (those 75 and older.)

This study explores these differences through a mixed-methods approach, combining semi-structured interviews of elderly individuals and causal models developed using data from the South Carolina hurricane evacuation survey. The findings of this study indicate that various influences—such as pet ownership and work status—affect potential evacuation behavior differently between groups. The findings of this study provide the basis for planning improvements and outreach for coastal elderly populations.

Return to Index

Brittany Brand, Boise State University
Pamela Mcmullin-Messier, Central Washington University
Melissa Schlegel, College of Western Idaho

Map Your Hazards! An Interdisciplinary, Place-Based Undergraduate Course Module Designed to Promote Informed and Resilient Communities

We are developing a "Map your Hazards!" undergraduate classroom module to provide students with an interactive mechanism to engage in place-based exploration of natural hazards, social vulnerabilities, and factors that shape perception of natural hazards and risk.

The module allows students to integrate geoscience and social science methodologies to understand societal impacts that result from natural hazards. Within the module, students: (1) identify and apply credible geologic and social science datasets to identify hazards and social vulnerabilities within their region; (2) collect and evaluate survey data on the knowledge, risk perception and preparedness within their social networks; and (3) make recommendations to potential community stakeholders to develop a prepared, resilient community.

Students gain insight into how our knowledge and perspectives of the world shape how we interact with it and how we promote and build resilient communities through understanding the relationship between human systems and natural systems. We will test the module this fall in our respective colleges, but preliminary results from a pilot study at University of Washington: (1) illustrate the increase in the students' interest and level of engagement; and (2) demonstrate the types of useful data collected within student communities.

Students found that a lack of accurate knowledge of regional natural hazards corresponds to a low perception of risk associated with these hazards, even among students who had grown up or had lived within the region. A broader objective of this module is to promote an informed community and bridge the gap between the scientific and hazard mitigation communities through the education of undergraduate students.

Return to Index

Divya Chandrasekhar, Texas Southern University
Yang Zhang, Virginia Tech

An Institutional Study of Disaster Recovery Management in Three Countries (China, India and the U.S.)

The increasing internationalization of the disaster aid industry has complicated the institutional setting within which local recovery must now be managed. Therefore, effective post-disaster recovery planning requires a thorough understanding of the institutional factors that influence its process and outcomes. Without it, negative externalities of post-disaster recovery management including involuntary displacement, delayed household recovery, and socio-political disempowerment will continue to occur.

The objective of this paper is to use an institutional perspective to examine how governance structures, predisaster local development practices, and post-disaster policy making affect processes and outcomes of local recovery planning. We compare post-disaster experiences in three countries: China after the 2008 Wenchuan earthquake; India after the 2004 Indian Ocean Tsunami; and the United States after the 2005 Hurricane Katrina. This comparison across three different types of political economies allows us to capture those institutional effects that are less visible to single-setting studies.

Data was collected as part of separate case study researches conducted by the two authors in the China, India, and the United States between 2008 and 2013. The case studies in India and the United States were qualitative. Data was collected through key informant interviewing, direct observation, and review of secondary documents. The case study in China was a mixed method study involving interviewing, review of secondary documents, and household survey.

Our findings compare the experiences of the three countries along the following cross-cutting themes: (1) predisaster development practices; (2) involvement of international actors in recovery; (3) design and implementation of ad-hoc post-disaster recovery policies; and (4) conditions of mechanisms of recovery aid funding.

Return to Index

Wenfang Chen, Beijing Normal University
Susan Cutter, University of South Carolina
Christopher Emrich, University of South Carolina
Peijun Shi, Beijing Normal University

Measuring Social Vulnerability to Natural Hazards in the Yangtze River Delta Region, China

Social vulnerability has gained much attention in disaster science as it emphasizes that human factors are more responsible for disaster losses. As a multi-dimensional concept, it requires adequate knowledge of the social, economic, and political background of targeted places. Research of this kind is scant in China, although China continuously suffers from devastating natural disasters. This paper presents a preliminary study on the quantifying and analyzing the social vulnerability in the Yangtze River Delta region in China, following the framework of the place-based Social Vulnerability Index (SoVI) developed initially for the United States, but incorporating social and political contexts of China and this specific region.

Thirty variables are selected, adapting the SoVI to China's social background for the 75 study units in the study area. Principal component analysis is used to reduce the data dimension. Finally, seven principle components—identified as "urban," "renters," "gender and age," "housing conditions," "medical service," "children," "special needs population," and "population change," respectively, are obtained. Factor scores are summed to get the final SoVI scores for each study unit.

The spatial patterns of both the components and SoVI reflect that some specific historical, social, and economic factors in the study area have a great influence on the social vulnerability, notably including the household registration system, the One Child Policy, and the fast urbanization and aging process. Suggestions are given to city level and county level policy makers for vulnerability reduction and risk mitigation.

Return to Index

John Cross, University of Wisconsin Oshkosh

Communication of Emergency Warnings on a University Campus

This paper reports the results of a survey of nearly 700 undergraduate students at the University of Wisconsin Oshkosh regarding the means to effectively communicate emergency messages. Students were questioned about their utilization of various messaging systems, including whether their cell phones, e-mail, or smart phones were on in the classroom setting. Students were asked about the ways they would prefer to be notified about both urgent emergencies and less critical situations, such as potential adverse weather or crime notifications. Data was collected to discern whether the preferred mode of communication differed between the classroom setting and the residence hall, and how students living on campus might differ from those living off campus or commuting from nearby communities.

Electronic media can quickly transmit messages, but they are often slowly received. During class, 95 percent of the students indicated their cell phones are off, set to vibrate, or operate in silent mode. If an e-mail was sent with the words "Emergency Message" in the subject line, only 41 percent expected to have noticed and read the message within 15 minutes of its transmission. Nevertheless, 42 percent of the students indicated their preferred way of receiving an emergency message in class was via "text message on cell-phone or similar device."

Students were questioned regarding the individual who would have the greatest credibility in communicating emergency messages. Although the university's news bureau has typically sent notices regarding weather-related class cancellations, 58 percent of students indicated that they "would [be]… most likely to believe and follow" messages from the university police chief. To obtain more information regarding the emergency message, two-thirds of the student indicated they would "contact friends or family on your mobile device" or "go to the internet with your mobile device."

Return to Index

Jacqueline Curtis, Kent State University
Andrea Szell, Kent State University

Viewsheds as a Geographic Scale for Mapping and Analyzing Neighborhood Psychosocial Stressors in Post-Disaster Environments

From the healing effects of gardens to depressive symptoms from neighborhood disorder, research across the social sciences and public health posits that what we see matters for our health. Indeed, what we see is actually a form of environmental exposure. Despite this general agreement, the details of the variables that constitute exposure and the pathways that link exposure to health outcomes are not fully understood.

The existing research on this subject has been mostly carried out in inner cities facing a suite of characteristics indicative of urban decline. These variables have been framed as neighborhood psychosocial stressors. However, empirical evidence from a number of post-disaster environments—including post-Hurricane Katrina New Orleans and post-Waldo Canyon Fire Colorado Springs—indicates that these characteristics can also emerge in places during recovery.
Furthermore, research addressing how visual exposure to neighborhood psychosocial stressors impacts health usually relies on census aggregation units (e.g., block, tract) as the scales of analysis, though these scales do not adequately capture a resident's visual exposure.

A more appropriate geography of analysis may be the viewshed, which represents the visible area surrounding a certain location, in this case, a home. This paper presents a review of viewshed analysis to identify its strengths and limitations as a scale at which to study exposure to psychosocial stressors in post-disaster environments. We then provides case studies from New Orleans and Colorado Springs to demonstrate how this approach can be integrated with other forms of spatial data to map and analyze exposure to neighborhood psychosocial stressors in post-disaster environments.

Findings indicate that exposure varies spatially within neighborhoods undergoing recovery, regardless of the cause of damage (hurricane vs. wildfire in this case) or the stage of recovery. Implications for spatially targeted public health intervention are discussed.

Return to Index

Emel Ganapati, Florida International University
Meredith Newman, Florida International University
Christa Remington, Florida International University

Emotionally Drained: Addressing Organizational Factors That Contribute to Burnout Among Response and Recovery Workers in Post-Disaster Haiti

Although demand for emotional labor is at its highest in times of disasters, there is a lack of studies in public administration on emotional labor involved with disaster response and recovery workers in international contexts. Focusing on emotional labor of those involved in response and recovery following the January 12, 2010 Haitian earthquake, this article addresses the following questions: What are the organizational factors that contribute to burnout among response and recovery workers who perform emotional labor in international post-disaster contexts? What can response and recovery organizations do to eliminate or reduce the impact of such factors?

Based on in-depth interviews with response and recovery workers, we identify five organizational factors that contribute to burnout among the response and recovery workers: (1) not being able to distance yourself from work; (2) not having places at work to decompress alone; (3) not seeing the direct impact of agency's programs on the ground; (4) conflict between the organizational culture and the post-disaster context; and, (5) lack of pre-disaster preparedness, in-field, and post-disaster training and support.

Overall, the study highlights the importance of providing training and support to disaster response and recovery workers before they enter the field, in the field, and after they exit the field, as well as creating a better work and living environment and designing special programs for them in post-disaster contexts.

Return to Index

June Gin, Veterans Emergency Management Evaluation Center
Diana Scarrott, Veterans Emergency Management Evaluation Center

Disaster Preparedness in Residential Organizations Serving Homeless Veterans in Los Angeles

People experiencing chronic homelessness can be particularly vulnerable to the adverse impacts of disasters due to lack of housing, dependence on trusted caregivers, extreme poverty, and health conditions. Veterans are twice as likely to be homeless as non-veterans, and homeless veterans suffer disproportionately from mental health conditions such as PTSD and substance use disorders. Homeless individuals are not always included in community disaster planning activities, and may experience resulting setbacks in their efforts toward self-sufficiency.

Although homeless shelters constitute a vital disaster safety net, little is known about the preparedness capabilities of nonprofit organizations and their ability to serve their clients during disasters. This pilot study examines disaster preparedness activities in nonprofit organizations providing shelter and housing to homeless individuals.

We conducted 12 semi-structured interviews at six nonprofit homeless providers in Los Angeles County. One executive and one emergency planner were interviewed at each site. The interviews examined three organizational preparedness domains: (1) their ability to protect staff and client life safety; (2) their ability to restore key services and continue housing residents; and (3) their connectedness to partners and viability as a community disaster resource.

