Improving the Shock Propagation Model of SPM3 by Incorporating the Solar Energetic Particle Flux and the Principal Component Analysis Method

Abstract Coronal mass ejections (CMEs), as one of the most destructive solar activity phenomena, are the primary driver of extreme space weather events. Accurately predicting the probability and arrival time of the corresponding interplanetary CMEs and their associated shocks is crucial in space weather forecasting. Currently, the primary shock propagation prediction models mainly input the initial kinematic parameters of CMEs near the Sun. The physics-based Shock Propagation Model Version 3 (SPM3) has demonstrated a good forecasting performance for the CME/shock events from 2013 January to 2023 July. To further improve the prediction accuracy, this study incorporates the flux of solar energetic particles (SEPs). A composite index ( I ), constructed using principal component analysis, quantifies the combined influence of SEP and CME’s source longitude. The analysis reveals that SPM3 systematically underestimates the shock’s transit times when the I < 0.2 and V CME < 650 km s −1 . For the events identified by these criteria, an empirical velocity-based bias correction is applied. Notably, the SEP flux data are specifically used to flag events where the correction is applied, rather than entering as a continuous predictor. The revised version SPM3v2 achieves a total prediction success rate of 61%, with mean absolute error and root mean square error reduced to 8.27 and 10.65 hr, outperforming the original SPM3.