Neural-Parameterized Cellular Automata for Wildfire Spread

2026-06-10

AI Summary

Traditional wildfire models struggle to accurately capture complex fire spread due to their reliance on static fuel maps and fixed low-dimensional parameters. This work proposes a novel approach that integrates deep learning with probabilistic cellular automata, introducing a multi-scale convolutional neural network into a three-state cellular automaton for the first time to dynamically generate spatially adaptive fire spread parameters. The method explicitly couples key environmental factors—such as wind direction and slope—while preserving physical interpretability and modeling nonlinear interactions. Implemented in JAX for efficient hardware acceleration and gradient-based parameter calibration, the model achieves over 0.6 IoU in 72-hour forecasts across six major wildfires in the western United States after assimilating 10 days of observational data.

Abstract

Traditional wildfire models rely on rigid, low-dimensional parameters and static fuel maps, frequently underpredicting fire spread. To address this weakness, we introduce a hybrid deep-learning parameterized Probabilistic Cellular Automata (CA) framework implemented in JAX. Our approach employs a Multi-Scale Convolutional Neural Network to dynamically generate spatially varying parameters that govern fire-spread probability, wind alignment, and slope influence. This hybrid design captures complex, nonlinear environmental interactions while preserving the physical interpretability of the underlying three-state CA. The JAX implementation enables hardware acceleration and gradient-based parameter calibration. Evaluated on six large-scale wildfires in the western United States, the model maintains IoU > 0.6 over 72-hour forecast horizons after a 10-day data assimilation window during which the model is fitted incrementally to observed perimeters; the resulting forecast is a conditional projection of fire growth under the suppression regime already ncoded in those observations.

Problem

• wildfire spread
• cellular automata
• parameterization
• fuel maps
• fire modeling

Innovation

• Neural-Parameterized Cellular Automata
• Multi-Scale CNN
• Wildfire Spread Prediction
• Probabilistic Cellular Automata
• JAX-based Differentiable Modeling