The homeless services providers in this sample were aware of the disruptive potential of disasters. Interviews suggest that providers had adopted comprehensive safety measures, but had not planned for continuity of care services. Disaster plans also appeared internal and insular. Organizations lacked strategies for outreach to street homeless, plans for relocating clients to alternate facilities, and pre-disaster collaborations with partners, peer organizations, and emergency management officials. Training, technical assistance, and opportunities to partner with government agencies were identified as needs. Government agencies can support the disaster capabilities of homeless shelters through increased emergency management collaboration and training.

Return to Index

Robert Gottlieb, University of South Carolina
Daniel Morath, University of South Carolina
Kevin Ash, University of South Carolina
Ronald Schumann, University of South Carolina
Christopher Emrich, University of South Carolina

Spatial Analysis of Sampling Error in the American Community

Spatial patterns of sampling error for estimates of socioeconomic variables by the American Community Survey are examined. Following the removal of long-form survey from the decadal U.S. Census, national five-year data for the ACS are now used in many applications related to disaster vulnerability and resilience by a variety of users. However, less attention has been paid to the error that the U.S. Census Bureau reports for this data. The reported margin of error for each analysis unit is converted to a coefficient of variation. Most users accept that a CV less than 0.1 or 0.15 indicates that the estimate is reliable. However, a large number of counties and census tracts do not meet this criterion.

Multiple statistical techniques are used to assess the properties of spatial clusters of the CV for several different variables for the most recent ACS versions. While both the county and census tract level of analysis are used, considering sampling error is particularly important at the tract level. Even though there is an overlap of four years for the 2006-10 and 2007-11, CV patterns change considerably, so users need to check the sampling error for each version used. There is a threshold based on the total population beyond at which ACS estimates are reliable, but below this threshold, one cannot predict the reliability using population alone.

Return to Index

Dana Greene, University of North Carolina at Chapel Hill
Tamara Gull, Oklahoma State University

Livestock Evacuations During the 2012 Oklahoma Wildfires

Oklahoma is subject to weather extremes ranging from snow and ice storms to tornadoes, dust storms, severe drought, and fire, creating challenges for humans and animals alike. The drought that has plagued the central United States for the past two years has had a significant impact on Oklahoma. Extreme heat, rainfall deficits, and continually shifting winds have provided a perfect atmosphere for wildfires.

Using a triangulated research design of in-depth qualitative interviews, observations, documents, spatial mapping, and visual data, we gathered information from three counties affected by wildfire in 2012. We focused on variables that influenced the ability to evacuate livestock versus shelter-in-place, such as the availability of transportation resources, an evacuation location, assistance with animal handling, the size of the herd, dispersal (pastured vs. penned/stabled), and the rapidity of wildfire onset.

We found that residents worked hard to evacuate their animals when the wildfire threat became apparent, but ran into difficulties with regard to transporting large numbers of livestock to safety, particularly with regard to dispersal and trailer availability. Our research findings emphasized the necessity for emergency management and mitigation plans to include safeguarding livestock.

We made specific recommendations for animal owners and veterinarians. Since social networks were found to be crucial in successful animal movement, we recommend that these networks be mobilized as a means of developing and testing evacuation plans for people and livestock. In addition, we recommend that animal owners create and practice an animal evacuation plan, and permanently identify their animals. Finally, we recommend that livestock owners have a priority list for evacuation.

Return to Index

Alex Greer, University of Delaware
Eric Best, University of Delaware

Size Doesn't Matter: The Complicated Relationship Between National Offshore Oil Spill Events, Framing, and Policy

This paper explores the relationship between the framing of and attention paid to U.S. offshore oil spill events and the resulting federal policy. It is essential to understand issue framing because it shapes how media, the general public, interest groups, and politicians define what issues led to failure and propose acceptable solutions that, through agenda setting, develop into policy.

To better understand this process within the offshore oil industry we examine nine offshore spills of over four million gallons each occurring less than 100 miles off the coast of the United States. We created a case study for each spill detailing media coverage and policy development resulting from each event.

Using nationwide databases, we categorized these spills based on the attention they received in the national media. These case studies underscored two important, interrelated findings. Interestingly, there is no direct correlation between the size of the spill and the amount of attention it receives in our dataset. The amount or relative strength of legislation was also not correlated to the size of the spill.

We recommend that further studies explore the suggested relationships between geographical distribution of spills and legislation.

Return to Index

Sara Hamideh, Texas A&M University

Collaborative Planning for Post-Disaster Recovery: Galveston After Ike

How could post-Ike recovery planning in Galveston been done better? This paper presents a case study of long-term recovery in Galveston, Texas, after Hurricane Ike and assesses its successes, failures, and opportunities for application of collaborative planning. There is a general consensus on the value of planning for disaster recovery and its timing as a critical factor. However, the existing literature fails to address a number of issues that might affect the success of the recovery process.

Collaboration techniques have been suggested for recovery planning processes as tools for capacity building. And yet the full spectrum of a collaborative planning approach has not been applied to recovery. Collaborative approach provides a framework to address a broad range of interests in recovery. It is also important to reflect on the connections between comprehensive planning and recovery planning. Rarely are these two processes integrated. The present paper seeks to address these issues and thereby provide a broader theoretical framework for recovery policy making.

Through a qualitative case study using participatory observation analysis, in-depth interviews and document analysis, this study explores Galveston's recovery. Recovery decision making in Galveston could be best described as a blend of technical-bureaucratic, political influence, and social movement which are different from a collaborative model with respect to diversity and interdependencies of involved interests and authenticity of dialogues among stakeholders. This led to discontinued community engagement; over-politicized recovery; and non-inclusive pursuit of opportunity for removing minorities.

Lessons for future recovery planning support: (1) establishing recovery collaboratives before disasters; (2) including all community groups in recovery collaboratives; (3) developing comprehensive consensus-based visions through recovery planning process integrated with the comprehensive plan.

Return to Index

Ziqiang Han, University of Delaware

The Effect of Social Capital on Perceived Disaster Recovery: A Longitudinal Study Based on the 2008 Wenchuan Earthquake in China

The importance of social capital in disaster management has attracted attention from disaster researchers in recent years. In this paper, the effects of social capital on perceived disaster recovery were analyzed using a dataset drawn from a two-wave longitudinal household study after the 2008 Wenchuan earthquake in China.

The first data collection was conducted at the very beginning of the recovery period, and the second wave of data was collected four years later. Logit and ordinal logit models were adopted for the analysis. The results revealed that the social capital had a significant and positive impact on the perceived recovery degree, but the effects of the three sub-indicators of social capital (organization membership, social support, community cohesion and solidarity) varied.

While the social support indicator had a consistent, positive, and significant effect on recovery perception, the community cohesion and solidarity culture indicator's positive effect was not consistent in all models, and the organization membership indicator was not significant at all. Meanwhile, financial capital and natural capital also had positive, significant effects on the perceived recovery, and these effects were actually more consistent than social capital in this analysis.

Finally, the reasons of such differences from the three social capital sub-indicators were discussed. The possible development of a comprehensive set of social capital measurements for disaster and emergency studies was proposed.

Return to Index

Shih-Kai Huang, Texas A&M University
Michael Lindell, Texas A&M University
Carla Prater, Texas A&M University

What Affects Households' Hurricane Evacuation Decisions? Review of 20 Years of Hurricane Evacuation Studies

Although empirical studies of warning response have made major contributions to the disaster research literature, studies of households' responses to hurricane warnings have not received a systematic literature review for nearly two decades.

To identify predictors of hurricane evacuation, this study collected 30 scientific articles (describing 35 actual and 11 hypothetical hurricanes) published since Baker's 1991 review. We categorized predictors of evacuation in terms of eight categories derived from Lindell and Perry's Protective Action Decision Model. These are demographic characteristics, information sources, geographic location and personal experience, official warnings, social and environmental cues, perceived storm characteristics, expected personal impacts, and perceived evacuation impediments.

The results revealed that, although households also rely on information from authorities and peers, the news media are consistently the most important source of information. Consistent with Baker's conclusion, demographic characteristics other than female gender and home ownership are not consistently useful in predicting household evacuation.

Also consistent with Baker, mobile home owners and risk area residents are consistently more likely to evacuate and are especially influenced by official evacuation warnings. Most social and environmental cues (e.g., observations of peers evacuating) and risk perceptions (e.g., beliefs about hurricane characteristics and personal impacts)—are consistently significantly correlated to household evacuation. However, many predictor variables have been examined in only few studies so the empirical support for these variables remains low. Thus, more studies are needed to build a solid empirical base for evaluating the effects of these predictors on household evacuation.

Return to Index

Beatriz Hummell, Pontifical Catholic University of Parana
Susan Cutter, University of South Carolina
Christopher Emrich, University of South Carolina

Constructing the Social Vulnerability Index (SoVI) to Natural Hazards in Brazil

Social vulnerability highlights differences in the human capacity to prepare for, respond to, and recover from disasters. It varies over space and time, and among and between social groups, due to differences in socioeconomic and demographic characteristics.

Brazil lacks indicators and frameworks to access vulnerability for the entire country. This research produced a Social Vulnerability Index replication study for Brazil, showing how SoVI concepts and indicators were adapted to the country, taking into consideration Brazil's social, economic and demographic particularities. Also, it accounts for issues regarding methodology and concept adaptation due to data availability.

SoVI Brazil follows the place-based framework adopted in the social vulnerability index initially developed for the United States. Forty-eight city-level indicators were reduced to 10 factors using a principal component analysis, explaining about 65 percent of the variance. SoVI is then calculated by the sum of the factors.

A few spatial patterns can be identified, showing a concentration of the most vulnerable cities in the north and northeast regions, as well as metropolitan regions and state capitals in the south and southeast regions, especially São Paulo and Rio de Janeiro. The least vulnerable cities are mainly concentrated in the inland regions of the Southeast.

Although factors contribute to the social vulnerability differently in each city, the overall results confirm that the social and economic disparities among Brazil's different regions are reflected in the population's vulnerability to natural hazards.

Return to Index

Beatriz Hummell, Pontifical Catholic University Of Parana
Clovis Ultramari, Pontifical Catholic University Of Parana

Geographies of Solidarity: Complexities in Priorities Definition and Aid Addressing after Natural Disasters—Preliminary Findings and Results

Although the mechanisms that coordinate relief and aid provision for regions struck by natural disasters are effective and organized, it is possible to notice that aid donors assist affected areas differently. The main objective is to verify the working hypothesis that resources addressed to natural disasters are related to the evidence and features of each case and not necessarily to the level of severity attained by the disaster.

