AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Scaling Laws of Global Weather Models

Yuejiang Yu, Langwen Huang, Alexandru Calotoiu, Torsten Hoefler

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Neural scaling laws · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Increase training data by 10xAurora · ERA5 validation lossvalidation loss: up to 3.2x lower lossImproved
Use Aurora instead of other models for data scalingAurora · ERA5 data-scaling exponentbeta: 0.51 vs 0.30-0.46 for othersImproved
Widen network at fixed parameter count (instead of deepening) (ablation)Aurora, GraphCast, Pangu, SFNO · ERA5 validation lossvalidation lossImproved
Increase width to 256 with fewer message-passing steps vs width 128 with more steps (ablation)GraphCast · ERA5 validation loss, 1.6M paramsvalidation loss: width=256 vs width=128Improved
Reduce depth to 1 (ablation)GraphCast, SFNO · ERA5 validation lossvalidation lossNo clear effect
Double dataset size (15.0 TB to 30.0 TB) (ablation)all five models · parameter-scaling lawexponent deltaImproved
Increase dataset from 15.0 TB to 30.0 TB (ablation)Pangu · parameter-scaling exponentdeltaImproved
Use maximal update parameterization (muP) (ablation)AIFS · AIFS largest model (width 1024) validation lossvalidation lossHurt
Allocate extra compute to data rather than model sizeall models · compute-optimal frontiervalidation lossImproved
Use transformer attention (Aurora) instead of GNN message passing (GraphCast)Aurora vs GraphCast · H100 training utilizationGPU utilization (%): 1.03 -> 37.2Improved
Use GraphCast for 2m temperatureGraphCast · ERA5 2T RMSERMSEImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy