Probabilistic wildfire risk assessment: A computationally efficient framework for systematic uncertainty propagation – hazard analysis

This paper introduces a Probabilistic Wildfire Risk Assessment (PWRA) framework that advances wildfire risk analysis by formulating risk as a spatial probability density function of loss and enabling systematic uncertainty propagation across the hazard-to-loss chain. Unlike prevailing wildfire risk assessment tools that rely on large ensembles of stochastic fire simulations, the PWRA employs the Generalized Unscented Transform to propagate uncertainty in ignition, weather, and fuel properties using far less simulations required by Monte Carlo-based methods. The PWRA’s modular design ensures consistent integration from hazard component to loss model, remaining flexible to accommodate different uncertainty sources and stakeholder-driven objectives. This study’s primary contribution is the establishment of the PWRA’s analytical and computational foundations, including a mathematically consistent calculation of exceedance-rate and hazard curves for wildfire demand parameters. Applied to the 2018 Camp Fire domain, the PWRA reproduces realistic spatial patterns of burn probability consistent with historical observations and produces exceedance-rate for multiple fire behavior metrics with substantial computational savings compared to the conventional methods. By offering a unified, flexible, computationally efficient, and uncertainty-aware formulation, the PWRA bridges a critical gap between wildfire hazard modeling and fire safety engineering, enabling the development of performance-based metrics for informed wildfire mitigation actions.