It is important to consider that a donor's decision to provide aid to a country or region result not only from solidarity, but also from a range of opportunities, restrictions, factors (social, political, geographic, cultural, etc.), and interests that intermediate the relation among donors and the affected areas.

Our methodology's quantitative approach consists of the analysis of 2000 to 2010 data collected from the Financial Tracking Service (FTS) regarding donations, donors, appealing agencies and affected countries, as well as data on victims, people affected and damage loss collected from the International Disasters Database (EM-DAT).

The qualitative approach consists of a survey applied to donor agencies, mainly nongovernmental organizations.

Preliminary findings indicate that, in most cases, the amount of aid is not proportional to the severity of the disasters, especially when considering the affected population and damage loss. Factors such as common language and previous experience in the affected area influenced donors' decisions about which regions or countries they will provide aid after natural disasters.

Return to Index

Rahmawati Husein, Universitas Muhammadiyah Yogyakarta

Assessing Hazard Mitigation Policies and Strategies in Java Coastal Areas

There has been a growing literature and studies on the role land use planning and development regulations can play in hazard mitigation. However, almost none of the research has been done in Indonesia, particularly in mitigating the coastal hazards although Indonesia has one of the longest coastal areas in the world.

Using tools developed by the Hazard Reduction and Recovery Center, Texas A&M University, this study investigates policy and practices of local jurisdictions, municipalities, and regencies along Java coastal areas where they are vulnerable to hazards. This study examines the adoption and implementation of land use planning regulations including building standards and information dissemination strategies at the local levels (Kabupaten and Kota).

The paper provides an overall assessment of the types of land use and development regulations and policies implemented by local jurisdictions based on surveys of leading planner (head of Bappeda) and leading public work officer (head of Dinas PU). In total, 31 municipalities and regencies participated in the survey from the total 65 regencies and municipalities in Java coastal areas. In addition, two provinces out of six provinces in Java also responded the survey.

Statistical models were developed to assess the adoption and extensiveness of using various forms of land use regulations, hazard mitigation programs, and hazard related building standards. Results suggest wide variation in the adoption and extensiveness of usage of strategies for reducing any impact of hazards in the coastal areas.

Return to Index

Kanako Iuchi, Tohoku University
Elizabeth Maly, Disaster Reduction and Human Renovation Institution

Initial Regional Assessment of Local Recovery Planning in the Tohoku Region after the Great East Japan Earthquake

What are the goals and planned procedures related to housing, resettlement and livelihood recovery of the tsunami-affected coastal municipalities? In particular, how is the current progress consistent with the initial master plans in terms of timeline and contents?

Over the two years since the Great East Japan Earthquake and tsunami, multiple levels of government have invested significant efforts in coastal municipalities towards a smooth recovery. Yet, the overarching recovery picture is difficult to understand, because of the broad scale of devastation and limited aggregated resources about planning. This research examines the status of recovery planning in municipalities two years after the quake, with emphasis on housing, resettlement, and livelihoods.

Three methodological steps are included. First, local governments' publicly available recovery plans were analyzed with respect to housing, land use, and re-settlement policies in the context of the planned timelines. Second, focused issues in recovery plans were qualitatively grouped, with reference to each municipality's local context. Finally, several selected communities were investigated to confirm their planning progress.

Our preliminary conclusion is that "one-size-fits-all" planning is complicating local planning processes. Three key findings support this idea: first, housing, resettlement, and urban development policies are similar in municipal recovery plans, although characteristics of municipalities vary. Second, different types of plans exist in support of, or in opposition to, official municipal recovery plans.

These plans represent outcomes of a consensus-based process. However, there are many localities without alternative plans. Lastly, while original recovery plans were quickly crafted, their implementation has not been as rapid.

Return to Index

Eliot Jennings, University of North Texas

Factors Influencing the Intent to Use and Use of Disaster Response Decision Support Software Technologies: An Empirical Examination of Local Emergency Managers in FEMA Region 6

A defining characteristic of today's world is technology. In recent years, government organizations have extended the application of technology to governmental functions in their desire to be more efficient, quicker and more reliable. As we continue to expand our use of technology in governmental affairs, it is important to understand how the increasing role of technology affects disaster management.

While research is plentiful on technology acceptance in the private sector, much less research has been conducted on the acceptance and use of technology in the public sector, and in particular, within the domain of disaster management. Utilizing Venkatesh et al's unified theory of acceptance and use of technology and a recent survey of city and county emergency management officials in Federal Emergency Management Agency Region 6, this study examines factors that influence the use of decision support technologies such as WebEOC and Eteam in disaster response.

The preliminary findings indicate that contrary to private sector technology acceptance studies in which performance expectancy is the strongest indicator, social influence is the strongest indicator of the intent to use decision support technology in disaster response by local emergency management officials. The research further indicates additional factors influence the transition from intent to use to actual usage of decision support technologies in disaster management.

Return to Index

Victoria Johnson, Massey University

An Impact Evaluation Of Shakeout, an Earthquake and Tsunami Drill in Two Coastal Washington State School Districts

In October 2012, Washington participated for the first time in ShakeOut, an annual, one-day event that informs the public about earthquake and tsunami preparedness and encourage residents to simultaneously practice "Drop, Cover and Hold On," the recommended protective action during an earthquake. The aim of this evaluation was to see how well children in grades six through 12 in two coastal Washington state school districts understood the objectives and consequences of the protective actions practiced during the ShakeOut drill, including a practice of vertical evacuation inside the schools.

Across both school districts, 29 teachers volunteered to participate and administered pretest and posttest questionnaires to 574 students, age 10 and older, in order to assess differences in students' knowledge, skills and attitudes about disaster preparedness and protective actions as a direct result of the ShakeOut drill.

The evaluation found that students had high levels of familiarity and correct knowledge about protective actions for earthquakes and tsunamis both before and after ShakeOut. However, significant portions of students had varying levels of knowledge and comprehension of the risks that cause injury and in some case had difficulty applying what they have practiced in the classroom to situations outside the classroom.

More than a quarter of students in both school districts did not know or were not sure if they participated a tsunami evacuation drill during ShakeOut. These results indicate that school-based disaster drills should be complemented with additional classroom lessons and discussion.

Return to Index

Nabil Kamel, Arizona State University

Community Resilience and Post-Disaster Recovery: The Role of Community-Based Organizations in Disadvantaged Neighborhoods

This paper presents broad lessons of the long-term recovery of several major disasters—Loma Prieta earthquake, Northridge earthquake, Hurricane Katrina, and the Sichuan earthquake. In addition to overall trends, we present cases of specific actions taken at the local level by community-based organizations.

In each of these cases, different courses for recovery were charted by local and community-based organizations, such as the Black Panthers in Loma Prieta, New Economics For Women and Comision Femenil in Northridge, the Vietnamese community in Katrina, and minority groups in the autonomous prefectures of the Sichuan province. The cases show that, through community planning, integration, and solidarity, marginalized communities can overcome institutionalized barriers in original ways, develop innovative solutions, and implement transformative progressive changes.

We propose that, in normal times, local and state authorities should develop strong and functional ties with housing producers, housing development corporations, and community-based organizations that address social needs. Similarly, learning from community-based organizations, local authorities should invest in affordable and rental housing, land banking, and community land trusts. They need to maintain consistent political support for proactive land use planning and affordable housing provisions, and prepare related plans that can be implemented in the aftermath of a disaster.

Return to Index

Anne-Lise Knox Velez, North Carolina State University

To What Extent Are NEPA Resource Categories Being Included in Local Disaster Planning Efforts?

The National Environmental Policy Act emphasizes the importance of preserving physical, biological, social and cultural resources in federally funded projects. Physical resources include geology, air, soils and water. Biological resources include plants and animals. Social resources include communities and economics, and cultural resources include archaeological and historic artifacts.

NEPA is procedural, meaning federal programs including Federal Emergency Management Agency should be NEPA compliant. Because FEMA requires states to adopt disaster plans to receive federal disaster aid—and county plans are often based on state plans and FEMA templates—one would expect NEPA resource categories to be reflected in county-level disaster planning. As disaster clean-up efforts are often federally funded, it is important to understand whether NEPA resources are included in local disaster planning efforts and to provide local emergency managers with information needed to plan around resources that are not currently emphasized.

This study uses combined data from both surveys of and telephone interviews with 86 county-level emergency managers in four states in the northwestern United States to assess: (1) Are NEPA resource categories being included in county disaster planning? (2) How much emphasis are counties putting on including each NEPA resource category in their planning efforts? (3) Which NEPA categories are most emphasized in county disaster planning efforts? and (4) Which factors contribute to the inclusion of NEPA categories in disaster planning efforts?

Analysis is conducted using IBM SPSS statistics software. ANOVA is used to determine emphasis on NEPA categories, and multinomial logistic regression to determine the strength of the degree of resource category inclusion. Binary logistic regression is used to determine which factors influence the inclusion of each NEPA category in local planning. Preliminary analysis indicates that social resources are most likely to be included in county level disaster plans, followed by biological and physical resources, then cultural resources.

Return to Index

Juliana Lam, University of South Carolina

Displacement in the United States Following Billion-Dollar Disasters

Disaster displacement and migration have become a greater research focus in the United States, particularly following major hurricanes. The number of disasters that have incurred over one billion dollars in losses continue to rise, especially in the last two decades. This research examines ten disasters selected from the Billion-Dollar Weather/Climate Disasters report published by the National Oceanic and Atmospheric Administration's National Climatic Data Center, and uses the United States Postal Service vacancy postal data as a sound proxy variable to reflect population displacement following the occurrence of each disaster chosen for this study.

Three questions inform this study. First, is there an observable population shift following a billion-dollar disaster? Second, how is social vulnerability related to displacement following a billion-dollar disaster? And third, what is the relationship between disaster types (hurricane, flood, tornado, drought) and displacement for billion-dollar disasters in the United States.

A GIS-based approach is used to delineate the impact region for each of the ten disasters and a corresponding non-impact control region to statistically examine the quantity and trend of residential vacancies for specific time steps before and after each disaster. Vacancy rates are used as a proxy for displacement. Preliminary results suggest a difference in vacancy rates before the disaster took place when compared to the rate after the event.

