Understanding failure propagation across a disrupted port network: the case of the Francis Scott Key Bridge

Ports are nodes in supply chains, and their operation is essential for sustaining trade flows. Due to their interconnected nature, a disruption can cascade across the port network and into the larger logistics system. Thus, it is important to understand how disruptions propagate. In this paper, a time-varying Granger causality (GC) methodology is used to analyze failure impact. The method identifies time-dependent causal relationships between ports and examines how disruption impacts propagate across the port network. It captures failure propagation and recovery dynamics over time. Using Automatic Identification System (AIS) vessel data, the methodology was applied to study the closure of the Port of Baltimore following the collapse of the Francis Scott Key Bridge on 26 March 2024. The findings reveal patterns of failure propagation and recovery, highlighting ports and transitions that stabilize or destabilize the network. The method showed high concavity and failure separation, reflecting its ability to identify a best-fit failure pattern.