Assessing perceived probability and damage of flood risk across the globe

In this study we aim to improve our understanding of household adaptation and migration responses to coastal hazards induced by sea level rise, such as coastal flooding. We apply a global ABM (agent-based model; DYNAMO-M), which simulates all ~200 million individual people and households in coastal flood zones around the world. The model simulates in yearly timesteps flood events and changing flood risk and whether residents migrate or adapt (e.g. flood proof their house) to lower their risk. Agents’ migration and adaptation decisions are based on the Subjective Expected Utility Theory (SEUT). Here, agents maximize their utility based on subjective risk assessments, such as their subjective perception of flood risk. However, the current risk perception parameter in the SEUT equation in DYNAMO-M is based on a single empirical study in France. Additional data is needed to address the heterogeneity of risk perceptions across different global coastal households. In order to assess the differences in risk perceptions in different areas around the world, we combine different data sets: (1) we conducted unique surveys on perceptions of flood risk and their determinants as well as people’s intention to adapt or migrate under future SLR scenarios in 6 countries with varying socio-economic backgrounds (Argentina, France, Mozambique, the Netherlands, the US, and Vietnam). Using these survey data, we identify the generic decision rule for the determinants of risk perception parameters such as the perceived probability and damage of flooding through regression analysis. (2) Next, we use additional global datasets on individual characteristics such as the World Value Survey (demographic and residential information data) and Cloud2Street (flood experience data) and use these data as explanatory variables for transferring risk perception parameters to countries where no primary survey data is available. This analysis may aid the understanding of global patterns in risk perceptions of people/agents. We believe our study serves as a basis for research on individual behavior under risk, the role of risk perception, and the use of the data in global ABMs.