An Agent-Based Model for Simulating Flood Governance and Community Resilience

Agent-based modeling (ABM) is a unique tool for understanding social mechanisms and emergent phenomena. The paper presents an empirically grounded agent-based model that simulates how stakeholders embedded in flood governance networks facilitate community loss-sharing and post-flood recovery. The model is designed and calibrated using extensive empirical data from communities in Guangzhou, China. Modeled agents include multi-level government agencies, NGOs, the private sector, and local clans, among others. The model integrates core processes (rainfall and flood impacts, network-based loss sharing and recovery, and the implementation of resilience measures) with modules about trust evolution and resource constraints. The purpose of this model is to evaluate the effects of different network structures, inter-stakeholder trust, and the diffusion of flood resilience measures on community flood resilience, and to advance the understanding of how resilience emerges as a macro-level attribute from micro-level interactions. Innovations are twofold: First, it moves beyond static analysis to simulate the dynamic, network-based collaborative processes among diverse institutional stakeholders; Second, it implements a process-based framework to measure community robustness and adaptivity, using these metrics to evaluate overall community resilience to floods. Key parameters, derived from literature and empirical research, were empirically validated and tested via sensitivity analysis. The model serves as an accessible tool for researchers and practitioners interested in stakeholder collaborations in community-level climate governance and identifying optimal intervention strategies.