There are three key research activities:
1. Model: Core of the project is the DYNAMO model which integrates a coastal flood risk model (GLOFRIS) with an agent-based model (ABM). DYNAMO is recently implemented for France and will be upscaled to other global coastal regions. The decision rules in the ABM are based on subjective discounted utility theory, which describes how people make decisions under flood risk. Based on this theory, the agent-based model simulates the adaptive behavior of agents under flood risk (simulated by the GLOFRIS module). Agents include: households in the coastal zone, governments and the private sector (e.g. insurance). A gravity module then simulates the aggregated migration flows between inland regions and the coast. The DYNAMO model is calibrated with empirical data from surveys in France. Results show that from the in total 260,000 people in flood zones in France, SLR increase net out-migration exceeding 10,000 coastal inhabitants. This number is low because it is assumed that still people move to the coast because of socio-economic pull factors. These projections of coastal migration under SLR are most sensitive to migration costs and whether the government has installed coastal flood protection.
2. Surveys: With the survey research we aim to understand the determinants of why people migrate, adapt or do not adapt or migrate, both at present and under future scenarios of sea-level rise. The survey data will serve as input for the agent-based model in DYNAMO. We have surveyed people in France, the United States (Florida/NYC), Buenos Aires, Vietnam, Mozambique. First results indicate that not only flooding is associated with increased migration flows, and place attachment, risk perception and financial drivers play a large role in migration decisions. Salinization and erosion do generally lead to more out-migration.
3. Up-scaling drivers: We apply novel global databases on flood protection, demography and socio-economy and machine learning to extrapolate behavioral rules on migration and adaptation decision from the regional coastal cases to other global coastal areas. We have started this activity by analyzing the impacts of natural hazards (floods, forest fire, earth quakes) on internal migration in the United States. In an econometric analysis we search for a correlation between observed migration flows, damage from natural hazards, and socio economic data on households (income, unemployment, etc.). Next, the aim is to create a database with global agent populations and their characteristics, which can serve as a basis for the DYNAMO ABM
We have communicated our results at conferences such as EGU and the migration conference at Columbia University in NYC in 2023 (
https://adaptation.ei.columbia.edu/retreat/home(s’ouvre dans une nouvelle fenêtre))
