Systemic Resilience under Compound Hazards: Insights into Multi-Hazard Earthquake–Flood Recovery Dynamics in Antakya

Disaster recovery is increasingly challenged by cascading and compounded hazards that unfold while urban systems are still partially restored. Yet, most resilience assessments focus on single hazards or static system performance, overlooking the intermediate recovery phase and its evolving dynamics over time. The intermediate recovery phase refers to the period when residents begin to resume activities such as commuting, accessing healthcare, and education, while infrastructure systems remain only partially functional and still vulnerable to further disruptions. This paper examines the impact of multi-hazard interactions on systemic resilience during the intermediate recovery window, using transportation accessibility as a proxy for understanding broader urban functioning.

We develop a network-based modelling approach to evaluate the impact of flooding and flood exposure on a transportation network undergoing recovery from earthquake-induced damage. The approach combines graph-theoretical network analysis with spatial flood modelling to evaluate how cascading disruptions impact connectivity and undermine access to critical services. Three complementary performance metrics are employed: (i) network centrality to identify structurally critical corridors, (ii) accessibility to essential amenities such as shelters, hospitals, and markets, and (iii) disruption-adjusted mobility flows that capture functional losses under inundation. The framework is applied to Antakya, Türkiye, following the February 2023 earthquake and subsequent flooding. 

Model results indicate that flooding exacerbates accessibility losses in Antakya’s earthquake-damaged transportation network, where recovery depends on a limited number of structurally critical corridors. Accessibility impacts are unevenly distributed, with temporary shelters and essential services, some of which are located in flood-exposed areas, becoming intermittently inaccessible when key routes are impassable. Our findings reveal that even seemingly localized flood events along these corridors can trigger systemic network effects, rerouting flows onto longer secondary paths, increasing travel distances, and isolating already vulnerable communities during the recovery process. 

Beyond physical damage, observations from the field trip suggest that residents and displaced communities have adapted to this uncertainty by informally sharing real-time routing information through social media and messaging platforms. This emergent bottom-up coordination reflects community adaptive capacity in navigating infrastructure constraints during the intermediate recovery phase, where road accessibility changes frequently due to ongoing reconstruction and intermittent disruptions. 

Overall, the results suggest that recovery trajectories can become more fragile during the intermediate recovery phase, when infrastructure systems are partially restored but remain structurally and operationally weakened by prior damage. In this state, systems have a reduced capacity to absorb any additional hazards, meaning that flooding can generate impacts that exceed those expected under normal system operations. The findings contribute empirical evidence to debates on systemic resilience and highlight the importance of moving beyond single-hazard recovery strategies toward multi-hazard, network-aware assessments.