10:45 AM - 12:15 PM
[SCG53-P06] Study on real-time prediction of long-period ground motion
The JMA’s earthquake early warning has targeted short period ground motion predictions based on seismic intensity. But from February 2023, The JMA will also target long-period ground motion predictions. The long-period ground motion prediction method used is only ‘No.1) method’. ‘No,2) method ‘ has not used now. Therefore, in this study, we examined the possibility of real-time prediction of long-period ground motion using a method such as the PLUM method, which is used to predict short-period ground motion.
For earthquakes for which the long-period ground motion class was observed in the past, we conducted case analysis using a method such as the PLUM method and examined it. Since the long-period ground motion class is calculated by the absolute velocity response spectrum, it was examined by analyzing the absolute velocity response time history. The prediction method and procedure are as follows. (1) Calculate the absolute velocity response time history at a certain observation station A. (2) Remove the station correction of station A from the absolute velocity response time history. (3) The time history of station A apply to station B to be predicted. (4) Station correction of station B are added to the time history to the time history of station A. This is an assumption that attenuation is not taken into consideration at the stage of (2)=>(3), and is based on the idea used in the PLUM method. For the correction of stations, the correction value used in the method of predicting long-period ground motion from the hypocenter (ground motion prediction equation) was used. These corrections are used in the ground motion prediction equation adopted by the Japan Meteorological Agency from February 2023. The ground motion prediction equation is derived by Dhakal et al(2015).
The long-period ground motion class which is the predicted value at station B, is compared with the long-period ground motion class observation value at actual station B. Since the assumption does not consider attenuation, the prediction and observation agree when the distance between station A and station B is short, but the prediction tends to be excessive as the distance increases.
*) Station corrections were provided by the Japan Meteorological Agency's Earthquake and Volcano Technology and Research Division.
References)
DHAKAL, Y.P, W.SUZUKI, T.KUNUGI, and S.AOI (2015), Ground Motion Prediction Equations for Absolute Velocity Response Spectra (1-10 S) in Japan for Earthquake Early Warning, Journal of Japan Association for Earthquake Engineering, 15, 91-111