Prediction of Halo Coronal Mass Ejections Using SDO/HMI Vector Magnetic Data Products and a Transformer Model

2025-02-25

Astrophysical Journal

AI Summary

To address the challenge of early warning for halo coronal mass ejections (CMEs), this paper proposes DeepHalo—the first Transformer-based model for halo CME prediction. It leverages SDO/HMI photospheric vector magnetograms, constructing 24-hour temporal sequences of active regions as input to perform end-to-end binary classification of same-day halo CME occurrence. The method innovatively integrates physics-informed feature engineering with data-driven modeling, enabling precise spatiotemporal matching between CME events and vector magnetograms using the SWNO-KI-LASCO joint catalog. Evaluated in an operational space weather forecasting setting, DeepHalo achieves a true skill statistic (TSS) of 0.907—outperforming an LSTM baseline by 10.5%—and delivers high-accuracy 24-hour advance warnings. This significantly enhances early identification capability for high-impact space weather events.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

We present a transformer model, named DeepHalo, to predict the occurrence of halo coronal mass ejections (CMEs). Our model takes as input an active region (AR) and a profile, where the profile contains a time series of data samples in the AR that are collected 24 hr before the beginning of a day, and predicts whether the AR would produce a halo CME during that day. Each data sample contains physical parameters, or features, derived from photospheric vector magnetic field data taken by the Helioseismic and Magnetic Imager on board the Solar Dynamics Observatory. We survey and match CME events in the Space Weather Database Of Notification, Knowledge, Information and the Large Angle and Spectrometric Coronagraph CME Catalog, and we compile a list of CMEs, including halo CMEs and nonhalo CMEs, associated with ARs in the period between 2010 November and 2023 August. We use the information gathered above to build the labels (positive vs. negative) of the data samples and profiles at hand, where the labels are needed for machine learning. Experimental results show that DeepHalo with a true skill statistic (TSS) score of 0.907 outperforms a closely related long short-term memory network with a TSS score of 0.821. To our knowledge, this is the first time that the transformer model has been used for halo CME prediction.

Problem

• Predict halo coronal mass ejections using SDO/HMI data.
• Develop DeepHalo transformer model for CME prediction.
• Evaluate model performance with TSS score comparison.

Innovation

• Transformer model predicts halo CMEs.
• Uses SDO/HMI vector magnetic field data.
• DeepHalo outperforms LSTM in CME prediction.