Derivation of a Bayesian fire spread model using large-scale wildfire observations

Models that predict wildfire rate of spread (ROS) play an important role in decision-making during firefighting operations, including fire crew placement and timing community evacuations. Here, we use a large set remotely sensed observations, explanatory data (focusing on weather), to demonstrate Bayesian probabilistic ROS modelling approach. Our approach has two major advantages: (1) Using actual instead controlled makes models developed well-suited prediction; (2) accounts for the complex nature by explicitly considering uncertainty produce predictions. We show highly informative predictions can be made from simple model containing wind speed, relative humidity soil moisture. provide current operational context our work calculating widely used deterministic Australia.