Examination of Community Responses to Hurricane Evacuation Orders Using High-Fidelity Mobility Data

Hurricanes pose a significant threat to a large portion of the U.S. population in coastal areas. Evacuation is the most common preventive risk reduction measure adopted by coastal communities in response to hurricanes. However, the success of evacuations depends on whether those at risk obey the directions. Therefore, evacuation decision-making and various factors that may impact it have been researched for over five decades. However, data collection and methodological limitations still hinder quantitative inferences on the impact of evacuation orders. In particular, there is a lack of comprehensive research exploring the causal effect of evacuation orders on influencing evacuation behavior. Notably, no historical records of evacuation orders exist to investigate such questions thoroughly. Although these orders are generally assumed to be followed in theory, in practice, evacuation rates vary widely among communities. Additionally, the understanding of the impact of socioeconomic disparities on evacuation is narrow. Specifically, the majority of studies on income and race disparities focus on a single hurricane event affecting a single state, and differences in their study designs prevent comparative or meta-analysis for broader understanding.

This research utilizes near-real-time high-fidelity mobility data to bridge existing gaps in understanding community responses to evacuation orders during hurricane events. To achieve this, we first developed the Hurricane Evacuation Order Database (HEvOD), a comprehensive, high-temporal-resolution repository that includes evacuation orders issued across the U.S. between 2016 and 2022, meticulously compiled from diverse sources including official announcements, social media, and news outlets. Equipped with this rich dataset, we assess the effectiveness of these orders by investigating the causal relationship between mandatory evacuation orders and observed community mobility patterns, specifically using data from Florida during Hurricane Dorian. Then, the role of socioeconomic and demographic factors in evacuation decisions of communities in response to government-issued evacuation orders is explored. This is done, by applying a consistent study design across seven hurricanes affecting twelve states, marking this as the first study in the literature that uses mobility data to study evacuation behavior across multiple hurricanes, providing a comprehensive understanding of income and race disparity in evacuations. The synthesis of these research objectives promises significant societal advancements, providing a foundation for future research to refine evacuation strategies and policies, leading to more inclusive and effective emergency response frameworks.