Examining Factors Influencing People's Perceived Vulnerability and Evacuation Decisions in Response to Hurricane Irma in Charlotte County, Florida

Hurricane evacuation refers to an individual’s or household’s response to a hurricane threat by temporarily relocating somewhere safe. However, it is rarely the case that everyone will comply when he/she is ordered to evacuate by authorities. Understanding why people fail to heed evacuation orders is vital in order to minimize non‐compliance. In the past, factors contributing to hurricane evacuation behavior have been studied mainly using survey instruments. The most significant variables include official evacuation orders, housing type, perceived risks, environmental cues, and social cues. Factors such as information sources, experience, and demographic characteristics have weaker impacts on evacuation decisions or are significant in only a limited number of studies. In this study, many of the literature‐documented factors that contribute to evacuation behavior were examined using survey data collected in Charlotte County, Florida, before and after Hurricane Irma (2017). Additional environmental factors, some of which were not included in a majority of the previous hurricane evacuation studies, were examined together with the survey‐based, socioeconomic, and risk‐perception variables. Environmental factors examined less frequently include variables such as the structural characteristics of people’s homes, storm surge zone, wind zone, and distance to shelters. Environmental variables examined in the previous literature include elevation, flood zone, and distance to the coastline. These data were collected through publicly available data sources and linked with the survey respondents’ home locations. The literature review points to the importance of risk perception on people’s hurricane evacuation decision‐making, so factors that may have contributed to people’s hurricane risk vi perception were also examined. These factors include both the survey‐based socioeconomic variables and the GIS‐derived environmental factors described above. Risk perception in this study was measured through the respondent’s answers to a question regarding their sense of safety in a hypothetical Category 3 hurricane. Since an individual’s perception of safety is an overall measure of whether people think they are threatened by particular hurricane‐induced risks (such as flood damage, wind damage, and power loss), the term perceived vulnerability was adopted and used throughout the entire thesis. In this study, perceived vulnerability refers to a respondent’s overall sense of safety when they consider all the potential threats associated with a Category 3 hurricane. Descriptive statistics, chi‐square test, and independent samples t‐test were performed to assess for associations between independent variables and the two dependent variables. One of the dependent variables is the evacuation decisions made by the survey respondents in Charlotte County in response to Hurricane Irma. The other dependent variable is perceived vulnerability. Logistic Regressions were performed for each dependent variable with only significant factors (according to the chi‐square or t‐test analysis) input as the independent variables. The results showed: 1) For survey respondents in Charlotte County living within evacuation zones A and B (both were under mandatory evacuation orders during Hurricane Irma), education level, perceived hurricane evacuation zone, perceived vulnerability, and distance of residence to the nearest coastline were statistically significant factors associated with the evacuation decisions made during Hurricane Irma. 2) According to Logistic Regression, perceived vulnerability and the distance from the subject’s home to the nearest coastline were vii the most significant factors influencing the evacuation decisions made during Hurricane Irma. 3) For survey respondents living across Charlotte County regardless of evacuation zones, many socioeconomic, risk perception, and environmental factors were statistically significantly associated with their perceived vulnerability. 4) When these factors were included in the Logistic Regression analysis, the respondent’s storm surge zone and their belief that their houses were located in a low‐lying area, were the factors found to be most significantly associated with perceived vulnerability. This study contributes to the assessment of people’s hurricane risk perception and evacuation decisions. It also demonstrates the benefit of including environmental variables in hurricane risk perception and evacuation decision‐making research. Many factors, including less‐researched environmental variables, were integrated into this comprehensive study. Perceived vulnerability, a measure of overall perceived risks, was found to be the most significant factor associated with people’s hurricane evacuation decisions. It was found more important than whether people believe they heard evacuation orders or what type of evacuation order they heard. Evaluating what determines people’s perceived vulnerability is a relatively new research direction in studies examining hurricane evacuation behavior. In this study, many of the environmental variables (such as elevation, the distance of the home to the nearest coastline, evacuation zone, and storm surge zone) were found to be significantly associated with people’s perceived vulnerability. For the subjects in this study, fear of flooding was identified through the two significant variables standing out in the Logistic Regression analysis – respondent’s storm surge zone (an environmental factor) and respondent’s belief that their home is located viii in a low‐lying area (a specific risk perception factor). This finding may be useful to government officials when they communicate with the public prior to a hurricane making landfall. Including information such as flooding risks may motivate residents to comply with evacuation orders.