Welcome to the overview page for Project EAGLE (Experimental AI Global and Limited-area Ensemble forecast system), which is a joint effort between NOAA Research Laboratories and the Earth Prediction Innovation Center (EPIC) in the Office of Oceanic and Atmospheric Research (OAR), and the National Weather Service (NWS). Project EAGLE provides NOAA and the broad weather enterprise with the ability to rapidly test, develop, and demonstrate in the near-real time Artificial Intelligence (AI) models for global ensemble forecasting. Once fully-developed, Project EAGLE will enable researchers within NOAA and across the weather enterprise to rapidly test their AI models against trusted forecast metrics that are currently used to evaluate the performance of NOAA flagship systems like the Global Forecast System (GFS), the Global Ensemble Forecast System (GEFS), High Resolution Rapid Refresh (HRRR), and Warn on Forecast System (WoFS). We envision that Project EAGLE will allow NOAA to identify the best performing AI innovations rapidly and advance them into near real-time demonstration that can deliver value to the Nation, save lives and property, and enhance the national economy.

Today, Project EAGLE includes two components: 

• Global-EAGLE-Solo is a demonstration environment for “deterministic” models that are initialized from a single GFS initial condition. Global-Eagle-Solo is a complement to the existing NOAA GFS physics-based forecast. In the first version of the Global-Eagle-Solo, NOAA EMC re-trained the GraphCast model using Global Data Assimilation System (GDAS) data as inputs and training targets (Tabas et.al 2025).
• Global-EAGLE-Ensemble is a demonstration environment for ensemble forecast systems initialized with the ensemble of GEFS initial conditions. Global-Eagle-Ensemble is a complement to the existing NOAA GEFS physics-based ensemble forecast system. The weights for the Global-Eagle-Ensemble members are generated by fine tuning the original GraphCast weights from DeepMind(c) with recent NOAA operational GDAS analyses. The resulting weights are effectively trained on the combination of the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis dataset, version5 (ERA5), European Centre for Medium-Range Weather Forecasts High RESolution (ECMWF HRES), and the NOAA’s Global Data Assimilation System (NOAA GDAS) analysis. Multiple checkpoints were saved to form 31 global ensemble members. The Global-Eagle-Ensemble is driven by the initial conditions of the operational GEFSv12.

The Global-EAGLE-Solo and the Global-EAGLE-Ensemble demonstrations are based on Google DeepMind’s GraphCast model (Lam et al., 2023) and are tuned by the NOAA Environmental Modeling Center using NOAA data. The model runs on a 0.25 degree latitude-longitude grid (about 28 km) and 13 pressure levels. The model produces 16-day forecasts 2 times a day at 00Z and 12Z. Major atmospheric and surface fields are available, including vertical velocity, geopotential height, specific humidity, 2-meter temperature, 10-meter wind components, and precipitation. The products are 6-hourly forecasts up to 16 days. The data format for version V1 of the system is Gridded Binary version 2 (GRIB2) but we are considering modernizing the output format to modern cloud native storage formats in later versions of the system.

Project EAGLE output is disseminated through the NOAA Open Data Dissemination Program (NODD). A limited set of visualization and verification graphics are available through Dynamic Ensemble-based Scenarios for IDSS (DESI).