Return to Index

Heather Lazrus, National Center for Atmospheric Research
Rebecca Morss, National Center for Atmospheric Research
Erin Towler, National Center for Atmospheric Research
Debasish Pai Mazumder, National Center for Atmospheric Research
Renee Mcpherson, University of Oklahoma

"Drought is a Relative Term:" Perceptions of Drought Risk Among Diverse Stakeholders in the Arbuckle-Simpson Aquifer

Drought is a challenge faced by communities across the United States, exacerbated by growing demands on water resources and climate variability and change. Recent hydrological studies of the Arbuckle-Simpson Aquifer in south-central Oklahoma indicate the need for sustainable management of the water extracted. The question of how to deal with that management against a backdrop of climate extremes, growing demand, and diverse stakeholder perceptions remains.

To examine this question, we are conducting an anthropological investigation of drought risk perceptions among diverse stakeholders in the vicinity of the aquifer, in conjunction with hydrological analysis of water resources, climatological analysis of regional drought, and meteorological analysis of past precipitation events.

The findings presented will focus around the anthropological component of the study, based on analysis of 38 in-depth interviews conducted with stakeholders in south-central Oklahoma. The interviews were designed to investigate conflict over water management by understanding how people perceive risk differently based on different opinions about the structure of the resource, varying levels of trust in authorities, and unequal access to resources. The Cultural Theory of Risk is used to explain how people view risks as part of their worldviews and why people who hold different worldviews disagree about risks associated with water availability.

Findings will contribute to understanding local perspectives and drivers which is important for sustainable management of the aquifer according to theories of successful governance of common pool resources such as groundwater in Oklahoma. The study is relevant beyond the field site as a novel interdisciplinary approach to understanding weather and climate risk.

Return to Index

Wen-Yen Lin, Ming Chuan University

A Case Study of Identifying Recovery Issues of Community Tourism Industry after a Typhoon Disaster

Typhoon Morakot hit Taiwan on August 7, 2009, and caused casualties totaling 643 dead, 60 missing, and 1,555 injured. This natural disaster has awakened the disaster consciousness of the Taiwanese. Academic disciplines are making further efforts to reducing losses from future disasters by learning from the past. After a devastating disaster struck, many communities of the impact area have to face the recovery problems of how to reinstate their livelihoods, especially the ones have lost their social and environmental resources.

Xinkai is a typical community, whose major basic industry is based on tourism and famous for its hot spring-related recreation activities. The community is struggling to survive after Typhoon Morakot. The record-breaking rainfall in this community caused three debris flows that destroyed houses, recreation facilities and infrastructures, and severed the only roads connecting to Xinkai from outside.

This study is an effort to examine Xinkai's tourism industry recovery issues through literature review, field surveys, and interviews with stakeholders. The issues were logged into six categories: capital, labor, supply, customer, hazard, and transportation. An analysis of relation among recovery issues were made by utilizing DEMATEL technique.

The results are expected to provide possible recovery suggestions of Xinkai's tourism industry, as well as to the communities with similar issues during their recovery processes.

Return to Index

Sophia Liu, U.S. Geological Survey

Formalizing Crowdsourcing in the Emergency Management Domain

Crowdsourcing is not new, but it is gaining significant attention in the emergency management domain through the innovative uses of social software and networking technologies like social media platforms. Ethnographic methods—namely interviewing 80 different stakeholders as well as observing and participating in multi-stakeholder disaster exercises, trainings, and experiments—were conducted to understand how crisis crowdsourcing has evolved since the 2010 Haiti earthquake.

Various findings emerged showing the formalization of crowdsourcing in the emergency management domain and how crowdsourced data can influence official decision making during disasters. Crowdsourcing training and protocols are being implemented and openly shared to standardize these efforts and provide benchmarks for monitoring and evaluating crowdsourcing workflows.

These workflows are the blueprints for facilitating distributed curatorial practices that support particular information management needs during crisis situations. The implementation of these crowdsourcing workflows have also provided a critical foundation for addressing the complex issues around data quality, as well as the ethical issues around privacy and security when collecting and publishing crowdsourced crisis data.

Although official emergency response agencies are interested in engaging with digital volunteer networks, these official agencies require instituting more formal protocols and changes before officially engaging with these porous communities of digital volunteers. Policies are also being incrementally adapted within government agencies to strengthen collaborations with particular digital volunteer communities. The paper presents these socio-cultural, technological, organizational, and political interface challenges that are emerging from the integration of crowdsourcing practices with official, authoritative practices in emergency management.

Return to Index

Bill Lovekamp, Eastern Illinois University
Gary Foster, Eastern Illinois University
Steven Dinaso, Eastern Illinois University
Vince Gutowski, Eastern Illinois University

Cemetery Preservation as Disaster Preparedness

Cades Cove was a thriving Appalachian mountain community prior to the establishment of the Great Smoky Mountains National Park. The creation of the park effectively brought an end to this community. It now exists within the park as a tourist destination (approximately 2 million visitors per year) with a few remaining primitive homesteads and cemeteries. Everyone knows what cemeteries are, but in that familiarity there is dismissal, for few know that cemeteries are far more than what they know them to be. Cemeteries are far more than sequestered repositories for the deceased. Cemeteries constitute libraries of stone and are proxies or microcosmic expressions of the communities they represent, an important element of community.

The first task of this research has been to conduct a sociological examination of Cades Cove and the associated cemeteries, archiving the socio-demographic data of the former communities and their residents. The second objective has been to collect precise GPS coordinates of all known stones and markers in the cemeteries of Cades Cove.

A major disaster can strike at any time, endangering precious cultural property. Our cultural preservation efforts are an important component of disaster preparedness in Cades Cove and to members of the surrounding communities with direct familial ties to the cove. Without these efforts and a major disaster, the community and these libraries of stone would be lost forever.

Return to Index

Brenda Mackie, Bushfire Cooperative Research Centre

Help or Hindrance: Repeated Warnings in the Context of Uncertainty

Disaster scenarios such as pandemics, floods, volcanic eruptions, earthquakes and bushfires, necessitate repeated warning messages in the absence of the actual event. Government and emergency management authorities have a twofold problem because they want to avoid the accusation of panicking the public while running a risk of under-preparing them at the same time. As a result, they may be tempted to downplay the severity of a potential disaster or delay issuing a warning because they are worried the public may get tired of the message.

Known as "cry wolf" or warning fatigue, the cynicism and apathy that can result from being over-warned has been presumptively relegated to "disaster myth," yet continues to be blamed by some for reduced vigilance, inadequate preparation and flawed decision-making. This paper argues that warning fatigue is a quantifiable multi-faceted construct, and can be demonstrated to influence risk perception in the context of uncertainty.

A warning fatigue measure comprising 10 sub-scales was completed once a month over a six-month period by residents of bushfire-prone Victoria, Australia. Results showed that risk perception of the threat from bushfires changed over time and scores of some sub-scales changed more than others.
This research suggests that far from being a myth, warning fatigue can influence the perception of risk and interpretation of risk messages, requiring emergency and disaster agencies to think differently about warnings in uncertain disaster contexts.

Return to Index

Praveen Maghelal, University of North Texas
Walter Gillis Peacock, Texas A&M University
Xiangyu Li, University of North Texas

Evacuating Together or Separately: Factors Influencing Split Evacuations Prior to Hurricane Rita

One of the long held assumptions of evacuation research is that "households constitute the basic deliberative unit for evacuation decision making." Indeed, Gladwin et al. in their ethnographic decision making analysis of evacuation, build their model around household decision-making. And yet, a recent national research council volume, Facing Hazards and Disasters, suggests that family/household evacuation patterns may well be undergoing change.

Citing work by Dow and Cutter, the report suggests that "even though families decide together to evacuate and wish to stay together, they increasing tend to use more that one vehicle to evacuate…"

This paper addresses directly these potential changes, by focusing on the issue of whether households stay together or leave separately, splitting a part of their response to evacuation. Our focus is on household evacuation in response to Hurricane Rita in 2005. We found that overall 9.3 percent of households reported evacuating in multiple groups at different times, although there were considerable variations, with nearly 17 percent of Galveston Island households splitting compared to 7.3 percent among shadow evacuee households.

Our findings suggest that decision timing, concern about reaching evacuation destination safely, income, number of vehicles and eligible drivers, and elderly household members were all significant determinants of splitting. Interestingly, perception of living in an evacuation zone had no effect on splitting. Consequences for future evacuation policies and research are discussed.

Return to Index

Sarah McCaffrey, USDA Forest Service
Robyn Wilson, Ohio State University
Greg Winter, Cornerstone Strategies

Understanding the Role of Risk Attitudes in Shaping Evacuation Decisions in Response to Wildfire

Research that has examined evacuation decision making has identified that, while some individuals leave early, others wait. Suggested reasons for differential response to when to evacuate have included location, evacuation costs, environmental and social cues, and desire for more information. Few studies have examined how risk attitudes (tolerance/aversion) may shape evacuation decision making. This presentation will report on preliminary results from a recent survey conducted in Texas, South Carolina, and Washington examining planned evacuation decision making in response to a wildfire event.

The study had two primary objectives: (1) understand how risk perception and risk attitudes influence planned response of homeowners to a wildfire; and (2) assess the relative role of risk attitudes to other issues of concern in shaping planned response during a wildfire.

Return to Index

Tara McGee, University of Alberta
Léanne Labossière, University of Alberta

Wildfire Mitigation by Local Governments in Alberta and British Columbia, Canada

Research question: What wildfire mitigation is being carried out by local governments in Alberta and British Columbia, Canada?

Methodology: An email survey was sent to fire chiefs or chief administrative officers in all municipalities in both provinces and regional districts in BC and municipal counties in Alberta, for a total of 371 potential survey participants. In Alberta, 44 surveys were completed and 23 were completed in British Columbia, for a total response rate of 18 percent. Data was analyzed by using descriptive statistics.

Findings: Most participating municipalities had been affected by brush, forest or grass fires in the past two decades. Most participating local governments disseminated information to residents to encourage wildfire mitigation by homeowners, though less than one quarter completed hazard assessments on private property or required residents to mitigate risks on their property.

Vegetation management and using fire resistant building materials for municipal buildings were completed by half of participating local governments, and in over half of participating municipalities wildfire mitigation was taken into account in decision making by land use planners. Emergency preparedness plans for most participating local governments specifically addressed wildfire mitigation.
In this presentation I will compare these findings with the survey completed in 2007, and examine the influence of fire experience on wildfire mitigation efforts by local governments.

