An internally consistent framework for calculating cascading probabilistic earthquake risk and its application to a case study in New Zealand

Quantifying the combined effects of earthquakes and their cascading hazards is essential for realistic risk assessment, yet most existing multi-hazard frameworks either neglect the correlations between cascading perils or represent them using simplified assumptions. Dynamic frameworks that explicitly correlate hazard intensities and their uncertainties across cascading perils provide more consistent and physically plausible impact estimates, offering greater value for resilience planning and risk management. This study introduces a probabilistic risk assessment framework that integrates ground shaking, tsunami inundation, liquefaction, landslides, and their combined impacts into a unified modelling approach. The key methodological innovation is a fully correlated Monte Carlo framework that propagates uncertainty from earthquake rupture through sequential hazard models (ground shaking, tsunami, liquefaction, and landslides) to asset damage, preserving the conditional dependence between primary and secondary hazards at every stage of the simulation. This enables physically consistent realization-by-realization estimates of cascading losses rather than combining independently simulated hazards. The framework was tested in Napier, a city of 65 000 inhabitants situated directly above the Hikurangi Subduction Zone (HSZ), New Zealand's largest earthquake source with an estimated maximum credible moment magnitude ( M _w ) ∼9.1. A 100 000-year stochastic catalogue of ruptures was generated and applied to ∼30 000 residential buildings, with ground shaking, tsunami inundation, liquefaction severity, and landslide runouts explicitly modelled. Results include damage state and damage ratio metrics for individual and combined perils, allowing attribution of losses across hazards. Ground shaking and liquefaction contribute approximately 60 % of the average annual building damage, with lateral spreading accounting for a further ∼35 %, while tsunami contributes less than 3 % on average. However, for very rare events (return periods >10 000 years), tsunami losses become comparable to those from ground shaking and liquefaction, demonstrating the importance of explicitly modelling cascading hazards. The framework provides decision makers, urban planners, and the (re)insurance sector with actionable metrics to guide resilience investments, refine underwriting, and minimize losses from cascading hazard events.