Speaker
Description
Forecasting extreme solar events remains one of the most critical challenges in space weather. The May 2024 superstorm serves as a compelling example, underscoring the urgent need for a comprehensive methodological approach capable of predicting the full sequence of events leading to severe geomagnetic storms impacting Earth. In particular, the extreme G5 geomagnetic storm observed in May 2024 was triggered by a fast and highly energetic Coronal Mass Ejection (CME) that cannibalized multiple preceding CMEs and reached Earth in less than two days. In this talk, we demonstrate the power of physics-driven machine learning in predicting the travel time of this CME, achieving unprecedented accuracy, down to a precision of one minute. We present the predictive performance and robustness of the method, supported by a sensitivity analysis that quantifies how uncertainties in input parameters affect the predicted travel time. Further, we illustrate the broader role of machine/deep learning in characterizing the chain of solar-terrestrial events associated with the storm, from active region classification, through flare forecasting to geomagnetic storm prediction. These results mark a significant advancement in the field, offering unprecedented accuracy in space weather forecasting, and also underscore the importance of integrating physics-based models with machine learning in providing new insights into the underlying physical processes.
| Do you plan to attend in-person or online? | In-person |
|---|
