Skip to main content

ORIGINAL RESEARCH article

Front. Astron. Space Sci.
Sec. Space Physics
Volume 12 - 2025 | doi: 10.3389/fspas.2025.1541299
This article is part of the Research Topic Solar Radio Bursts and their Applications in Space Weather Forecasting View all 5 articles

Forecasting Long-term Sunspot Numbers Using the LSTM-WGAN Model

Provisionally accepted
Hao Yang Hao Yang 1,2Pingbing Zuo Pingbing Zuo 1,2*Kun Zhang Kun Zhang 1,2Zhenning Shen Zhenning Shen 3Zhengyang Zou Zhengyang Zou 3Xueshang Feng Xueshang Feng 1,2
  • 1 Harbin Institute of Technology, Shenzhen, Shenzhen, Guangdong, China
  • 2 State Key Laboratory of Space Weather, National Space Science Center, Chinese Academy of Sciences (CAS), Beijing, Beijing Municipality, China
  • 3 State Key Laboratory of Lunar and Planetary Science, Macau University of Science and Technology, Macau, China

The final, formatted version of the article will be published soon.

    The sunspot number, an indicator of solar activity, is vital for forecasting variations in solar activity and predicting disturbances of the geomagnetic field, which can impact satellite communications and navigation systems. This study proposes a hybrid model that combines Long Short-Term Memory (LSTM) with the Wasserstein Generative Adversarial Network (WGAN) for sunspot number prediction. The LSTM-WGAN model performs better than the LSTM model in forecasting long-term sunspot numbers using single-step forecasting methods. To further evaluate its effectiveness, we conducted a comparative analysis, comparing predictions of LSTM-WGAN with those provided by the European Space Agency (ESA). This analysis confirmed the accuracy and reliability of LSTM-WGAN model in predicting the sunspot numbers. In particular, our model successfully predicted that the peak of sunspot numbers during the 25th Solar Cycle is slightly higher than that of the 24th Solar Cycle, which aligns with current observations.

    Keywords: Sunspot number, LSTM, WGAN, deep learning, time series forecasting

    Received: 07 Dec 2024; Accepted: 17 Jan 2025.

    Copyright: © 2025 Yang, Zuo, Zhang, Shen, Zou and Feng. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Pingbing Zuo, Harbin Institute of Technology, Shenzhen, Shenzhen, Guangdong, China

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.