FireCastNet: Earth-as-a-Graph for Seasonal Fire Prediction

2025-02-03

AI Summary

Long-term wildfire forecasting under intensifying climate change remains challenging due to complex, multi-scale spatiotemporal dependencies across climate, ecological, and anthropogenic systems. Method: This paper introduces FireCastNet, the first adaptation of the short-term weather forecasting graph neural network GraphCast to six-month global seasonal wildfire probability prediction. It establishes a novel “Earth-as-a-Graph” paradigm, representing the Earth system as a dynamic spatiotemporal graph over a global grid. The architecture integrates a 3D convolutional encoder with graph neural networks to jointly model heterogeneous climate–ecosystem–socioeconomic variables across scales. Contribution/Results: Evaluated on the SeasFire dataset, FireCastNet significantly improves accuracy and robustness in predicting burn areas six months in advance. Ablation studies confirm the critical importance of longer input sequences and larger spatial receptive fields. This work pioneers a graph-based modeling framework for long-horizon wildfire forecasting and provides a foundation for global fire risk early warning systems.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Computer Vision: Remote Sensing / Geospatial AI
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

With climate change expected to exacerbate fire weather conditions, the accurate and timely anticipation of wildfires becomes increasingly crucial for disaster mitigation. In this study, we utilize SeasFire, a comprehensive global wildfire dataset with climate, vegetation, oceanic indices, and human-related variables, to enable seasonal wildfire forecasting with machine learning. For the predictive analysis, we present FireCastNet, a novel architecture which combines a 3D convolutional encoder with GraphCast, originally developed for global short-term weather forecasting using graph neural networks. FireCastNet is trained to capture the context leading to wildfires, at different spatial and temporal scales. Our investigation focuses on assessing the effectiveness of our model in predicting the presence of burned areas at varying forecasting time horizons globally, extending up to six months into the future, and on how different spatial or/and temporal context affects the performance. Our findings demonstrate the potential of deep learning models in seasonal fire forecasting; longer input time-series leads to more robust predictions, while integrating spatial information to capture wildfire spatio-temporal dynamics boosts performance. Finally, our results hint that in order to enhance performance at longer forecasting horizons, a larger receptive field spatially needs to be considered.

Problem

• Wildfire Prediction
• Climate Change
• Seasonal Forecasting

Innovation

• FireCastNet
• 3D Image Processing
• GraphCast Technology