A Novel Approach for Forecasting Cycle 25

Abstract

A novel approach to forecast the ongoing solar cycle, Cycle 25, is proposed in this article. The new (Version 2.0) of the smoothed monthly sunspot number is well fitted by our four-parameter function, with a mean correlation coefficient of $\overline{r} = 0.984$ for all the past cycles. This function can be simplified into different reduced functions, which are more suitable for making predictions. The free parameters of these reduced functions are either $B$, the key parameter, as it is linked to the amplitude, or both $B$ and $\alpha$ (the parameter linked to the rising time). Three predictions are made. First, relying on the available data (i.e., 25 months) from the sunspot series, we use a two-parameter function to estimate the peak value to be $A = 172 \pm 18$ SSN (sunspot number) around $2024.7 \pm 0.7$ year. Then, we propose a new model as the foundation of the other two forecasts. A new three-parameter function is introduced to fit the $B$-parameters of the previous cycles and forecast the ongoing ones. The predictive power of our two functions is added to make two more predictions about the peak of Cycle 25; one that considers the available data ($A = 147 \pm 27$ SSN around $2024.6 \pm 0.7$ year) and another that does not ($A = 156 \pm 31$ SSN around $2024.3 \pm 0.7$ year). By taking the crossing of the confidence intervals, we estimate Cycle 25 to reach its peak $A = 164 \pm 10$ SSN around $2024.5 \pm 0.7$ year.