A Novel Bimodal Forecasting Model for Solar Cycle 25

Abstract In this paper, a novel bimodal model to predict complete sunspot cycle based on comprehensive precursor information is proposed. We compare the traditional 13 month moving average with Gaussian filter and find that latter has less missing can better describe overall trend of raw data. Unlike previous models usually only use one precursor, here we combine implicit geometric solar (peak skewness start value predicted cycle) method geomagnetic index adopt multivariate linear approach higher goodness fit (>0.85) in fitting. Verifications for cycles 22–24 demonstrate good performance predicting peak occurrence time. It also successfully predicts structure 22 24, showing certain ability whether next unimodal or bimodal. shows 25 single-peak will come 2024 October 145.3.