Network Resilience in Disaster Response: Integration and Fragmentation in Polycentric Governance Systems

• In Event: Disaster Responses in the US States

Abstract

As disasters become more frequent and complex, effective response increasingly depends on interorganizational networks that coordinate across agencies, sectors, and levels of government. Network performance rests not only on formal structure but also on resilience—the capacity to anticipate risks, absorb shocks, and adapt while sustaining critical functions during disruption. Yet the role of state-level governance arrangements in shaping the resilience of interorganizational response networks remains underexamined. This gap raises a central question with direct implications for state policy design: under what governance conditions do response networks remain coordinated rather than fragmented when stressed? This study addresses that question by comparing two recent disasters that unfolded under contrasting state response systems: the 2024 Baltimore Key Bridge collapse in Maryland and the 2025 Los Angeles wildfires in California. Although the hazards differ in scale and type, both required rapid multi-jurisdictional coordination. Importantly, both cases exhibit polycentric characteristics, but differ in the degree of integration, coordination coherence, and clarity of leadership. Maryland's response was organized through a tightly coordinated, functionally differentiated structure in which multiple unified commands led distinct operational domains while remaining integrated within an overall network. California's response unfolded through a broader and more fragmented polycentric system shaped by overlapping jurisdictions, strong local autonomy, and extensive but loosely integrated mutual-aid arrangements. This contrast renders the paired cases analytically valuable for assessing how different configurations of polycentric governance condition network behavior and resilience under stress. The study employs a comparative mixed-method design. Qualitative content analysis draws on after-action reports, news articles, and press briefings to trace leadership practices and communication flows. Social network analysis, constructed from documented interorganizational ties, measures structural properties relevant to resilience, including density and centralization, and tracks how these properties evolve. Resilience is operationalized along three dimensions: anticipatory capacity, reflected in pre-event arrangements and early warning; absorptive capacity, reflected in the maintenance of core functions under shock; and adaptive capacity, reflected in the reconfiguration of ties and roles under changing conditions. Preliminary analysis suggests that the two systems differed sharply in both coordination pathways and resilience outcomes. Maryland’s tightly coordinated and functionally differentiated polycentric structure supported rapid convergence, strong information integration, and coherent interorganizational action during the acute phase. California's broader and more fragmented polycentric structure revealed serious coordination failures, including under-activation, leadership ambiguity, and communication breakdowns, which undermined the network's ability to respond coherently in the early stages of the wildfire crisis. These patterns indicate that the critical distinction is not simply between centralized and decentralized systems, but between more integrated and more fragmented forms of polycentric coordination and their consequences for response network resilience. The study contributes to network governance and disaster policy scholarship by identifying the structural and leadership conditions that distinguish integrated from fragmented polycentric responses, and by situating these findings within cross-state policy variation. For practice, it highlights the risks of loosely integrated polycentric systems and offers guidance on strengthening coordination coherence without sacrificing polycentric flexibility—an increasingly urgent concern as compound and cascading disasters tests existing response architectures.

Authors

• Hanvit Kim, Soongsil University
  Presenting Author

• Naim Kapucu, University of Central Florida
  Non-Presenting Co-Author