PyTorchFire: A GPU-Accelerated Wildfire Simulator with Differentiable Cellular Automata

TL;DR

PyTorchFire tackles the need for fast, accurate wildfire predictions by integrating differentiable cellular automata with GPU-accelerated computation and gradient-based parameter calibration. The approach enables real-time estimation of ignition probabilities and model parameters via gradient descent, using a step-ring backpropagation strategy to manage memory while preserving gradient diversity. Empirical results demonstrate substantial speedups over CPU baselines and improved alignment with both simulated and real wildfire observations, validating the method's potential for real-time decision support. The work provides an open-source, portable tool that can be extended toward global-scale applications and more sophisticated fire models, bridging research and practical wildfire management.

Abstract

Accurate and rapid prediction of wildfire trends is crucial for effective management and mitigation. However, the stochastic nature of fire propagation poses significant challenges in developing reliable simulators. In this paper, we introduce PyTorchFire, an open-access, PyTorch-based software that leverages GPU acceleration. With our redesigned differentiable wildfire Cellular Automata (CA) model, we achieve millisecond-level computational efficiency, significantly outperforming traditional CPU-based wildfire simulators on real-world-scale fires at high resolution. Real-time parameter calibration is made possible through gradient descent on our model, aligning simulations closely with observed wildfire behavior both temporally and spatially, thereby enhancing the realism of the simulations. Our PyTorchFire simulator, combined with real-world environmental data, demonstrates superior generalizability compared to supervised learning surrogate models. Its ability to predict and calibrate wildfire behavior in real-time ensures accuracy, stability, and efficiency. PyTorchFire has the potential to revolutionize wildfire simulation, serving as a powerful tool for wildfire prediction and management.

Figures (14)

• Figure 1: State transition diagram illustrating dynamic changes between wildfire states.
• Figure 2: Comparison of our probability-like normalization function $y=f_p\xspace(x)$ with $y=x$, illustrating its effectiveness in maintaining valid probabilities within $[0, 1]$ and smoothly adjusting values outside this range.
• Figure 3: Propagation of $p_\text{propagate}\xspace$ to neighboring cells in a Moore neighborhood and the application of the inclusion-exclusion principle to determine $p_\text{ignite}\xspace$.
• Figure 4: Flow chart depicting data flow and dependencies in PyTorchFire, emphasizing connections between data, parameters, and outputs.
• Figure 5: Selected rings of cells for back-propagation, illustrating the balance between memory usage and diversity cell coverage.