Bridging short- and medium-range weather forecasting with machine learning
physics.ao-ph, cs.LG
Submitted: 2026-08-27
Updated: 2026-08-27
Code: https://github.com/NOAA-PSL/nested-eagle
License: http://creativecommons.org/licenses/by/4.0/
The gist: The National Oceanic and Atmospheric Administration (NOAA) employs independent prediction systems for distinct forecast products.
Terminology
Abstract
The National Oceanic and Atmospheric Administration (NOAA) employs independent prediction systems for distinct forecast products. While some separation is practical, we argue that combining short- and medium-range weather into a single prediction system would provide the public with a useful distillation of global weather and its impacts. To this end, we present Nested-EAGLE (Experimental Artificial intelligence Global and Limited-area Ensemble): a 0.25° global weather model with a 6 km refinement over the Contiguous United States (CONUS). The model achieves significantly lower mean-squared error in near-surface and low-level quantities over CONUS compared to NOAA's Global Forecast System and High-Resolution Rapid Refresh (HRRR), while remaining competitive throughout the rest of the global atmosphere. We show that the skill gains for near-surface fields stem from incorporating high-resolution regional analysis data into training through the nesting process. Forecasts of precipitation amounts are less skillful than those from HRRR, owing to deterministic training. However, we show that Nested-EAGLE provides the most accurate forecasts of storm locations at longer leads, despite blurred extrema. Our results motivate future work to extend the skill gains beyond CONUS and improve precipitation representation.
Sources
- Building Machine Learning Limited Area Models: Kilometer-Scale Weather Forecasting in Realistic Settings
- Skillful joint probabilistic weather forecasting from marginals
- FourCastNet 3: A geometric approach to probabilistic machine-learning weather forecasting at scale
- Climate in a Bottle: Towards a Generative Foundation Model for the Kilometer-Scale Global Atmosphere
- HealDA: Highlighting the importance of initial errors in end-to-end AI weather forecasts
- HourGlass: A probabilistic data-driven temporal downscaler for global and regional weather forecasting
- AIFS -- ECMWF's data-driven forecasting system
- Pointwise is Pointless? A Multimodal Ablation Study for Precipitation Nowcasting with Graph Neural Networks
- High-Resolution Probabilistic Data-Driven Weather Modeling with a Stretched-Grid
- Graph-based Neural Weather Prediction for Limited Area Modeling
- Learning Accurate Storm-Scale Evolution from Observations
- Learning to Advect: A Neural Semi-Lagrangian Architecture for Weather Forecasting
- AIFS-DOP: End-to-End Medium-Range Weather Prediction from Observations Alone with Machine Learning
- XiChen: A global weather observation-to-forecast machine learning system via four-dimensional variational gradient-guided flexible assimilation
- Skillful high-resolution weather forecasting independent of physical models
Related papers
- NORi: An ML-Augmented Ocean Boundary Layer Parameterization
- A Checklist to assess the energy and carbon impacts of ML/AI applications in Earth System Modeling
- On the Predictive Skill of Artificial Intelligence-based Weather Models for Extreme Events using Uncertainty Quantification
- A Mechanism-Coupled Split Window Network for Medium- to High-Resolution Land Surface Temperature Retrieval
- Composable multi-satellite precipitation estimation for evolving observing systems
- Towards accurate extreme event likelihoods from diffusion model climate emulators