Investigating the importance of critical infrastructures' interdependencies during recovery; lessons from Hurricane Irma in Saint-Martin's island

Abstract

Most research on critical infrastructure (CI) interdependencies focus primarily on their role in propagating system disruptions, neglecting their role during the restoration and recovery processes. The few studies incorporating interdependencies to model restoration, often consider hypothetical situations, cover a maximum of five CIs, or focus on short-term recovery. In an effort to bridge this gap, the presented study investigates the influence of CI interdependencies on the restoration of Saint Martin's CI network-of-networks after Hurricane Irma. Daily restoration data of 15 CIs was collected to build service restoration curves over a period of 28 months (medium-to long-term recovery). Obtained recovery curves revealed that while Irma took a proportionate toll on all Saint-Martin's CI, the latter's recovery took a surprisingly long time. Our investigation found underlying interdependencies that were ignored in the plans, which could have led to a slower system restoration process. Accordingly, a holistic “network-of-networks”-based approach coupled with statistical modeling was followed to rank CI according to their criticality degree and correspondingly simulate a criticality-based systemic restoration plan. Key findings reveal that, in order to optimize CI recovery, the conventional silo-based and decentralized CI-by-CI restoration approaches should be complemented by a well-coordinated restoration strategy that recognizes and considers interdependencies.