Welcome to EAGLE!

This repository contains various configurations to guide users through a full machine learning pipeline for weather prediction! You will find multiple directories showcasing various model configurations ranging from a "hello world" setup to operational quality models. While various model configurations will differ, they will all follow the same steps of an end-to-end machine learning pipeline. This includes:

1. Proper environment setup: Users are provided with ready-to-use conda environments
2. Prepare training data: Use ufs2arco to create training, validation, and test datasets
3. Train an AI model: use anemoi-core modules to train a graph-based model for weather prediction
4. Generate a forecast: use anemoi-inference to run inference from a model checkpoint
5. Verify model performance: use wxvx to verify forecasts against gridded analysis or observervations
6. View model output: use eagle-tools to visualize model output and scores

For more information about model configurations or the various steps of the pipeline, please see our documentation.

---

Acknowledgments

ufs2arco: Tim Smith (NOAA Physical Sciences Laboratory)

• Github
• Documentation

Anemoi: European Centre for Medium-Range Weather Forecasts

• anemoi-core github
• anemoi-inference github
• Documentation: anemoi-models, anemoi-graphs, anemoi-training, anemoi-inference

wxvx: Paul Madden (NOAA Global Systems Laboratory/Cooperative Institute for Research In Environmental Sciences)

• Github

eagle-tools: Tim Smith (NOAA Physical Sciences Laboratory)

• Github