Scaling transformer neural networks for skillful and reliable medium-range weather forecasting
read the original abstract
Weather forecasting is a fundamental problem for anticipating and mitigating the impacts of climate change. Recently, data-driven approaches for weather forecasting based on deep learning have shown great promise, achieving accuracies that are competitive with operational systems. However, those methods often employ complex, customized architectures without sufficient ablation analysis, making it difficult to understand what truly contributes to their success. Here we introduce Stormer, a simple transformer model that achieves state-of-the-art performance on weather forecasting with minimal changes to the standard transformer backbone. We identify the key components of Stormer through careful empirical analyses, including weather-specific embedding, randomized dynamics forecast, and pressure-weighted loss. At the core of Stormer is a randomized forecasting objective that trains the model to forecast the weather dynamics over varying time intervals. During inference, this allows us to produce multiple forecasts for a target lead time and combine them to obtain better forecast accuracy. On WeatherBench 2, Stormer performs competitively at short to medium-range forecasts and outperforms current methods beyond 7 days, while requiring orders-of-magnitude less training data and compute. Additionally, we demonstrate Stormer's favorable scaling properties, showing consistent improvements in forecast accuracy with increases in model size and training tokens. Code and checkpoints are available at https://github.com/tung-nd/stormer.
This paper has not been read by Pith yet.
Forward citations
Cited by 7 Pith papers
-
Beyond Linear Superposition: Discovering Climate Features in AI Weather Models with KAN-SAE
KAN-SAE applies nonlinear per-feature B-spline activations in sparse autoencoders to discover 72% more alive climate features and interpretable patterns such as European heatwaves and Pacific typhoons in deep learning...
-
AdaWeather: Adaptively Mixing Probabilistic Weather Forecasts with Logarithmic Regret
AdaWeather adaptively mixes probabilistic weather forecasts and achieves logarithmic regret relative to the best static mixture of experts in hindsight.
-
U-Cast: A Surprisingly Simple and Efficient Frontier Probabilistic AI Weather Forecaster
A standard U-Net with MAE pre-training followed by short CRPS fine-tuning via Monte Carlo Dropout matches or exceeds GenCast and IFS ENS probabilistic skill at 1.5° resolution while cutting training compute and infere...
-
Otter Weather: Skillful and Computationally Efficient Medium-Range Weather Forecasting
Otter Weather is a spatiotemporal model that outperforms NWP baselines by 9.6% at 24h lead with under 3.5 A100-days training and extends efficiency gains to probabilistic forecasting via CRPS.
-
Hybrid Neural World Models
Hybrid neural world models train one network with horizon conditioning to predict multi-horizon physical states and extract a per-trajectory error map from forward passes alone for hybrid accuracy-speed operation acro...
-
PINN-Cast: Exploring the Role of Continuous-Depth NODE in Transformers and Physics Informed Loss as Soft Physical Constraints in Short-term Weather Forecasting
PINN-Cast combines continuous-depth Neural ODEs inside transformer blocks with a two-branch attention module and physics-informed loss to produce short-term weather forecasts that respect governing physical principles.
-
Accurate, Efficient, and Explainable Deep Learning Approaches for Environmental Science Problems
The work introduces WaLeF/FIDLAr for flood forecasting, CoDiCast for probabilistic weather, and Hypercube-RAG for explainable environmental QA, claiming superior accuracy, efficiency, and interpretability over baselines.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.