Solar Corona Live! Predicting the 2024 Eclipse in Real Time
Earlier this year, the team at Predictive Science Inc. (PSI) used a near real-time magnetohydrodynamic (MHD) simulation to predict the appearance of the Sun’s atmosphere—the solar corona—leading up to the April 8, 2024 total solar eclipse. For the first time, they provided a continuously updated prediction based on a data-driven time-evolving MHD model. Measurements of the Sun’s surface magnetic field were used to update the boundary conditions of the simulation as new data became available from observational sources like NASA’s Solar Dynamics Observatory. Computational resources provided by the NASA Advanced Supercomputing Division at NASA’s Ames Research Center were critical to the success of this first ever real-time global thermodynamic MHD model of the corona, requiring the continuous use of over 200 compute nodes for a full three weeks.
Without the resources and flexibility afforded to us by the NASA Advanced Supercomputing facility, this unique data assimilative predictive modeling would simply not have been possible. Thanks to access to their resources and assistance from their staff, we were able to incorporate surface magnetic field data into our high-resolution science-quality model almost as quickly as it was taken, and stay within 10 hours of real time.
Cooper Downs,
Predictive Science Inc.
Quick Facts
- Using a dedicated system reservation, the PSI team were able to run their Magnetohydrodynamic Algorithm outside a Sphere (MAS) MHD code in real time continuously for three weeks, using over 200 Cascade Lake 40-core compute nodes of NASA’s Electra supercomputer.
- This real-time data-assimilative model of the corona allowed the team to explore inherently time-dependent phenomena that cannot be studied with a steady-state model, such as coronal mass ejections, prominences, and streamer evolution.
- The resulting model was used by both scientists and the public to plan observational efforts, and demonstrated that real-time, data-driven modeling of the solar corona is possible. The three weeks of simulation data is being used for multiple research projects by researchers across the heliophysics and astrophysics communities, in addition to informing improvements to PSI’s MHD model.
- The team plans to use the methods and experience of this three-week effort to create a 24/7, continuously running real-time model of the solar corona, similar to terrestrial weather models, for use in space weather and heliophysics applications.
Researchers
- Ronald Caplan, Predictive Science Inc. Please enable Javascript to see email address.
- Cooper Downs, Predictive Science Inc. Please enable Javascript to see email address.
