Flood resilience through multi-actor coordination: Insights from the 1870 mega-flood in the upper Yangtze River

The participation of diverse stakeholders is considered essential in building flood resilience. However, empirical studies remain insufficient to clarify the roles and interactions within multi-actor networks and their implications for resilience, especially in the historical contexts. Taking the 1870 M-flood in the upper Yangtze River as an analytical window, this study integrates social network analysis with historical analysis to examine how flood resilience was manifested through the coordinated actions and interactions of multiple actors. Extensive historical materials provide the data sources, including documentary records, flood inscriptions, and archived interview transcripts. The results reveal a centralized and hierarchical governance network involving multiple actors. Actors with greater scale and resources operated at higher levels of the social-ecological system. Individuals and households focused mainly on self-preservation, while family, clan, and community ties facilitated mutual assistance in emergencies. Local governments served as both decision-makers and implementers of response measures, coordinating the contributions of non-governmental actors, including local gentry and charitable organizations. Provincial governments set strategic direction and reallocated financial resources, whereas the central government shaped responses indirectly through a highly institutionalized relief policy framework. Information flow within the governance network supported decision-making and enabled the redistribution of material resources. Disaster experience also fostered social learning, reflected in heightened risk perception, the embedding of disaster memory in local knowledge systems, and the incorporation of effective practices into routine risk management. Despite this institutionalized governance structure, its effective functioning depended heavily on local fiscal and administrative capacities, resulting in uneven performance across counties.