Write a comment
A new study led by Prof. Jiansen He's team at Peking University, in collaboration with researchers from the Chinese Academy of Sciences, introduces a novel approach for predicting solar activities. The team combines spatiotemporal decomposition methods with machine learning to forecast sunspot numbers and solar magnetic synoptic maps for Solar Cycle 25.
The global distribution of the solar by Simon Mansfield
Sydney, Australia (SPX) Sep 27, 2024
A new study led by Prof. Jiansen He's team at Peking University, in collaboration with researchers from the Chinese Academy of Sciences, introduces a novel approach for predicting solar activities. The team combines spatiotemporal decomposition methods with machine learning to forecast sunspot numbers and solar magnetic synoptic maps for Solar Cycle 25.
The global distribution of the solar magnetic field plays a crucial role in influencing solar activity, which has significant impacts on human society. Understanding the complexity and topology of the solar magnetic field is essential for predicting solar activity and eruptions. The challenge of predicting the evolution of the solar magnetic field has been a long-standing issue in solar research.
In the study, wavelet analysis was applied to the spherical harmonic coefficients of solar synoptic maps. This analysis uncovered complex disturbances in the photospheric magnetic field, particularly during the solar maximum. Notably, a correlation between the peak in harmonic coefficients and sunspot numbers was identified, suggesting a link to the Sun's meridional circulation.
To predict sunspot numbers for Solar Cycle 25, the researchers employed a long short-term memory (LSTM) neural network model. The model forecasts that the peak of sunspot numbers will occur around June 2024, within an 8-month range, with a predicted peak intensity of 166.9+/-22.6. This indicates that Solar Cycle 25 is expected to be stronger than Solar Cycle 24, but slightly weaker than Solar Cycle 23.
The team also developed an integrated approach to predict future 5th-order magnetic synoptic maps. By decomposing harmonic coefficients using empirical mode decomposition (EMD) and predicting each component series through LSTM, the researchers reconstructed low-order synoptic maps. These predictions align with known polarity laws, and a quantitative analysis confirms a reasonable level of reliability.
While some deviations were observed between predicted maps and actual observations, this study provides an important step forward in the empirical prediction of global solar magnetic fields. The findings offer valuable insights that will benefit future solar observation efforts.
Research Report:Prediction of solar activities: Sunspot numbers and solar magnetic synoptic maps
Related Links
School of Earth and Space Sciences, Peking University
Solar Science News at SpaceDaily