A Probabilistic Approach to Wildfire Spread Prediction Using a Denoising Diffusion Surrogate Model

Abstract. We propose a stochastic framework for wildfire spread prediction using deep generative diffusion models with ensemble sampling. In contrast to traditional deterministic approaches that struggle to capture the inherent uncertainty and variability of wildfire dynamics, our method generates probabilistic forecasts by sampling multiple plausible future scenarios conditioned on the same initial state. As a proof-of-concept, the model is trained on synthetic wildfire data generated by a probabilistic cellular automata-based simulator, which integrates realistic environmental features such as canopy cover, vegetation density, and terrain slope, and is grounded in historical fire events including the Chimney and Ferguson fires. To assess predictive performance and uncertainty modelling, we compare two surrogate models with identical network architecture: one trained via conventional supervised regression, and the other using a conditional diffusion framework with ensemble sampling. In the diffusion-based emulator, multiple inference passes are performed for the same input state by resampling the initial latent variable, allowing the model to capture a distribution of possible outcomes. Both models are evaluated on an independent ensemble testing dataset, ensuring robustness and fair comparison under unseen wildfire scenarios. Experimental results show that the diffusion model significantly outperforms its deterministic counterpart across various metrics. At a training size of 900, the diffusion model outperforms the deterministic baseline by a substantial margin. Averaged across the Chimney fire and Ferguson fire datasets, the diffusion model achieves a 67.6 % reduction in mean squared error (MSE), a 5.4 % improvement in structural similarity index (SSIM), and a 69.7 % reduction in Fréchet Inception Distance (FID). These findings demonstrate that diffusion-based ensemble modelling provides a more flexible and effective approach for wildfire forecasting. By capturing the distributional characteristics of future fire states, our framework supports the generation of fire susceptibility maps that offer actionable insights for risk assessment and resource planning in fire-prone environments.