Abstract: The impact of natural disasters has been increasing in the recent years. Despite developing international interest in multi-hazard events, few studies quantify the dynamic interaction that characterise these phenomena. It is argued that without taking into account the dynamic complexity of natural catastrophes, impact assessment will underestimate risk, leading to increased vulnerability and distorted emergency management priorities. The work presented herein demonstrates how we can use graphs and networks to assess the complex impact of multi-hazard scenarios. First, the combination of maximal hazard footprints and exposed nodes (e.g. infrastructure) is used to create the hazard network. Iterative simulation of the subnetwork defined by actual hazard magnitudes is then exploited to provide an estimate of the overall compounded impact from a sequence of hazards. In order to illustrate this novel method, the Kaikoura earthquake event that occurred in 2016 is used as a calibrating event to validate the method and further study the cascading events that might threaten other parts of New Zealand. The cascading hazards include numerous landslides events. The results of the impact model on the road system will then be compared with the recorded level of service following the 2016 events. This technique is intended to inform the basis of challenging scenarios for preparing communities and emergency services.