The Earth Prediction Innovation Center (EPIC) and the NOAA Artificial Intelligence (AI) for Numerical Weather Prediction (AI4NWP) Working Group are proud to announce the public release of the nested-Experimental AI Global and Limited-area Ensemble (nested-EAGLE) Application v1.1.0. This release is an update to the v1.0.0 release from April 22, 2026, which was designed for training and evaluating a global-nested configuration of the European Centre for Medium-Range Weather Forecasts (ECMWF) and partner agencies’ Anemoi machine learning (ML) forecast system. It expands the retrospective forecast system into a global-nested near real-time (NRT) forecast system capable of being fully trained or run from a pretrained checkpoint file. It also incorporates several other updates currently available in the EAGLE repository. The new release includes the following updates:
- Integrated a global configuration into the nested-EAGLE framework
- Integrated NRT forecast capability, for both global and global-nested configurations, into the nested-EAGLE framework
- Introduced a retrospective scorecard to compare various AI- and physics-based forecast systems
- Added a NRT forecast verification page for global-nested configurations
The EAGLE User’s Guide has been updated to reflect these improvements to the Application. Data files required to run the nested-EAGLE Application, including verification datasets, are publicly available in Amazon Web Services (AWS) S3 buckets. These include the Global Forecast System (GFS) analysis dataset, High Resolution Rapid Refresh (HRRR) analysis dataset, and the AI-ready archive of conventional observations. Users seeking assistance can submit questions or feedback through EAGLE’s GitHub Discussions or file an issue through GitHub Issues. For further details, please consult the EPIC Nested EAGLE webpage available at https://epic.noaa.gov/nested-eagle/.
As detailed in the User’s Guide, the nested-EAGLE Application runs on Research and Development High Performance Computing System (RDHPCS) Ursa and the native Microsoft Azure Cloud, both of which provide the GPU resources required for a timely execution of this current iteration of the EAGLE machine learning model training workflow.
Contributors:
The nested-EAGLE Application Release team includes members from NOAA laboratories, centers, cooperative institutes, and community partners — notably NOAA Oceanic and Atmospheric Research’s Earth Prediction Innovation Center (EPIC) within the Weather Program Office, Physical Science Laboratory (PSL), Global Systems Laboratory (GSL), and the Office of Modeling and Development (OMD). It uses software developed by PSL, GSL, and ECMWF with EPIC providing support for software infrastructure development. The code is publicly available, hosted on GitHub. Publicly available data is provided through an AWS S3 bucket established as part of the NOAA Open Data Dissemination (NODD) Program and NCAR Research Data Archive. Future research will expand the developmental and operational capabilities of nested-EAGLE to support both on-prem and cloud-based computing environments, as well as additional model applications.
Acknowledgements:
This release was funded by NOAA programs, including the Water in the West Project, Weather Program Office’s Earth Prediction Innovation Center and Joint Technology Transfer Initiative, NWS Office of Science and Technology Integration, and Software Engineering for Novel Architectures project. The team acknowledges the NOAA RDHPCS program and the Microsoft Azure team for providing dedicated computing resources and pipeline integration support for this release.
