Shannon Entropy-Based Prediction of Solar Cycle 25

Abstract

A new model is proposed to forecast the peak sunspot activity of the upcoming solar cycle (SC) using Shannon entropy estimates related to the declining phase of the preceding SC. Daily and monthly smoothed international sunspot numbers are used in the present study. The Shannon entropy is the measure of inherent randomness in the SC and is found to vary with the phase of an SC as it progresses. In this model each SC with length Tcy is divided into five equal parts of duration Tcy/ 5. Each part is considered as one phase, and they are sequentially termed P1, P2, P3, P4, and P5. The Shannon entropy estimates for each of these five phases are obtained for the nth SC starting from n=10–23. We find that the Shannon entropy during the ending phase (P5) of the nth SC can be efficiently used to predict the peak smoothed sunspot number of the (n+ 1) th SC, i.e.Smaxn+1. The prediction equation derived in this study has a good correlation coefficient of 0.94. A noticeable decrease in entropy from 4.66 to 3.89 is encountered during P5 of SCs 22 to 23. The entropy value for P5 of the present SC 24 is not available as it has not yet ceased. However, if we assume that the fall in entropy continues for SC 24 at the same rate as that for SC 23, then we predict the peak smoothed sunspot number of 63±11.3 for SC 25. It is suggested that the upcoming SC 25 will be significantly weaker and comparable to the solar activity observed during the Dalton minimum in the past.