Abstract
Climate change has increased the destructive force of natural hazards and the occurrence of disasters that damage housing and infrastructure and threaten the health and wellbeing of human populations. Demographic research on disasters advances understanding of climate impacts on coastal populations exposed to tropical cyclones, storm surge, and flooding. However, rigorous study designs that allow for inferences about causal mechanisms are needed. We review natural experimental study design elements and research findings from four demographic surveys disrupted by large-scale disasters. Two studies focus on the effects of Hurricane Katrina on New Orleans' population and the other two on the 2004 Indian Ocean and 2011 Tōhoku Tsunami effects in Indonesia and Japan, respectively. Three elements of the studies' designs support the most novel findings: measures of pre- and post-disaster demographic and health outcomes; multiple post-disaster follow-up data waves; and measures of variability in hazard exposure and impact. Two other design elements, population representation and the existence of an unexposed control group, are less critical. Study results advanced knowledge of the effects of damage-related displacement on mental and physical health and wellbeing, as well as harder-to-observe effects of disaster exposure on mortality and fertility outcomes. Acknowledging that natural experiments are rare, we evaluate opportunities for research on hazard exposures and population outcomes using administrative data, existing panel surveys, and new retrospective surveys. Linkage of these data sources to hazard exposure data can expand geographic and population coverage in this field and accelerate understanding of climate-related impacts on population outcomes. This article is categorized under: The Social Status of Climate Change Knowledge > Sociology/Anthropology of Climate Knowledge Assessing Impacts of Climate Change > Observed Impacts of Climate Change Perceptions, Behavior, and Communication of Climate Change > Behavior Change and Responses.