Prediction of Solar Cycle 25 using F10.7 cm Radio Flux And Machine Learning Techniques

Abstract

It is well known that solar activity, which changes periodically, referred to commonly as the solar cycle, modulates solar wind variations in interplanetary space and, in turn, space weather activity in near-Earth space. The study of the solar cycle and its prediction is, thus, crucial. The Sun’s magnetic activity changes are responsible for the variations in solar cycle activity that influence the activity in the corona. Here in this study, we use a time series forecasting model of Seasonal Autoregressive Integrated Moving Average (SARIMA) and predict a value of 133±12 for the peak SSN for the upcoming solar cycle 25 that is expected to occur in August 2024. Next, we employ the SARIMA model to the F10.7cm solar radio flux data and predict the F10.7cm flux for solar cycle 25. The sunspot number and the F10.7cm radio flux are strongly related with a correlation coefficient of 98%. Based on the relation, we obtain the SSN for cycle 25 and predict its peak SSN to be 183±11, which is expected in August 2024. Finally, we also predict the F10.7cm flux for solar cycle 25 by employing various machine learning models and use the predicted F10.7 cm flux to obtain a peak SSN of 167±11 which is expected to occur in August 2024. Thus, the predicted SSN obtained using the F10.7 cm flux, SARIMA model, and machine learning models indicates that the solar cycle 25 is expected to occur in the near soon and will be comparatively stronger than the previous solar cycle 24, but it will be similar to the solar cycle 23.