Forecasting solar cycle 25 using comprehensive precursor combination and multiple regression technique

Abstract

ABSTRACT In this paper, we propose a new model to predict the complete sunspot cycle based on the comprehensive precursor information (peak, skewness, maximum geomagnetic index aa of the previous cycle, and start value of predicted cycle). The monthly average sunspot original data are processed by Gaussian smoothing and the new model is validated by the observed sunspots of cycle 24. Compared with the traditional 13-month moving average, the Gaussian filter has less missing information and is better to describe the overall trend of the raw data. Through the permutation and combination of multiple parameters in precursor methods of solar cycle forecasting, the multiple regression technique is used to successfully achieve the peak prediction. The regression coefficient (R) of the empirical model established in this paper can reach 0.95. By adding a new parameter to the original HWR function, we provide a complete solar cycle profile showing unimodal structure. It shows that the peak value of cycle 25 will come in March 2024, with a peak of 140.2.