Return to Index

Martina McGuinness, University of Sheffield

Planning for Floods and Droughts in the Face of Climate Change—A Continuum Perspective

Water-related risks have emerged as key symptoms of climate change faced by United Kingdom society. Repeated high profile flooding events have resulted in heightened public and policy awareness of this risk. However, the counterpart of flooding, drought, has received much less attention despite recent experience of it in England.

While it appears counter-intuitive that the hazard of too much water might co-exist with the threat of too little water, this is an emerging phenomenon in the UK. The 2007 floods were preceded by two years of heat waves and droughts and in 2009-10, the UK faced rainfall levels lower than the long term average and accompanying drought in some areas.

Just as global warming is leading to milder, wetter winters, so too, increases in global temperatures are also likely to lead to greater levels of evaporation from the land surface resulting in drought in some areas as well as subsidence. Novel responses that are multi-dimensional, involving relevant stakeholders and spanning regulative, cultural-cognitive and normative domains are required for effective mitigation strategies.

This paper presents some initial findings from a collaborative project funded by the UK's Natural Environment Research Council and the Environment Agency. It adopts an integrative approach to understanding climate change perturbations to the water environment. Through a continuum perspective, commonalities which extend across different forms of water related risk are identified as well as barriers to effective management and mitigation.

Return to Index

Martina McGuinness, University of Sheffield
Eve Coles, University of Leeds

In Search of Resilience: Exploring Shifting Paradigms of Contingency Management

Many of the hazards faced by 21st century society reflect changes within the broader risk domain and they are characterized by high levels of uncertainty, unpredictability and emergence: a "risk society." Such hazards can require novel response and regulation.

A major review of contingency planning arrangements in England and Wales took place in 2000-01 resulting in a range of initiatives which spanned the creation of new structures (the Civil Contingencies Secretariat in 2001), statute (the Civil Contingencies Act (CCA) 2004) and process (the UK Capabilities Programme and National Risk Register) aimed at enhancing societal resilience.

While the role of government remains that of guarantor of security for its citizens, this is not a task that can be tackled by government alone. For the effective management of risk, and for resilience to be achieved, participation by a range of public and private sector stakeholders is required.

However, in seeking to provide institutional stability and allay the fears of key stakeholders, government intervention may have the potential for unintended consequences. Arguably, this reflects the dynamic and complex nature of risk and security in late modernity and key challenges facing institutional actors such as uncertainty and the limits of knowledge. These problems, in turn, can adversely affect the way institutions respond to, or prepare for, crisis, particularly in the context of a networked governance framework which is oriented around resilience.

This paper derives from a series of seminars funded by the UK's Economic and Social Research Council between 2011 and 2013 with participants from academia and practice, policy development and implementation, and the public sector and private sector. Based upon these seminars, insights are presented into key themes impacting upon the development of policy, structures and processes in contingency management in the UK, and internationally.

Return to Index

Michelle Meyer, Texas A&M University

Social Capital and Collective Efficacy: Developing Disaster-specific Measures

Two commonly proposed, but under-theorized, components of community disaster resilience are social capital and collective efficacy. These two capture the interactive aspects of a community that imply a capacity to respond, adapt, learn from a disaster, and effectively reorganize community life quickly following an event. Social capital describes the resources available through a social network that can be activated to outcomes for the individual members and the entire network. Collective efficacy, in contrast, refers to the capacity of a group of people to work together for shared goals and has been linked to a variety of collective outcomes such as delinquency rates or disaster recovery.

Much research uses routine measures of these two concepts to predict disaster-specific outcomes, such as Putnam's indicators. Yet, social capital and collective efficacy theorists argue that research should be specified to particular contexts and situations. Based on surveys and interviews with residents of two Florida counties, one rural and one urban, I developed disaster-specific measures of both concepts and discuss how they compare and contrast with routine measures.

My findings highlight distinctions in the practice of disaster-specific social capital and collective efficacy from routine conceptualizations of each. My network-based approach to disaster social capital shows the importance of family, and how these networks may not transform into economic capital. The collective efficacy results highlight how disasters are perceived to bring out the best in neighbors and community members, but that disaster response organizations are assumed to work effectively and efficiently to coordinate collective action, forming the basis of collective efficacy and community resilience. These results have implications for the conceptualization and measurement of both concepts in disaster resilience.

Return to Index

Michelle Meyer, Texas A&M University

The Practice of Disaster Social Capital among Community Organizations: Comparisons and Contrasts Between Two Florida Counties

Disaster social capital represents the resources available through social networks that specifically address disaster mitigation, preparedness, response, or recovery. Social capital is a way to understand disaster impacts and outcomes and is a component of community resilience. In this paper, I focus on organizational social capital to compare and contrast the disaster social capital among community organizations in two Florida counties, one rural and one suburban/urban.

Organizational social capital has been shown to provide resources that benefit the individual organizations and produce collective benefits for the whole network including knowledge creation and sharing, financial resources, and innovations. Further, it can produce greater efficiency and effectiveness especially important to social service and nongovernmental organizations focused on community outcomes. In disaster settings, social capital among emergency management organizations fosters more efficient response. Social capital is central to Federal Emergency Management Agency's "Whole Community Approach" for resiliency (in which four of the six themes focus on social capital).Yet, less research has focused on non-disaster response organizations use of networks for disaster situations.

Based on interviews with nonprofit, emergency management, religious, and social service agencies, I present two typologies of disaster social capital and discuss the perceived benefits of each on community disaster resilience, especially for vulnerable populations. One shows a durable, disaster-specific network, while the other involves assumed response from routine social capital. My results highlight the importance of specifying social capital to disaster concerns and how qualitative aspects of the practice of social capital provides more information about social capital's influence on resilience than quantitative counts of organizations.

Return to Index

Scott Miles, Western Washington University

Theorizing Community Resilience

This presentation describes a new theoretical framework of community resilience based on existing studies on the topic, in addition to other relevant literature. This new framework can be used to propose and evaluate new scholarly hypotheses, empirical data schema, indicator and metric methodologies, computational models, and geovisualization techniques for better evaluating community resilience to hazards.

The framework is composed of two conceptual models for guiding research and tool design. One deals with baseline conditions for resilience and one deals with the dynamics of resilience. The static model of the proposed framework represents two objects of human settlements: community and infrastructure. The static model of the framework includes four variables, each comprised of multiple constructs. The object of community is made up of the variables well-being and identity. The constructs of well-being are satisfaction, affiliation, autonomy, health, security, and material needs. The constructs of identity are equity, esteem, empowerment, diversity, continuity, efficacy, distinctiveness, and adaptability.

As an object within this framework, infrastructure is comprised of the variables services and capitals. The constructs of services include rivalry, excludability, marketability, redundancy, robustness, substitutability, centrality, gravity, and connectedness. The constructs of capitals are social, political, human, economic, built, and natural. Similar to the static model, the dynamic model of the framework comprises three variables: metabolism, geography, and sufficiency. The constructs of metabolism are governance, intervention, provision, consumption, and disruption. The constructs of geography are space, time, scale, and topology. Lastly, the constructs of sufficiency are supply, demand, speed, adequacy, and externality.

Return to Index

Scott Miles, Western Washington University
Nora Jagielo, Western Washington University

Socio-Technical Analysis of Hurricane Isaac Power Restoration

This paper focuses on power outages caused by Hurricane Isaac on August 29, 2012. The hurricane resulted in loss of power to customers in at least 24 Louisiana parishes. The power outage and restoration process led to broad, vocal criticism of the two major private power companies serving Louisiana from the public and government officials.

The goal of this work is to understand how power restoration speed, power utility actions, and other factors influence the real and perceived impacts to customers, as well as reactions of the public and government officials. In other words, are public and government criticisms of power restoration speed rooted in measurable impacts to customers with increasing outage times?

This paper presents an analysis of power outage patterns in context of the social, economic, and ecological impacts related to them. These results are compared with codified reactions to the power restoration activities and performance from a sampling of customers, emergency managers, and government officials. Results provide insights as to whether increasing outage time results in a quantifiable difference in business impacts, as well as whether outage patterns affected some types of customers more so than others.

We characterize restoration decisions of the two major power companies, as well as the context that might have influenced these decisions, to understand organizational practices for improving power system resilience. The results of this work can be used to reduce power outage impacts to businesses, households, and government services. The results of the project will help to set reasonable, objective guidelines and criteria for restoration performance of power companies.

Utility providers and emergency management agencies will benefit from this research through eventual improvement of strategies for power system resilience practices with respect to technical performance, socio-economic impacts, and public or government relations.

Return to Index

Daniel Morath, University of South Carolina

Tweeting from the Front Line: Promises and Pitfalls of Citizen Sensors in Disasters

The shifting nature of communication technology presents new opportunities for monitoring disasters in near-real time. Recent events, including superstorm Sandy, Hurricane Irene, and the Haiti earthquake signify a coming-of-age for citizen sensors in disasters. Crowd-sourced data from popular social networking sites like Twitter, Facebook, and Flickr form the basis of spatially-contextual disaster information referred to as volunteered geographic information. Since VGI data are a compilation of users' contributions, representing a range of intentions, backgrounds, and locally observed conditions, researchers note its potential to facilitate emergency response activities.

Scholars also point to an ethical pitfall in using VGI for decision support because of the digital divide, which refers to differential access to communication media and the information economy as a result of social position. This could lead to a privileging of information produced by affluent users and contribute to increased marginalization of vulnerable social groups. While recent investigations of disaster VGI focus on technical aspects of collecting, standardizing, and visualizing data, relatively few empirical studies have considered both the applicability of message content for emergency management and the ethical issues that complicate the inclusion of such information in decision support.

This study poses two research questions: (1) What types of geographic information (e.g. warnings, impacts, needs, updates) are communicated via Twitter during disasters? (2) In what proportion do various locations (and social groups) contribute to disaster VGI via Twitter?

Data collected during Winter Storm Nemo is examined to illuminate these practical aspects of VGI in disasters. Using an inductive thematic content analysis, tweets are categorized and visualized spatiotemporally. Subsequent comparison of tweet content and frequency in relation to population density and demographic indicators of social vulnerability is used to draw conclusions about the effect of the digital divide on the resulting VGI.

