Can we predict the solar wind before it reaches Earth?
This challenge invites participants to apply AI and physics-informed models to identify, classify, and forecast solar wind structures using the latest data from Parker Solar Probe, Solar Orbiter, and SDO.
The ultimate goal: linking solar disk activity to solar wind structures to improve space weather forecasting and protect our technology-dependent society.
Stage 1 – Classification & Clustering
Extend prior FDL 2021 work on solar wind structure classification (switchbacks, high-speed streams).
Use unsupervised & interpretable ML to explore the complexity of solar wind data.
Leverage in-situ datasets from Parker Solar Probe & Solar Orbiter.
Stage 2 – Linking Drivers & Precursors
Combine identified wind structures with solar disk imagery.
Apply foundation model embeddings from SDO.
Connect precursors on the Sun with target structures in the solar wind.
Integrate with physics-based models (e.g., HUXt, coronal magnetic fields).
Advance understanding of magnetic reconnection and switchbacks.
Improve predictive capability of coronal hole evolution and solar wind preconditioning.
Provide better forecasts of space weather impacts on Earth.
Parker Solar Probe & Solar Orbiter in-situ measurements
SDO solar disk imagery & embeddings
Physics-based solar wind models (HUXt)
Open-source ML frameworks for interpretable AI
AI/ML researchers interested in physics-informed learning
Space scientists exploring solar-terrestrial interactions
Students and professionals in data science, astrophysics, heliophysics
Innovators eager to tackle real-world forecasting challenges
Stage 1 Launch: [TBD]
Stage 1 Submission Deadline: [TBD]
Stage 2 Launch: [TBD]
Final Presentations & Awards: [TBD]
Accuracy of classification (Stage 1)
Ability to link solar precursors with solar wind outcomes (Stage 2)
Use of interpretable ML & physics-informed approaches
Scored by Macro F1-score, domain expert review, and robustness tests
Total Prize Pool: $XX,000
Publication opportunities in AI + Space Science venues
Mentorship from heliophysics & AI experts
Exposure to global space weather research communities
👉 Ready to predict the solar wind and push the frontier of AI in space science?
