Abstract:Real-time estimation of earthquake source parameters combined with ground motion prediction equations for ground motion field prediction is the most commonly used and most time-efficient method in earthquake early warning systems. For medium to large earthquakes, obtaining accurate source parameters (magnitude, fault length, fault strike, and the proportion of unilateral or bilateral rupture) is a crucial factor affecting the effectiveness of this method, due to the influence of finite fault effect. A real-time fault rupture scale estimation method for large earthquakes was introduced, including the rupture direction and rupture length of the main fault, which can be used for real-time ground motion prediction and near-real-time seismic influence field mapping. The optimal line source model was determined by minimizing the error function between the observed ground motion parameters and the theoretical predicted values incorporating fault model parameters. By real-time tracking of fault rupture and correcting the source model, the underestimation of near-fault regions by traditional point source models during large earthquakes could be significantly improved. Furthermore, the introduction of the Akaike Information Criterion could automatically determine when to use the line source model, so as to predict the ground motion field more accurately. As a supplement to traditional early warning methods for large earthquakes, this method not only maintained the timeliness but also improved the prediction accuracy. The method was tested in the 2016 Mw 7.0 Kumamoto earthquake. The results showed that at approximately 13 seconds, the angle of the fault model began to converge and adding the line source model after 18 seconds significantly improved the prediction accuracy. This method can rapidly and accurately estimate the magnitude and fault rupture parameters of the earthquake and demonstrates good stability.