arXiv:2606. 11676v1 Announce Type: cross Abstract: Traditional wildfire models rely on rigid, low-dimensional parameters and static fuel maps, frequently underpredicting fire spread.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.
arXiv:2609.36315v1 Announce Type: cross Abstract: Wildfires are an increasing hazard to ecosystems, air quality, and human systems, creating a growing need for datasets that support systematic develo...
The paper introduces a hybrid CNN‑cellular automaton (CNN‑CA) simulator for planning aerial wildfire suppression, trained on six historical fires. It jointly optimizes binary drop decisions and continuous aircraft trajectories while accounting for aircraft constraints, wind drift, and fuel reduction effects of water and retardant. The authors evaluate the method against random, tactical, greedy, and derivative‑free planners, showing significant reductions in simulated fire extent under a 2020 Bear Fire case study, though noting the results are conditional on the frozen simulator and not operational evidence.
The study evaluates deep learning surrogates for wildfire spread prediction, training four architectures on 10,584 high‑resolution simulations from Catalonia. Results show that only surface fuel load significantly predicts burn probability, and convolutional models mainly use distance to the fire front while a transformer model emphasizes fuel and terrain. When applied to a new region without retraining, the models still perform reasonably, with only a modest accuracy drop.
The paper presents a three‑stage framework that merges a Random Forest (RF) model with a cellular automaton (CA) to model large‑scale wildfire spread. First, an RF trained on the 2021 Canadian fire season predicts daily pixel‑level fire‑occurrence probabilities. Second, optional spread‑rate priors are supplied by quantile gradient boosting models for sensitivity analysis. Third, an RF‑informed CA integrates the RF probability layer with neighbourhood‑driven spread on a 5 km grid, achieving higher spatial overlap than CA‑only baselines in 2023 simulations and demonstrating improved performance on 2022–2024 datasets (AUC 0.725–0.795).
The paper presents modular deep learning augmentations for next‑day wildfire spread prediction, including wind‑ and slope‑conditioned attention biases, physics‑feature retrieval‑augmented output correction, and fire‑conditioned dual‑stream gating. These modules are evaluated on five backbone models using the Next Day Wildfire Spread benchmark, with staged ablations, directional audits, retrieval perturbations, calibration measures, and computational comparisons. The best augmented SwinUNETR model achieves an F1 score of 0.4216 and an AUC‑PR of 0.3673, while a mixed ensemble reaches 0.4292 and 0.3790, demonstrating that predictive performance, operational trustworthiness, and computational practicality can be simultaneously improved.
arXiv:2607. 21080v1 Announce Type: new Abstract: Long-horizon weather forecasting is a fundamental challenge in atmospheric science, for which autoregressive Deep Learning Weather Prediction (DLWP) has emerged as the primary paradigm.