Migration patterns and household adaptation dynamics in response to hurricanes and floods

In recent years, anthropogenic climate change has moved from a future concern to a current reality (Trenberth 2018). Now impacting every region on the planet, the effects of climate change are “widespread, rapid, and intensifying” (IPCC 2021). It is therefore pertinent to think of how we can adapt to a changing and more hostile world (Robinson et al. 2020). To this end, this dissertation focuses on environmental migration outcomes. This is accomplished with three distinct analyses. The first study uses a multi-level, mixed-effects analysis on responses to two waves of a survey distributed across the Albemarle Pamlico Peninsula, North Carolina to consider how residents’ opinions and beliefs change as they are exposed compounding disasters. The findings show that saltwater intrusion, a slow-onset hazard, is associated with heightened risk perception; and that repeated hurricane exposure, saltwater intrusion, and heightened risk perceptions are associated with a greater acceptance of migration as an adaptive strategy. The second study utilizes survival analyses on a novel dataset to track movement for all individuals with credit records in eastern North Carolina, considering compounding hazard exposure and pre-established vulnerability. It finds that the co-location of exposure and vulnerability is associated with more resilient migration patterns. However, state-level vulnerability remained constant as immigration into floodplains occurred at approximately the same rate as out-migration. This “vulnerability replacement” dynamic represents a missed opportunity to capitalize on the organic post-hurricane process of flood vulnerability reduction to achieve broader improvements. The third study utilizes propensity score matching on another novel dataset to analyze post-buyout migration in Harris County, Texas, after Hurricane Harvey (2017) in contrast with a counter-factual group of residents who moved from similar areas within the same time frame without the benefit of federal support. The results show that buyout participants were less likely to move into a floodplain than members from our control group. However, buyout participants did not experience significantly different outcomes to those of non-buyout movers. The findings suggest that buyout programs have a limited influence on the established migration system within which they operate