![[2602.22962] Scaling Laws of Global Weather Models](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22962] Scaling Laws of Global Weather Models
Summary
This article examines the scaling laws of global weather models, focusing on the relationship between model performance, dataset size, and compute budget, revealing insights into optimizing weather forecasting models.
Why It Matters
Understanding scaling laws in weather models is crucial for improving predictive accuracy and efficiency in weather forecasting, which has significant implications for climate science, disaster preparedness, and resource management. This research highlights the importance of model architecture and training data size in enhancing performance.
Key Takeaways
- Model performance improves significantly with larger training datasets.
- Wider model architectures are preferred over deeper ones for weather forecasting.
- Allocating compute resources to longer training durations yields better performance than increasing model size.
Computer Science > Machine Learning arXiv:2602.22962 (cs) [Submitted on 26 Feb 2026] Title:Scaling Laws of Global Weather Models Authors:Yuejiang Yu, Langwen Huang, Alexandru Calotoiu, Torsten Hoefler View a PDF of the paper titled Scaling Laws of Global Weather Models, by Yuejiang Yu and 3 other authors View PDF HTML (experimental) Abstract:Data-driven models are revolutionizing weather forecasting. To optimize training efficiency and model performance, this paper analyzes empirical scaling laws within this domain. We investigate the relationship between model performance (validation loss) and three key factors: model size ($N$), dataset size ($D$), and compute budget ($C$). Across a range of models, we find that Aurora exhibits the strongest data-scaling behavior: increasing the training dataset by 10x reduces validation loss by up to 3.2x. GraphCast demonstrates the highest parameter efficiency, yet suffers from limited hardware utilization. Our compute-optimal analysis indicates that, under fixed compute budgets, allocating resources to longer training durations yields greater performance gains than increasing model size. Furthermore, we analyze model shape and uncover scaling behaviors that differ fundamentally from those observed in language models: weather forecasting models consistently favor increased width over depth. These findings suggest that future weather models should prioritize wider architectures and larger effective training datasets to maximize predicti...
