Spatially explicit agent-based approach for human–flood interaction modeling under external support

Abstract

Two-way feedback between humans and flood systems is important for a better understanding of flood resilience and how flood risk evolves over time through human adaptation. However, most studies on this subject have assumed human traits and actions to be uniform, overlooking the diversity of human responses to floods. Furthermore, in many developing countries, the flood protection infrastructure is co-managed and co-maintained through the involvement of local communities as well as external or governmental support. However, the influences of differing levels of external intervention on the long-term outcome of co-management and on community behavior remain underexplored. In this study, we aimed to overcome this issue by considering a community as a heterogeneous group under flood risk via geographically explicit agent-based modeling. This method allowed us to computationally experiment on how human–flood interactions are conditioned by variations in household-level attributes and governmental support. Specifically, our study considered the community-based flood protection system in the Ganges-Brahmaputra River basin in Bangladesh. Our results suggest that heterogeneities in the geographic location, income, and flood damage have significant effects on the outcome of human–flood interactions, specifically cooperative efforts toward minimizing flood damage and maintaining flood protection structures. Our results also highlight that a moderate level of governmental support (i.e., assisting with only a part of the required labor) is necessary to maintain the infrastructure built to promote cooperative action in a community-scale flood protection system. Overall, we believe our agent-based model provides a potential for investigating the community dynamics from modeling individual-level human–flood interactions, ultimately provide a guidance for planning future governmental support for self-managed flood protection system.