Interpretable Patch-Based Deep Learning for Wildfire Spread Prediction
Interpretable Patch-Based Deep Learning for Wildfire Spread Prediction from Ensemble Simulations
This research provides a detailed comparison of different deep learning architectures for wildfire prediction, which could be useful for anyone working on or interested in the application of AI for disaster management.
A new study compares four deep learning architectures for predicting wildfire spread, using 10,584 simulations from Catalonia.
The research finds that only surface fuel load significantly predicts burn probability, reducing prediction error by 21%.
Convolutional models primarily use distance from the fire front, while Swin-Unet focuses more on fuel and terrain.
The models were tested on a second region, Pedriza, with minimal retraining.
Interpretable Patch-Based Deep Learning for Wildfire Spread Prediction from Ensemble Simulations
Wildfire spread is traditionally predicted using physics-based simulators, which are physically interpretable but whose cost increases with each additional ensemble member. We ask how well deep learning surrogates can reproduce these simulations at a fraction of this cost, training them on 10,584 fire spread simulations at 2m resolution for the Rectoret region in Catalonia, Spain. Four architectures are compared: a patch-based U-Net, a transfer-learned ResNet-50, a physics-informed network constrained by the wind-driven advection equation and a Swin-Unet transformer. Among the terrain and vegetation variables, only surface fuel load predicts burn probability with any strength (r = 0.27) and including it lowers prediction error by 21%. The remaining variables correlate weakly and are highly duplicative. Next, an experiment with saliency, occlusion and rotation demonstrates the models' learning. Convolutional models rely primarily on distance from the current fire front, while Swin-Unet assigns more weight to fuel and terrain, a finding also noted in an unrelated wildfire dataset. When applied without retraining to the second region, Pedriza, all three convolutional models still predict fire spread, losing accuracy by a small but systematic margin.
This report is categorized under AITOP's "Papers" desk, focusing on Europe's AI governance and technology ecosystem. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include deep learning.
Source attribution: curated from arXiv cs.AI (@Marcin Lawenda, Aleksandra Krasicka, David Caballero, Luis Torres, Łukasz Szustak).
