| description abstract | Evacuations require people to leave their homes and seek temporary accommodations for varying lengths of time. However, although studies of evacuation and return-entry examine travel to and from short-term evacuation destinations, prior work overlooks relocations from one destination to another during long-term evacuations. In this study we use novel spatial-social network data from a household mail survey administered in three US regions (Akron, OH, Philadelphia, PA, and State College, PA) to examine people’s decisions to relocate from short-term to long-term evacuation destinations. To understand these decisions, we used a CatBoost classification algorithm to examine the significance of demographic and evacuation-related factors, including social support from family and friends at evacuation and relocation destinations, that affect relocation decision-making during evacuations lasting up to two months. We find that a majority of respondents plan to remain at a single destination during a long-term evacuation and tend to receive support, such as accommodations, from family and friends at these destinations. In contrast, respondents who plan to relocate tend to lack support from family and friends at long-term destinations. While relocation decisions are influenced by multiple factors, travel to remote evacuation destinations and lack of social support at these destinations have the greatest influence on relocation decision-making. In contrast, no demographic attributes show statistically significant correlation with relocation decisions. By revealing evacuees’ intended short- and long-term evacuation and relocation destinations, and social support available to evacuees at these locations, the study’s findings offer implications for the strategic management of large-scale evacuation and return-entry processes. | |
