Modeling the Real-Time Decision to Evacuate from a Hurricane

While there has been a significant amount of research dedicated to better understanding a household's decision to evacuate from a hurricane, routinely these studies are conducted in a post-storm environment months or even years after the event. Unfortunately, this post-storm research is subject to the significant survey issue of recall, whereby appropriate recollections of the facts surrounding one's decision to evacuate are likely suspect stemming from the passage of time. In order to remedy the recall issue we have collected survey information on a household's decision to evacuate while a hurricane was actually threatening. We present an analysis of this intended household evacuation decision for two hurricanes that impacted both the mid-Atlantic and northeastern regions of the United States – Hurricane Earl in 2010 and Hurricane Irene in 2011. Significantly, our survey covers varied geographic areas – North Carolina, Massachusetts, and New York – allowing for the evacuation decision analysis of two different types of hurricane experienced residents from a general perspective. Moreover, the survey instrument specificity allows for the analysis of the evacuation decision accounting for a multitude of relevant factors not traditionally analyzed including storm and forecast awareness, perceived vulnerability and concern, preparation, and sources of forecast information. Thus, the results from our study allow for a better understanding of the actual factors that various types of households consider as they make their evacuation decision in varied geographic areas as the storm threatens and evolves. Understanding this is critical in moving forward on the construction of forecast information that households can and actually do use to make better evacuation decisions.