PROPAGATOR

PROPAGATOR is a stochastic, cellular automata–based wildfire spread simulator designed for operational fire risk assessment and emergency management support. The model represents the landscape as a regular raster grid in which each cell can be in one of three states: unburned, burning, or burned. Fire spread is simulated as a probabilistic contagion process between neighboring cells (Moore neighborhood), where the fire spread probability depends on a nominal vegetation-dependent spread potential modulated by wind, topography (slope and aspect derived from a Digital Elevation Model), and dead fine fuel moisture content. The rate of spread controls the timing of state transitions and is computed through quasi-empirical relationships adapted from fire behavior literature.

PROPAGATOR is stochastic: multiple independent simulations are performed and aggregated to produce spatial probability maps of fire arrival, iso-probability contours, and time-evolving spread patterns. Statistics of fireline intensity and rate of spread spatial distributions are available as well. Required inputs include ignition location(s), a fuel/land cover map, a DEM, wind speed and direction, and fuel moisture; optional layers allow representation of suppression actions. The model is computationally efficient, making it suitable for near-real-time applications, and has been validated against observed wildfire perimeters in Mediterranean case studies. While it simplifies detailed physical fire processes in favor of speed and operational usability, it provides robust probabilistic outputs that explicitly represent uncertainty in wildfire spread forecasting.