3:30 PM - 5:00 PM
[SCG60-P03] Broadband Strong Ground Motion Simulation of the 2000 Tottori-ken Seibu Earthquake Using the Empirical Green's Function Method.
Keywords:Empirical Green's Function Method, 2000 Tottori-ken Seibu Earthquake, Broadband Strong Ground Motion Simulation, Surface Wave, Long Period Ground Motion
In Japan, the National Research Institute for Earth Science and Disaster Resilience and other organizations have obtained high-precision records that include long period components, but these records are not being fully utilized. The empirical Green's function method (EGF) can predict broadband strong ground motions if strong motion records with sufficient signal to noise ratios are available. However, while the EGF has a good track record for short period seismic motions, mainly entity waves, it does not have a good track record for long period seismic motions, including surface waves. In recent years, attempts have been made to take shallow slip into account for evaluation near the hypocenter fault pole, but the extent to which this affects the evaluation of seismic motion, including surface waves, has not been fully investigated. In this study, we simulate broadband seismic motion including surface waves using EGF for the 2000 Tottori-ken Seibu earthquake (Mj=7.3), for which aftershock records are abundant.
Figure 1 shows the source model of the 2000 Tottori-ken Seibu earthquake using EGF according to Ikeda et al. (2002). In the following, we attempt to use this model to reproduce broadband seismic motions including surface waves. Figure 2 shows the location of the main shock and aftershocks, their source mechanisms, and observation points. The aftershock used was 10/8 (2000/10/8/20:51, Mj=5.0, depth 5 km) in addition to 10/17 (2000/10/17/22:16, Mj=4.2, depth 8 km) used by Ikeda et al. (2002). First, the main shock was reproduced at stations near the epicenter. (Excluding TTRH02, which is strongly affected by ground non-linearity) Figure 3 shows a comparison between the underground observed waveform and the synthetic waveform. (Using a bandpass filter of 0.125~10.0 Hz) The 10/17 aftershock reproduced well the envelope shape of the acceleration waveform and the pulse waveform characteristics of the velocity waveform. On the other hand, the acceleration waveform of the 10/8 aftershock is underestimated, but the phase and amplitude of the velocity waveform are reproduced relatively well.
Next, we will attempt to reproduce broadband seismic motions, especially surface waves, at stations far from the epicenter. Figure 4 shows a comparison between the underground observed waveform and the synthetic waveform at HRSH06, and Figure 5 shows a comparison of Fourier spectra. (Using a bandpass filter of 0.4~10.0 Hz) In the 10/17 aftershock, the short period seismic motion is reproduced well, but the long period seismic motion is not reproduced. On the other hand, in the aftershock of 10/8, the amplitude of the acceleration waveform is underestimated, but the long period seismic motion is reproduced well. The reason for this is thought to be that the surface wave excitation is weak at the slightly deeper 10/17, while it is strongly excited at the relatively shallower 10/8. Therefore, we attempted to reproduce long period seismic motions due to shallow Asperity2 using the wave number integration method (Hisada (1997)). However, the source model was modified based on Iwata and Sekiguchi (2002). Figure 6 compares the observed and calculated waveforms (frequencies of 0.1 to 10 Hz), and shows that the Radial and UD components reproduce long period seismic motions, which are most likely surface waves.
In this study, we attempted to reproduce broadband seismic motions by EGF using the source model by Ikeda et al. (2002). As a result, the aftershock of 10/17 could not reproduce the long period seismic motion mainly composed of surface waves, although the short period seismic motion was reproduced well. On the other hand, the relatively shallow aftershock of 10/8 revealed a high reproducibility of long period ground motions, but a low reproducibility of short period ground motions. Future calculations will include the effects of other stations and the background region of the epicenter model.
Figure 1 shows the source model of the 2000 Tottori-ken Seibu earthquake using EGF according to Ikeda et al. (2002). In the following, we attempt to use this model to reproduce broadband seismic motions including surface waves. Figure 2 shows the location of the main shock and aftershocks, their source mechanisms, and observation points. The aftershock used was 10/8 (2000/10/8/20:51, Mj=5.0, depth 5 km) in addition to 10/17 (2000/10/17/22:16, Mj=4.2, depth 8 km) used by Ikeda et al. (2002). First, the main shock was reproduced at stations near the epicenter. (Excluding TTRH02, which is strongly affected by ground non-linearity) Figure 3 shows a comparison between the underground observed waveform and the synthetic waveform. (Using a bandpass filter of 0.125~10.0 Hz) The 10/17 aftershock reproduced well the envelope shape of the acceleration waveform and the pulse waveform characteristics of the velocity waveform. On the other hand, the acceleration waveform of the 10/8 aftershock is underestimated, but the phase and amplitude of the velocity waveform are reproduced relatively well.
Next, we will attempt to reproduce broadband seismic motions, especially surface waves, at stations far from the epicenter. Figure 4 shows a comparison between the underground observed waveform and the synthetic waveform at HRSH06, and Figure 5 shows a comparison of Fourier spectra. (Using a bandpass filter of 0.4~10.0 Hz) In the 10/17 aftershock, the short period seismic motion is reproduced well, but the long period seismic motion is not reproduced. On the other hand, in the aftershock of 10/8, the amplitude of the acceleration waveform is underestimated, but the long period seismic motion is reproduced well. The reason for this is thought to be that the surface wave excitation is weak at the slightly deeper 10/17, while it is strongly excited at the relatively shallower 10/8. Therefore, we attempted to reproduce long period seismic motions due to shallow Asperity2 using the wave number integration method (Hisada (1997)). However, the source model was modified based on Iwata and Sekiguchi (2002). Figure 6 compares the observed and calculated waveforms (frequencies of 0.1 to 10 Hz), and shows that the Radial and UD components reproduce long period seismic motions, which are most likely surface waves.
In this study, we attempted to reproduce broadband seismic motions by EGF using the source model by Ikeda et al. (2002). As a result, the aftershock of 10/17 could not reproduce the long period seismic motion mainly composed of surface waves, although the short period seismic motion was reproduced well. On the other hand, the relatively shallow aftershock of 10/8 revealed a high reproducibility of long period ground motions, but a low reproducibility of short period ground motions. Future calculations will include the effects of other stations and the background region of the epicenter model.