image: We are now entering the solar maximum of Solar Cycle 25. The cover features a time series of sunspot numbers over the past five solar cycles and their predicted results for the next 12 years, along with the magnetic synoptic maps and their predicted results for each solar maximum and minimum. The background of the cover consists of an image of solar extreme ultraviolet radiation and the magnetic field obtained by the potential field source surface model (from the Helioviewer Project), as well as the auroras on Earth associated with solar activities (from the International Space Station; NASA website).
Credit: ©Science China Press
This study is led by Prof. Jiansen He’s team at Peking University and their cooperators from the Chinese Academy of Sciences, sheds light on the prediction of solar activities. The researchers analyze and discuss the potential laws in the spherical harmonic coefficients of solar synoptic magnetic maps. By combining machine learning, mode decomposition, and harmonic reconstruction methods, they achieve predictions for the sunspot numbers and solar magnetic synoptic maps for Solar Cycle 25.
The global spatiotemporal distribution of the solar magnetic field is a crucial factor in determining solar activity, which is closely connected to human society. The topology and complexity of the solar magnetic field are key to understanding solar eruptions and predicting solar activity levels. The study of the photospheric magnetic field’s evolution has a long history, and predicting solar magnetic activity remains a hot topic in the field. However, understanding the global spatiotemporal distribution of the solar magnetic field and how to predict its evolution remains an unresolved and challenging issue.
They first apply wavelet analysis to the spherical harmonic coefficients of synoptic maps, revealing complex short-period disturbances in the photospheric magnetic field around the solar maximum. Furthermore, the harmonic coefficient almost always reaches its peak simultaneously with sunspot numbers, suggesting a potential link to the Sun’s meridional circulation.
Next, the researchers construct a long short-term memory neural network (LSTM) model to predict sunspot numbers for Solar Cycle 25. According to the model, the peak sunspot number for Solar Cycle 25 is expected to occur around June 2024 within an 8-month window, with a peak intensity of 166.9±22.6. Therefore, Solar Cycle 25 is predicted to be stronger than Solar Cycle 24 but slightly weaker than Solar Cycle 23.
The researchers further apply an integrated method to predict the future 5-order magnetic synoptic maps. Using empirical mode decomposition (EMD), each harmonic coefficient is decomposed into several component series, which are then predicted using LSTM. The predicted low-order synoptic maps are reconstructed finally through spherical harmonics reconstruction. The predicted synoptic maps are validated to be consistent with known polarity laws, and quantitative analysis suggests a certain level of reliability.
Although there are still some deviations between the predicted maps and observations, this study fills a gap in the empirical prediction of the global distribution of solar magnetic fields, and offers valuable insights for the future solar observation programs.
See the article:
Prediction of solar activities: Sunspot numbers and solar magnetic synoptic maps
https://doi.org/10.1007/s11430-023-1354-4
Journal
Science China Earth Sciences