COFLAB: Coupling Flood and Agent-Based models for interpreting centuries of River-Urbanism Interactions
Land use changes, rapid urban expansion, and increased flood risks under climate change are challenging settlements along the world’s big rivers, where some of the earliest examples of urbanism emerged. Despite billions of investments in levees, floods are still the most damaging disasters globally (~US$60 bn in a year). Indeed, it has even been argued that some technocratic solutions may weaken long-term resilience. This project seeks to investigate long-term flood risk, resilience, and adaptation strategies in the Yellow River Basin (YRB) by combining Earth system modelling, agent-based modelling (ABM), and historical data over centuries.
Using hydro models, Dr. Shuang Song will simulate flood inundations across the YRB, reconstructing flood dynamics from 1400 CE. A central focus of this study is the “levee effect,” where protective structures like levees can increase long-term vulnerability by diminishing collective memory and reducing adaptive flexibility. Furthermore, this project will develop a novel ABM to test the “levee effect” with spatial heterogeneity based on multidisciplinary data from paleoclimatic, archaeological, and paleo-environmental records in the YRB. Through further scenario analysis, the coupled model will allow me to explore the spatial and temporal variations in resilience caused by land use and flood adaptations, thus inspiring modern flood risk management.
Research Objectives
This project includes three research objectives (ROs):
- This project proposes to develop a long-term flood inundation model driven by paleoclimate forcing to interpret how deforestation led to spatial heterogeneity in flood risks by comparing different land use scenarios.
- This project will collect data on disasters and adaptations of floodplain cities, and compare their spatiotemporal patterns with simulated palaeo-floods, to assess the resilience of long-term urban flood risk management.
- This project will develop an ABM and test the long-term nature of the “levee effect” to observe how different flood loss patterns emerged from strategies of building levees and cultural memory.
This project is undertaken within the Max Planck Postdoctoral Program framework (2025 cohort).