Return to Index

Rebecca Morss, National Center for Atmospheric Research
Julie Demuth, National Center for Atmospheric Research
Jeffrey Lazo, National Center for Atmospheric Research
Katherine Dickinson, National Center for Atmospheric Research
Heather Lazrus, National Center for Atmospheric Research
Betty Morrow, Socresearch Miami

Understanding Public Responses to Hurricane Risk Messages

Recent events such as hurricanes Isaac and Sandy illustrate that weather forecasters and public officials still face major challenges in communicating the risk of an approaching hurricane to members of the public. To improve hurricane risk communication, it is important to understand how different members of the public perceive and respond to hurricane forecast and warning messages, as well as which types of hurricane messages help motivate people to take protective action.

To address these issues, we conducted a survey of members of the public who reside in areas of coastal south Florida who are at risk from hurricanes and storm surge. The survey presented a hypothetical hurricane situation and provided different respondents with different hurricane risk messages, based on modifications of the National Hurricane Center's track forecast cone product. Respondents were then asked about their intentions to take different protective actions (e.g., evacuate, withdraw cash from the bank) as well as questions about their perceptions of the risk and the risk message.

The survey also included questions on respondents' sociodemographics, worldviews, past hurricane experiences, and perceived barriers to evacuation. The analysis examines how respondents' intended protective responses were influenced by their characteristics, their experiences and perceptions, and the different message elements tested. This presentation will discuss initial results from the analysis, informed by risk theories (cultural theory of risk, extended parallel process model) to help understand the findings.

Return to Index

Danielle Nagele, University of Delaware
Julie Demuth, National Center for Atmospheric Research

The Current Weather Warning System in High Impact Areas

Thousands of natural hazards affect the United States each year making obvious the need for an effective warning system with the ability to reduce losses. A "warning system" is typically thought of as the actors, resources, and processes involved in detection, prediction, and communication of impending hazards. Understanding the way this system works and the interactions between each component is imperative in determining what is effective and what needs to be improved.

This research explores a conceptual model of the warning system in order to extend our understanding of the organizations and tasks involved. In addition, this analysis examines the organizational variations that can arise among warning systems in different regions. It is proposed that repetitive impacts from the same hazard can lead to changes in communication structure, actor roles, and use of resources.

This research was conducted using a multiple case study design where organizations located within National Weather Service Warning Forecast Office regions were interviewed. Two of the cases represent areas in which the organizations face repetitive impacts from the same damaging hazard. While all regions experience hazards to some degree, the other two cases represent areas that do not have a particular agent impacting them on a significantly regular basis. The subjects in each case study were drawn from six types of organizations within each area. Interviews addressed tasks and activities associated with the warning system, communication within and across organizations, roles and responsibilities, and the use of resources.

Preliminary findings suggest that new technologies, such as NWS chat and smart phones, have greatly impacted the communication structure of the warning system. Recent years have also seen a stronger integration of the storm spotter network. In general, this is a system that has not remained stagnant over the years, but has grown and evolved much like the public it services.

Return to Index

Hari Krishna Nibanupudi, International Center for Integrated Mountain Development

Dignity and Resilience

Natural hazards are termed disasters when they kill and displace humans, damage livelihood and development infrastructure. Families and societies with adequate resources and capacities are able to avoid or limit the disastrous consequences of natural hazards to an extent and are able to recover whatever the losses they face. People with inadequate resources are forced to live at the front line of the hazards and are unable to prepare adequately for lives and livelihoods in the event of a disaster.

This resource poverty making them vulnerable to disasters, doesn't however, make them poor in dignity and resilience. Women and men from poor communities have shown enormous resilience every time they have faced crises created by natural hazards or political conflicts. However, just because they deal with crisis stoically, doesn't mean they should be left to fend for themselves. In the absence of an international legal framework for the protection of rights and dignity of disaster affected populations, the fate of millions is left to the discretion of local authorities, capacities of humanitarian agencies and multi-dimensional objectives of political and religious charities.

This paper investigates the following research questions:

• What is the meaning of risk and vulnerability from the perspective of people who are exposed to natural hazards?
• What is most significant for people that they can't trade off in the face of a disaster risk?
• Do current risk and vulnerability frameworks understand these terms from people perspectives?
• Do disaster risk management strategies address the risk and vulnerability of people based on their own priorities and needs?

With the help of literature research, interviews with vulnerable communities and select case studies, this paper highlights the concepts of risk and vulnerability from the perspective of communities who are exposed to hazards. The paper argues that it is time that humanitarian actors' push for an international legal framework on the lines of human rights law with the teeth to push for the rights and dignity of natural disaster-affected people.

Return to Index

Jing Pan, Purdue University
Robert Cox, Purdue University
Randy Rapp, Purdue University
James Dietz, Purdue University
Bryan Hubbard, Purdue University

An Exploratory Study of Trust Dynamics in Disaster Recovery Projects

Trust is a core factor in disaster recovery projects on two layers. It first measures project performance by predicting work relationships. It also indicates disaster recovery efficiency on community level. However, no quantitative studies have been conducted regarding trust in disaster recovery projects. As a social capital, trust is vital to the disaster recovery process because of the great uncertainties and vulnerability of disaster survivors.

This research proposes a two-dimensional trust dynamic model. The spatial dimension of the trust dynamic model is a trust measurement model. It includes three latent types of trust—dispositional trust, cognitive trust, and structural trust. The temporal dimension is comprised of short-term trust and long-term trust which is sustained from projects into business alliances or interpersonal relationship. The spatial dimension of the model was visualized using social network analysis.

The model can be applied to many community members in disaster-prone areas to visualize the direct and indirect connections among government, business and communities. The research demonstrates the importance of trust in improving performance in the aftermath of a disaster. By adopting the set of suggested strategies, communities and business could achieve higher efficiency in resource allocation during the recovery process.

Return to Index

Samantha Penta, University of Delaware
James Cox, University of Delaware
Tricia Wachtendorf, University of Delaware

The Value of Preparedness: Organizational Culture and Disaster Preparedness in Delaware Nursing Homes

In August of 2011, Hurricane Irene made landfall on the coast of North Carolina and made its way up the East Coast of the United States. In anticipation of potential flooding because of this storm, one Delaware nursing home evacuated its facility. As a result, the Delaware Department of Health and Social Services Division of Long Term Care Residents Protection sponsored a study to examine the challenges congregate care facilities face regarding disaster preparedness, emergency evacuation, and sheltering.

Interviews and focus groups were conducted with administrators and senior level personnel from 17 skilled facilities in Delaware regarding experiences during Hurricane Irene, and preparedness activity more generally. Taking an inductive approach, this study uses organizational culture theory to explore disaster preparedness in skilled facilities.

The study determines if skilled facilities have a culture of preparedness, and if they do, what that culture looks like, if this culture is industrywide, or if it varies by facility type. Several themes emerged in the analysis, including a capacity for flexibility when necessary, valuing their own experience and the experiences of others in the healthcare setting, and a grounding of risk perception in familiar experiences.

Data suggests that the characteristics of their culture are not universal industrywide, but rather that differences emerge for reasons beyond network involvement.

Return to Index

Laurie Pearce, Justice Institute of British Columbia

Cold Weather Self-Decontamination: Establishing Protocols and Including Psychosocial Considerations for Disabled and Special Populations

Unaddressed, the psychosocial consequences of working in crisis situations can increase the risk of adverse health outcomes, post-traumatic stress, and exacerbate economic and social disruption. Despite potential costs, psychosocial consequence management is rarely systematically or comprehensively addressed in drills, exercises or training or acknowledged as a critical component of effective disaster leadership.

The Simulation Training and Exercise Collaboratory (SIMTEC): Enhancing CBRNE Psychosocial Capacity and Capability Management research project is multi-faceted and includes the development of a series of tabletop exercises. In order to test newly developed cold weather self-decontamination protocols, a drill was conducted in March, 2013 in North Vancouver, B.C., Canada. The purpose of the drill was to determine if self-decontamination protocols could reduce participant trauma, reduce the number of people leaving prior to decontamination and arrange for people to be self-decontaminated in a timely manner.

Those participating in the exercise included persons with disabilities (e.g., paraplegic, blind), children, and special populations (e.g., anxiety disorder). For high risk individuals, professional actors were coached by persons with high-risk conditions. The exercise also involved gross decontamination and technical decontamination (e.g., mobile shower unit).

This presentation will present the preliminary findings of the decontamination drill, some of the lessons learned, and some of the challenges that exist. In all cases, the participants were able to undergo a self-decontamination process which allowed them to remove 80 percent to 90 percent of contaminants and to be in a warming bus before first responders and a hazmat team were able to begin having people come through the mobile shower unit.

Return to Index

Barbara Ryan, University of South Australia

Does Measurement of Intentions Allow Us to Assume Behavior? A Disaster Information Seeking Perspective

The research of disaster behavior and information seeking has shown actual behavior during a disaster and intended behavior, but rarely are the two compared outside disaster preparedness research. The aim of this study is to compare actual behavior with intentions and draw conclusions about whether reported intentions can be used to predict behavior in a disaster.

An online survey of Australians and a mail survey of Queenslanders conducted between September, 2012 and February, 2013 secured 349 responses. Responses came from 107 people with no disaster experience who reported their information seeking intentions. The remaining 242 experienced a disaster in the past two years and reported their actual behavior. Questions included how people learned or would learn of a disaster, where they went or would go for more information, and sources they used or would use during the impact and initial response phase of a disaster.

Limitations of the research included sampling methods (snowball and convenience), which led to a gender bias of 69.9 percent female, a heavy representation of tertiary educated people (56.2 percent of the sample compared with 14.3 percent of the Australian population) and household income (32.8 percent reported household income above AUD$100,000 a year compared with the national average of 14 percent).

Analysis of the results showed little difference between actual behavior and behavioral intentions, even in source preferences and time people took or thought they would take to seek further information after the initial alert. However, in the questions that asked about the importance of sources across the period of information seeking, there were variations of up to 12 percent across the considered importance of some sources.

While more rigorous research needs to be undertaken to confirm or refute the results, this study supports research techniques that use respondent intentions in order to predict information seeking behavior in the impact and response phase of a disaster.

Return to Index

Jenniffer Santos Hernandez, University of Puerto Rico-Rio Piedras

Social Vulnerability as a Social Process: Theoretical and Practical Insights

The study of social vulnerability to disasters has generally focused on understanding how social stratification relates to disasters and how social forces can create the possibility of a hazard to become a disaster. Ben Wisner and his colleagues assembled a working definition of social vulnerability as they were studying the famine that unfolded in the Sahel from 1967 to 1973.

