Including dynamics in a network-based stochastic multihazard model: A virtual testbed for volcanic ashfall and flood risk assessment

Network models have been previously proposed for spatial cascades of natural hazard events. These have generally not taken time into account, with the cascade of events effectively assumed to occur instantaneously. This study introduces a dynamic, network-based stochastic model developed as a virtual testbed to simulate complex multihazard interactions between multiple temporal processes, often occurring on different time scales. Since state of the art physical models generally involve heavy computation, the use of computationally simple probability distributions to describe the dynamics and interaction of the hazard events enables a larger number of model simulations, promoting greater robustness of model forecasts. The network modelling approach aims to allow the identification of key elements of the system that are most vulnerable, develop risk mitigation strategies, and examine restoration plans. We exemplify our methodology by investigating impacts of volcanic ashfall on river flow dynamics in the Rangitaiki and Tarawera river systems in New Zealand, simulating hydrological processes over a 365-day period with a volcanic eruption. Our results demonstrate how testbeds can be use to explore “what-if” cascading impacts scenarios, by providing a flexible, computationally efficient framework, offering crucial support for Disaster Risk Management (DRM) in volcanic regions. • We outline a conceptual approach to forecast multihazard natural hazard cascades. • Interactions are modelled using a dynamic temporal network including delays. • Hazard inputs and triggering use published phenomenological hazard models. • All model components have estimable parameters, for transferable application. • The case study virtual testbed explores cascading impacts of ashfall on flood risk.