Authors: Ruizhu Chen (Stanford University), Junwei Zhao (Stanford University)

Reliable knowledge of the Sun’s global magnetic field—including its far side—is essential for accurate coronal and solar wind modeling. Direct and continuous far-side magnetic measurements, however, remain unavailable. In this work, we present a data-driven framework for inferring far-side unsigned magnetic flux maps using helioseismic observations. This framework builds on two prior efforts: (1) a time-distance helioseismic imaging technique incorporating 14 multi-skip acoustic wave schemes, significantly improving active region (AR) detection near the far-side limbs and newly emerged ARs (Zhao et al. 2019); and (2) a machine-learning method that maps far-side helioseismic images into magnetic flux maps, trained via an intermediate step using EUV 304 Å images (Chen et al. 2022).

We use this framework to generate a daily time series of far-side magnetic maps based solely on near-side observations. These products are then applied to two case studies: (1) the May 2024 geomagnetic storm, in which a far-side AR rotated into view and drove a major space weather event, and (2) the 2011 November 3 CME event, for which far-side magnetic context is critical in studying solar energetic particle (SEP) transport and acceleration.

Together, these results demonstrate the usefulness of combining helioseismic far-side imaging and machine learning for continuous global solar magnetic monitoring, and show how these far-side magnetic products can be integrated into both solar storm interpretation and SEP event analysis.