Jaime Carrasco-Barra

Title — Hands-on Workshop on Recent Advances in Cellular Automata–Based Wildfire Simulation with Cell2Fire

Abstract — Spatial wildfire spread is a paradigmatic example of complex spatio-temporal dynamics that can be effectively represented using discrete and cellular automata–based models. This hands-on session presents Cell2Fire (C2F), a spatially explicit wildfire simulation framework grounded in cellular automata concepts, where landscape heterogeneity, fuel properties, topography, and weather conditions are encoded on regular grids, and local transition rules govern fire propagation.

Originally developed in C++ for forest fuel management in Canadian landscapes, Cell2Fire has evolved into a flexible modeling framework supporting multiple fire behavior prediction systems, including the Canadian Forest Fire Behavior Prediction (FBP) System, KITRAL, and BehavePlus. These systems define fire spread transition rules that operate at the cell level, enabling the simulation of fire growth as an emergent phenomenon arising from local interactions.

Beyond fire spread simulation, Cell2Fire integrates spatial decision-making components that combine cellular automata outputs with optimization and multi-criteria analysis. Probabilistic ignitions, coupled with fire growth simulations, are used to generate prioritization metric maps based on user-defined landscape values (e.g., downstream protection value, DPV). These maps serve as inputs for resource-constrained optimization of fuel treatments and firebreak placement using mixed-integer programming, allowing the evaluation of intervention strategies through before-and-after comparisons.

The framework is distributed as an open-source, cross-platform QGIS Processing Toolbox plugin (AGPL-3.0), supporting both graphical user interfaces and script-based workflows in Python or Bash. Parallel computing is employed to efficiently execute large ensembles of stochastic simulations, enabling the analysis of uncertainty and variability inherent in wildfire dynamics. During the session, participants will interactively explore the theoretical foundations of cellular automata–based wildfire modeling while running Cell2Fire on their own computers, linking discrete modeling principles with real-world spatial applications.