There is no firm definition of vulnerability. Rather it can be considered as an evolving concept. In the book At Risk: Natural Hazards, People's Vulnerability and Disasters, Wisner defined social vulnerability as "the characteristics of a person or group and their situation that influence their capacity to anticipate, cope with, resist, and recover from the impact of a natural hazard." Their definition includes the difficulties and chances that people may encounter as they mitigate, prepare for, deal with, respond to, and recover from the impact of a hazard.

Despite its focus on process, applications of the concept of vulnerability in research often fall short in capturing the elements and dynamics of social vulnerability. This research uses Puerto Rico as a case study in which social vulnerability is conceptualized as stemming from evolving structural and behavioral forces. In a macrocosm, this research explores how the practice of emergency management may impact, address, or fail to address social vulnerability to disasters at the community level.

The findings provide insights that could assist emergency management practitioners and disaster researchers working in the areas of development, emergency management, bureaucratic change, rationalization, decision making, and policy making.

Return to Index

Wendy Saunders, GNS Science
James Beban, GNS Science
Margaret Kilvington, Independent Social Research

Risk-Based Approach to Land Use Planning for Natural Hazards

Traditionally the planning approach for addressing natural hazards in New Zealand has been based on the likelihood of an event occurring. There has been little consideration of the consequences associated with a natural hazard event where it exceeds the design occurrence interval. For example, the recent earthquake sequence in Christchurch demonstrated the potential damage that can occur to buildings and communities when an earthquake that exceeds the Building Act (2004) design standards occurs.

A risk-based planning assessment can be used to address the effects of a particular natural hazard. A risk-based assessment ensures that the economic, environmental, social and cultural consequences of a specific development are explored and quantified as part of future planning decisions. The advantage of a risk-based assessment is that once it has been incorporated into a district plan, it allows for the consideration of the risks associated with both the construction of buildings and a change in use to an existing building. This in turn allows for more robust planning decisions to be made when determining the risks arising from natural hazards on a particular activity.

A risk-based approach to land-use planning is based around five interlinked steps, with associated community engagement and consultation at each step:

• Know your hazard;
• Determine the severity of the consequences;
• Evaluate the likelihood of an event;
• Take a risk-based approach; and
• Monitor and evaluate.

This approach has been developed using participatory action research, which has ensured the output—an online toolbox for risk-based land use planning—has support from planners across New Zealand.

Return to Index

Riccardo Scarpa, University of Waikato
Douglass Beban, Texas A&M University
Mara Thiene, University of Padova

Perceived Risks of Mountain Landslides in Italy: Preliminary Results

Landslides in mountain regions of Italy have regularly occurred throughout history, often resulting in fatalities. Debris can destroy homes and other structures. Policies are being contemplated to reduce the risks of these landslides and a gauge of public support for particular programs is needed.

As a first step, we survey a sample of 168 individuals to obtain estimates of their perceived risks or subjective probabilities of landslides, and mortality risks from these. About half of the sample are visitors to the mountain regions, and the other half own homes there. A split sample design is used so that one group receives science-based risk information, while the other does not. The group that receives the information is allowed to update their initial risk estimates based on the new information they receive. Subjects estimate the risks for others, and separately, the risks for themselves. On average, both the visitors and homeowners who receive the information slightly increase their risk estimate, although the updated estimate remains lower than the implied science-based estimate.

Both the visitors and homeowners underestimate the annual mortality risks from landslides, compared to the science based estimates (about nine per million deaths annually, from 1960 to 2011). Simple regression models (ordinary least squares) are estimated to explain the variation in stated subjective probabilities. For the simple landslide risk (not conditioned on mortality), higher education, fear of avalanches, and having lost a friend or relative result in significant and higher risk estimates. Being female and having a fear of avalanches contributes significantly, and positively to the mortality risk estimate, while more education reduces this estimate.

The subjective probability estimates are used as explanatory variables in a simple, preliminary stated-choice model of programs to reduce the risks. The higher the stated risks, the more likely the subjects are to choose the alternatives.

Return to Index

Ronald Schumann III, University of South Carolina
Susan Cutter, University of South Carolina
Christopher Emrich, University of South Carolina
Daniel Morath, University of South Carolina
Kevin Ash, University of South Carolina

Extending the Recovery Divide: Long-Term Trends and Local Perspectives in Post-Katrina Mississippi

Recovery is the least understood phase of the disaster cycle due to its lengthy timeframe, its variable speed at different spatial scales, and its geographically specific context.  Recovery processes at the household or neighborhood level often lag behind those happening at the larger community or regional scales, necessitating an adjustment in techniques used for monitoring progress in the long-term.

Findings from a seven-year longitudinal study of post-Hurricane Katrina Mississippi show stagnation in the recovery process. Based on regional level longitudinal data, this research aimed to answer the central question "recovery to what?" at the neighborhood scale.

Two research questions frame the study: (1) What spatial disparities in recovery form in the long-term?  (2) What are the driving forces behind these disparities?

Return to Index

Jeannette Sutton, University of Colorado Colorado Springs
Emma Cutter, University of California, Irvine
Britta Johnson, University of Colorado Colorado Springs
Sean Fitzhugh, University of California, Irvine
Ben Gibson, University of California, Irvine
Carter Butts, University of California, Irvine

Warnings Online: A Cross-Hazard Analysis of Twitter-based Warning Message Transmission

Empirical research on hazard warnings has emphasized message content and style as predictors of public protective action taking. Milling activity is one step leading to clarifying, understanding, and personalizing the message before taking action and includes seeking out and passing information along to others. The serial transmission of messages across online networks is one example of milling in response to warnings.

In this research we address the following research question: How does message content, message style, and public attention to tweets relate to the behavioral activity of retweeting a warning message in disaster? We investigate Twitter messages shared by public officials across four hazard events: a wildfire, hurricane, blizzard, and acts of violence.

Using the HEROIC project data collection system, we collect Tweets and network information from a set of targeted accounts in four different geographical areas in the domestic United States. Tweets were coded for their thematic content and message style. We also take into consideration the network characteristics of each official account and any key environmental changes occurring at the time of the warning event. We then create predictive models to show how these variables correlate with retweeting behavior.

Serial transmission and message amplification during the warning period of disaster have the potential to increase message diffusion to reach greater numbers of the population at risk. This study helps to identify the important factors influencing message dissemination, across a variety of hazards, and has important implications for organizational learning at differing scales.

Outcomes from this research will be used for developing guidance for short text warning messages and strategies to increase the reach of messages through networked social media under periods of imminent threat.

Return to Index

Jeannette Sutton, University of Colorado Colorado Springs
Charles Spiro, University of Colorado Colorado Springs
Kotaro Shoji, University of Colorado Colorado Springs
Jessica Lambert, Alliant International University
Britta Johnson, University of Colorado Colorado Springs

Disaster Resilient Rural Communities: Perceptions of Collective Efficacy Following Two Rural Mountain Disaster Events

Community resilience in disaster has become a key issue among those facing seasonal hazard events and unexpected environmental jolts. Strategies to identify what makes a community resilient and to utilize these findings to increase adaptive capacities have increased as disaster losses have grown. In rural communities, where populations tend to be geographically isolated and must rely on local resources, the focus on resiliency is critical. One intriguing community variable with little scientific evaluation post-disaster is the Internet.

This research examines the question: "How does access to online information affect the perception of individual and collective resilience in rural communities across all phases of disaster?" We investigate the interplay among disaster exposure, social capital, individual, and family level coping ability perceptions, and access to information to better understand individual and community resilience.

We present findings from face-to-face interviews with key informants and representative survey of the public population in two rural mountain communities, both of which were affected by wildfire in 2007. We test theoretically based models that focus on social and individual resources in predicting community resilience and compare this between those who indicated a strong use of the internet and those who did not.

Findings from this research will increase knowledge about critical dimensions of rural community resiliency, including community resources, information and communication infrastructures, social capital, and community competence across all phases of disaster, taking into account the individual and community level responses.

Return to Index

Yulia Tyshchuk, Rensselaer Polytechnic Institute
William Benight, Rensselaer Polytechnic Institute

Social Media in Extreme Events

Social media has become important for emergency management. During disasters, it is imperative for people in affected areas to obtain accurate information. People using social media make a conscious decision to trust, act on, propagate, or disregard relevant information. The researchers are studying the linkages among social media message characteristics—such as content, sender-receiver, etc.,—with human behavior in response to the actionable information during extreme events. In order to understand people's behaviors in response to social media information, we must understand the content of their messages and how the messages travel through social media.

First, Facebook was evaluated in the context of 2011 Japan Tsunami to demonstrate the ways emergency management was able to successfully partner with local media. The emergency manager was able to utilize Facebook to develop important relationships with the affected community via social media. The study proposed ways to successfully use Facebook during disasters: by closing a feedback loop between first responders and the public, by monitoring information flow, and by providing regular updates to the public.
The study of Twitter and Twitter networks was conducted in the specific case of 2011 Japan tsunami. The focus was to identify key actors using social network analysis techniques. The formation of cohesive groups in the network and the role of key actors in those groups were evaluated.

Our conclusions show that the key actors played an important role in diffusion of actionable information. The diffusion began with the central and prominent actors and propagated through the gatekeepers of information to the rest of the network. The key actors exhibited intentions to take the prescribed action and urged others to do so.

Finally, the key actors played an important role in bridging the cohesive groups in the network in order to facilitate the diffusion of actionable information.

Return to Index

Shannon Van Zandt, Texas A&M University
Yang Wallace, Virgina Tech
Wesley Highfield, Texas A&M University
Walter Gillis Peacock, Texas A&M University

Inequities in Long-term Housing Recovery After Disasters

Disaster impacts are due to interactions between hazard exposure, physical vulnerability, and social vulnerability. In contrast to their reputation as random and acute events, disasters both magnify and accelerate processes already occurring in communities. During the recovery process, this compression of time and space may exacerbate pre-existing inequities related to household characteristics, housing type, housing tenure, and housing value. Differences in exposure, insurance, financing, and structural quality of homes may further widen gaps between population groups. We examine trajectories of housing recovery as well as obstacles to building back better to answer the question, "Recovery for whom?"

