Advancement in ice-jam flood risk management: Integrating dynamic adaptive behavior by an agent-based modeling in Fort McMurray, Canada

•Traditional ice-jam risk assessment lacks a consideration of evolving human behavior.•We built an agent-based model to assess the long-term ice-jam flood risk.•This study highlights the role of dynamic adaptive behavior in ice-jam flood risk.•Artificial breakage can lead to a regime shift in the ice-jam flood risk.•Artificial breakage can decrease the number of adapted residents to flood-proofing. Ice-jam flooding is a prevalent extreme event that impacts flood hazard and vulnerability. We introduce a conceptual model framework for Dynamic Ice-jam Flood Risk Assessment (DIFRA). DIFRA integrates ice-jam flood hazard, ice-jam flood risk, and human adaptation. Using agent-based modeling, we captured top-down (artificial breakup) and bottom-up (flood-proofing) adaptive behavior. Our study in Fort McMurray, Canada, shows the complex interaction between micro-level behaviors and macro-level phenomena over time. Our variance-based global sensitivity analysis shows the role of dynamic adaptive behavior in ice-jam flood risk, where the artificial breakage by the government can lead to a regime shift and a decrease in the ice-jam flood risk. However, it can also decrease the number of newly adapted residents to flood-proofing and the role of residents in ice-jam flood risk. DIFRA offers a comprehensive approach to understanding and managing ice-jam flood risk, with potential applications to similar riverine communities in cold regions.