MagFormer is a cutting-edge project aimed at enhancing our understanding of Coronal Mass Ejections (CMEs) and their interactions with Earth's magnetosphere. This tool leverages advanced Machine Learning techniques, particularly MHD-informed multi-modal networks, to provide accurate geomagnetic forecasting. The foundations of this project were laid with the development of CorKit and StarStream, which streamline the data pipeline for researchers in this domain.
MagFormer relies on a rich dataset sourced from multiple space-based observatories and missions. The primary data sources include:
- LASCO (Large Angle and Spectrometric Coronagraph): Provides imagery of the solar corona, crucial for identifying and tracking CMEs.
- SDO (Solar Dynamics Observatory): Offers high-resolution data on solar activities, such as sunspots and flares, that may trigger CMEs.
- DSCOVR (Deep Space Climate Observatory): Monitors solar wind conditions in real-time, providing essential input for predicting geomagnetic storms.
- ACE (Advanced Composition Explorer): Measures solar wind parameters, magnetic fields, and cosmic rays, contributing to the understanding of CME impact on Earth.
- SOHO (Solar and Heliospheric Observatory): Observes the sun's outer atmosphere, aiding in the early detection of CMEs.
- WIND: Records solar wind interactions with the Earth's magnetosphere, offering valuable insights for modeling geomagnetic responses.
- SWARM: Provides precise measurements of the Earth's magnetic field and its variations, which are crucial for understanding geomagnetic storm dynamics.
- STEREO (Solar TErrestrial RElations Observatory): Offers stereoscopic views of the sun and the solar wind, helping to improve CME forecasting by providing 3D observations of solar events.
This project would not be possible without the invaluable data provided by the following missions and institutions:
- NASA Goddard Space Flight Center
- NOAA's Space Weather Prediction Center
- European Space Agency (ESA)
- US Air Force Research Laboratory (AFRL)
We extend our deepest gratitude to the teams behind these missions for their dedication to advancing space science. Acknowledgements
This project owes its success to several key contributors and organizations:
- Arasci Team: The collaboration initiated at the NASA Space Apps Challenge has been instrumental in scaling this solution for broader applications.
- Universidad Comunera (UCOM), Paraguay: Your unwavering support has enabled us to bring this research to the attention of high-impact journals. We are immensely grateful for the encouragement and resources provided.
This project is licensed under the MIT License - see the LICENSE file for details.
MagFormer is not yet published in a peer-reviewed journal. However, if you find this project useful, please cite it as follows:```
@misc{magformer,
author = {Jorge Enciso},
title = {MagFormer: Geomagnetic Forecasting with MHD informed Multi-Modal Networks},
howpublished = {\url{https://github.com/Jorgedavyd/MagFormer}},
year = {2024}
}
- Email: [email protected]
- GitHub