The authors review findings from the literature, including much of their own empirical work from the 1992 Hurricane Andrew in Miami-Dade, Florida and 2008 Hurricane Ike in Galveston, Texas (two of the top five most costly storms in U.S. history) to assess and characterize long term trends in housing recovery. Longitudinal, parcel-level data on both housing units and households permits careful modeling and analysis of the pace of housing recovery for different populations, housing types, as well as the change in the spatial distribution of populations and turnover in housing stock. Supplementing primary data with secondary data from county appraisal districts and city building permits enhances our understanding of the forces at work.

Data show that households and neighborhoods with high levels of social vulnerability suffer greater levels of damage, are less likely to have insurance, are less likely to apply for certain forms of federal aid, are slower to undertake significant repairs to their homes, are slower to pull permits for repairs, and on the whole, are slower to recover.

Over time, the differences in these recovery trajectories may lead to permanent displacement of vulnerable residents and redevelopment of previously affordable housing into less affordable housing types, favoring affluent stakeholders but limiting the availability of much needed housing.

Return to Index

Tamara Wall, Desert Research Institute
Thomas Zang, USDA
Brian D'agostino, San Diego Gas & Electric
Steve Vanderburg, San Diego Gas & Electric

Santa Ana Wind Classification System with Respect to Large Fire Potential

Santa Ana wind events (a type of Foehn wind) generally occur across southern California October-May each year. When the conditions for extreme wildfire events are present, high winds during this time period have led to devastating wildfires in recent years (1993, 2003, 2007, and 2008). To improve decision support services to fire agencies and the general public living in the wildland urban interface areas of southern California, an innovative partnership between San Diego Gas & Electric, USFS, National Weather Service, UCLA, and the Desert Research Institute are working on a classification system and tool to categorize these events.

Integrated with the decision support tool development is a multi-phase social science research project to assess the how this decision support tool can be introduced to and utilized by the general public and regional media outlets. The first phase of this project was a survey (n=459) that was undertaken in 2012 of WUI residents in the Los Angeles and San Diego WUI. The survey identified residents' sources for information about weather-related events, differences in the credibility and utilization of these sources, weather and fire-related information residents believe to be relevant just before and during a wildfire event, and their perceptions of their level of risk from wildfires.

Initial results suggest that WUI residents in the study area have experienced high levels of exposure to wildfire risk and are aware of atmospheric and vegetative conditions that contribute to wildfire events. Perceptions of wildfire risk and mitigation actions, however, vary across the study population.

Return to Index

Lynn Weber, Barnwell College
Anna Rolinski, Barnwell College

Power Relations and Low-Income Housing in Mississippi after Hurricane Katrina: People vs. Ports

On Dec. 31, 2005, Congress allocated $5.058 billion to Mississippi in CDBG block grant funds to be administered through the Department of Housing and Urban Development for "necessary expenses related to disaster relief, long-term recovery, and restoration of infrastructure directly related to [Hurricane Katrina]."

This research examines the power relationships shaping the allocation and distribution processes for these recovery dollars, illustrating who gained, who lost, and how the contest played out. Conducted as a part of an in-depth study of inequalities in Mississippi's recovery process, the research employed a variety of data sources and methods: extensive analysis of public documents, direct observations in multiple locations, and attendance at public forums as well as 56 in-depth, one- to four-hour interviews, conducted in 2007 and 2008 with people active in the recovery process, leaders from—state local governments, businesses, 32 community-based, nonprofit organizations—and residents from vulnerable populations.

The data reveal the contest between a government/corporate/elite alliance promoting large-scale economic development and housing for middle to upper income residents pitted against the ongoing needs for affordable housing among the poor, the working class, the elderly, the single mothers, and the people of color on the Gulf Coast.

This contest is nowhere seen more clearly than in the protracted battle over Mississippi's expenditure of $600 million in CDBG funds to repair the Port of Gulfport and to expand it to four times its pre-Katrina size while the housing needs of Mississippi's most vulnerable populations went largely unmet. Evidence shows how power brokers structured their arguments to justify the port expansion plan as well as the extent to which vulnerable groups and the community-based organizations advocating on their behalf rejected those arguments, offering their own counterarguments and organized resistance.

Return to Index

Hao-Che Wu, Texas A&M University
Michael Smith, Texas A&M University
Carla Prater, Texas A&M University

Perceptions on Hurricane Information and Tracking Maps

It is important to study the process by which people track hurricanes because these storms can cause extensive casualties and damage. The existing literatures are lacking of a scientific understanding of dynamic decision making—especially hurricane information seeking behavior—during events in which additional information is available over time.

The hurricane evacuation decision context is well understood. the National Hurricane Center issues hurricane forecast advisories every six hours until a hurricane turns into a tropical depression. Emergency managers and residents in the risk area are most likely to make decisions on their protective actions based on these hurricane forecast advisories. This study uses the DynaSearch program to conduct a computer-based experiment that examines hurricane information seeking pattern of students playing the roles of county emergency managers, their understanding of hurricane strike probabilities and their choices of protective action recommendations during four different hurricane scenarios.

This study simulates the approach of a hurricane by providing experiment participants a sequence of hurricane forecast advisories and examining how they search for information, change their threat perceptions and implement protective actions over time. The results show: (1) hurricane intensity was the parameter that was most commonly viewed and hurricane wind radius was the parameter that was least commonly viewed; (2) forecast track had a large number of clicks and long click durations, on the other hand, Uncertainty cone had fewer clicks but longer click durations; (3) participants' judgments of the extent to which they used each of the parameters are not entirely consistent with their search patterns; and (4) participants were able to utilize the available information in the tables and tracking maps to make reasonable judgments about each city's relative strike probability.

Thus, these results suggest the problem of misinterpretation of the uncertainty cone is less severe than some might have concluded from the evidence provided by Broad et al. (2007).

Return to Index

Clayton Wukich, Sam Houston State University
Michael Lindell, University of Illinois At Chicago

Texas Wildfires: Measuring Interorganizational Response and Recovery Networks

In 2011, the people of Texas endured a historic drought that, in part, led to devastating wildfires largely in rural areas. Over 25,000 separate fires burned nearly four million acres and destroyed 5,599 structures, including 2,862 homes. This paper evaluates interorganizational response and recovery operations during the wildfires, mapping the dynamic, shifting models of governance employed.

The authors identify the capacity of emergency managers and their communities to recognize risk and coordinate with other actors across levels of government and social sectors. Using social network analysis, the authors model the structure of these response and recovery efforts at the macro, meso, and micro levels and make propositions that link these network structures to performance.

Data are derived from 150 Texas situation reports, which serve as official records of action and interaction during operations. In all, they report over 20,000 interactions that differ by a range of indicators including emergency support function and intensity of the fire. The authors employ an innovative data coding and analysis strategy, which integrates multiple techniques from Koliba, Meek, and Zia (2011), Comfort and Haase (2006), and Axelrod and Cohen (1999).

This paper explores a variety of approaches used by organizations to manage these large-scale events. In doing so, the authors identify, in part, the levels of resilience demonstrated by emergency managers and their communities. In addition, by exploring multiple performance indicators, the paper illustrates the complexity of the action arena and considers a wide range of strategies required to manage future incidents. Policy recommendations are developed to augment future preparedness and response operations.

Return to Index

Yu Xiao, Texas A&M University
Walter Gillis Peacock, Texas A&M University

Do Hazard Mitigation and Preparedness Reduce Physical Damage to Businesses in Disasters: The Critical Role of Business Disaster Planning

This paper provides fresh evidence on the status of business disaster planning, mitigation, and preparedness, and the effectiveness of such activities for loss reduction.

We analyzed data from a survey of businesses in Galveston County, Texas conducted seven month after Hurricane Ike struck the area in September 2008. Unlike previous studies, we found business disaster planning significantly promoted hazard mitigation and preparedness, and therefore, protected businesses from physical damage.

Return to Index

Joslyn Zale, The University of Southern Mississippi
Joby Bass, The University of Southern Mississippi
Bandana Kar, The University of Southern Mississippi
James Dickens, The University of Southern Mississippi

Availability and Coverage of Communication Platforms for Warning and Alert in the Mississippi Coastal Communities

Technological advancements have led to development of different communication platforms, both hierarchical and network, to provide alert and warning messages to at-risk populations to enable effective and efficient remedial actions be taken by impacted populations. Irrespective of the advancements in communication platforms, the success of any alert/warning system depends on the response of the recipients to alert and warning messages. The main purposes of this project are: (1) to identify communication platforms available for warning and alert dissemination during emergencies; and (2) to determine the comprehensive coverage provided by all platforms so that steps can be taken to address the need of at-risk populations lacking coverage by any communication platform.

The main research questions explored are: (1) What are the available platforms for alerting public in the study counties? (2) Which warning platforms are used by emergency management agencies/emergency responders to warn and alert public? (3) What is the coverage area of each platform and all platforms? (4) What is the comprehensive coverage of all the platforms? and (5) What is the ranking of each platform based on area covered by them?

This study was conducted for the three coastal counties of south Mississippi (Hancock, Harrison, and Jackson counties), a region susceptible to tropical storms and hurricanes every year during hurricane season: June 1st to November 1st. In addition to its physical vulnerability to natural and human-made hazards, the region is occupied by multi-cultural communities and expansive rural areas along with heavily populated coastline.

Return to Index

Haibo Zhang, Nanjing University

How Do Rural Residents Improve Readiness for Disasters in Eastern China? Three Possible Social Mechanisms and Empirical Testing

How to facilitate the public to improve their readiness for disasters is a permanent research issue. Since the SARS outbreak in 2003, China has begun to establish a comprehensive emergency management system, which did improve the readiness of the government.

However, efficient disaster management depends on joint efforts by governments and the public. Compared to the rural residents who suffered and learned from the Wenchuan Earthquake and the Yushu Earthquake, the rural residents in eastern China improve their readiness for disasters in everyday life.

How do they do that? Although eastern China has not seen a large scale natural disaster serving as the focusing event in recent decades, it is definitely at a high risk of large disasters like earthquakes.

The paper primarily proposes three kinds of social mechanisms which may be able to work to improve their readiness for disasters in the everyday life: (1) self-learning based on knowledge level; (2) economy driven based on income promotion; (3) political motivation based on policy implementation.
Cooperating with the Provincial Emergency Office of Jiangsu Province, the paper used a quota sampling survey and to collect data in Jiangsu in 2011, and tested all the three hypotheses. The primary finding is that the political mechanism works significantly, while the others need to be developed.

Return to Index