Departure Timing Preference during Extreme Weather Events: Evidence from Hurricane Evacuation Behavior

Abstract

Hurricane evacuation has become an increasingly complicated activity in the U.S. as it involves moving many people who live along the Atlantic coast and Gulf coast within a very limited time. A good deal of research has been conducted on hurricane evacuation, but only a limited number of studies have looked into the timing aspect of evacuation. This paper intends to contribute to the literature on evacuation timing decisions by investigating what factors influence the time preference at the household level. Two hurricane survey data sets were used to analyze household evacuation behaviors across the Gulf coast as well as the Northeast and Mid-Atlantic coast in a comparative perspective. Using the Heckman selection model, we examined various factors identified in the literature on the two possible outcomes (evacuation and early evacuation). We found that the most important determinants of evacuation were prior evacuation experience, evacuation orders, and risk perceptions, while the most important determinants of early evacuation were prior evacuation experiences, days spent at the evacuation destination, and the cost of evacuation. Socioeconomic factors also influenced the two decisions but differently. These results provide implications for future hurricane evacuation planning and for improving emergency management practices.

A data‐driven method for identifying the locations of hurricane evacuations from mobile phone location data

“…The increase in the frequency and intensity of hurricanes due to climate change highlights the need for robust evacuation planning. Such plans require knowledge and assumptions of the mobility patterns of individuals and households (Lindell & Prater, 2007b), which may include information on departure time estimates (Jiang et al., 2022; Ng et al., 2015; Sorensen, 1991), the number of vehicles used per household (Lindell & Prater, 2007b), trips taken to prepare for an evacuation (Lindell et al., 2019; Lindell & Prater, 2007b; Urbanik, 2000; Urbanik et al., 1980), picking up children (Liu et al., 2012; Murray‐Tuite, Schweitzer, et al., 2012), and the likelihood of using a shelter (Li et al., 2012; Lindell et al., 2019). These mobility patterns have often been assessed through the use of surveys collected after a hurricane; however, in recent years, mobile phone location data has been used to examine some of these behaviors (e.g., Long et al., 2020; Tao, 2021; Yabe & Ukkusuri, 2020).…”