5:15 PM - 6:45 PM
[U15-P27] Estimation of source rupture process of the 2024 Noto-hanto earthquake sequence focusing on rupture propagation complexity
Keywords:Source rupture processes, rupture propagation complexity, strong ground motions, the waveform inversion
Introduction
The 2024 Noto-hanto earthquake sequence, the hypocenters of which are the tip of the Noto peninsula, started with an Mj5.9 foreshock (the first event) at 16:10:9.54 on 1 January, 2024, and another mainshock Mj7.6 (the second event) occurred subsequently 5km away from the hypocenter of the Mj5.9 event at 16:10:22.57 just after 13 second (JMA, 2024). The aftershock distribution was spread approximately 60km southwest and 60km northeast from the hypocenter (Tohoku univ., 2024), which is assumed to be the rupture area. We estimated the source rupture process of this earthquake by the waveform inversion analysis using the strong motions records.
The strong motions at near the hypocenter, at ISK001(Ohtani) and ISK002(Shoin), showed clear two wave-packets corresponding the first and second events. Other strong motions at ISKH02(Yanagida) and ISK003(Wajima), where located approximately 25km southwest away from hypocenters demonstrated multiple pulses with several wave-packets and long duration of 30 seconds all together. The source rupture process is assumed to be very complex because the waveforms observed from station to station near the assumed rupture area.
Analysis
The waveform inversion was used by the multiple-time-window linear waveform inversion technique (Hartzell and Heaton, 1983). The rupture speed of the first time window was used to be 2.0km/s based on the results of preliminary analysis.
We assumed a fault model which consisted of three fault segments based on the aftershock distribution by DD method (Tohoku univ., 2024). Both of the central segment (str:52 deg) including the hypocenters of the first and second and the West segment (str:65 deg) near the Wajima are southeast-dipping. On the other hand, the east segment (str:200 deg) locating between the Noto peninsula and the Sado island is northwest-dipping.
The theoretical Green’s functions were calculated by the discrete wavenumber method (Bouchon, 1981) with the reflection and transmission matrix method (Kennett and Kerry, 1979) using a one-dimensional-velocity-structure model from the JIVSM (Koketsu et al., 2012).
We analyzed velocity waveforms bandpass-filtered between 0.05 and 0.2 Hz recorded at 22 strong motion stations of K-NET, KiK-net, F-net and JMA. The time window of observed records for inversion analysis was the 60 sec of time histories beginning from 2 sec prior to the direct S-wave onset of the first event because it is difficult to separate between the first and second events.
Results
First, we assumed the rupture process of this earthquake proceed bilaterally from hypocenter along the entire fault plane, then large slips were estimated on the west and east segments. However, the synthesized waveforms at ADM and 47251 stations located at the Sado island near the east segment were not well reproduced.
Next, we analyzed the rupture process models with time lag (10 to 16 seconds) from rupture start time on the east segment. As a result, the residual between the observed and synthesized waveforms were minimized when the rupture on the east segment was delayed by 13 sec.
Large slips were estimated at shallow and deep parts on the west segment and shallow part of the east segment and less slips on the central segment. The total seismic moment is 4.0×1020Nm(Mw7.6), and the average slip over the entire fault plane is 3.0 m. The largest slip is estimated to be 11.0 m at the shallow part of the west segment. This slip distribution may change depending on the assumed velocity structure model. We should be examined validity of this result in the future.
Overall, the source rupture model in this paper reproduces the synthetic waveforms that match the observe ones. However, there are stations where the synthetics do not match always the observations, for example, the ISKH01 (Suzu) near the hypocenter and WJM(Wajima) west of the Noto peninsula.
In order to reproduce the observed waveforms at more stations, more complex fault planes and rupture processes are needed.
The 2024 Noto-hanto earthquake sequence, the hypocenters of which are the tip of the Noto peninsula, started with an Mj5.9 foreshock (the first event) at 16:10:9.54 on 1 January, 2024, and another mainshock Mj7.6 (the second event) occurred subsequently 5km away from the hypocenter of the Mj5.9 event at 16:10:22.57 just after 13 second (JMA, 2024). The aftershock distribution was spread approximately 60km southwest and 60km northeast from the hypocenter (Tohoku univ., 2024), which is assumed to be the rupture area. We estimated the source rupture process of this earthquake by the waveform inversion analysis using the strong motions records.
The strong motions at near the hypocenter, at ISK001(Ohtani) and ISK002(Shoin), showed clear two wave-packets corresponding the first and second events. Other strong motions at ISKH02(Yanagida) and ISK003(Wajima), where located approximately 25km southwest away from hypocenters demonstrated multiple pulses with several wave-packets and long duration of 30 seconds all together. The source rupture process is assumed to be very complex because the waveforms observed from station to station near the assumed rupture area.
Analysis
The waveform inversion was used by the multiple-time-window linear waveform inversion technique (Hartzell and Heaton, 1983). The rupture speed of the first time window was used to be 2.0km/s based on the results of preliminary analysis.
We assumed a fault model which consisted of three fault segments based on the aftershock distribution by DD method (Tohoku univ., 2024). Both of the central segment (str:52 deg) including the hypocenters of the first and second and the West segment (str:65 deg) near the Wajima are southeast-dipping. On the other hand, the east segment (str:200 deg) locating between the Noto peninsula and the Sado island is northwest-dipping.
The theoretical Green’s functions were calculated by the discrete wavenumber method (Bouchon, 1981) with the reflection and transmission matrix method (Kennett and Kerry, 1979) using a one-dimensional-velocity-structure model from the JIVSM (Koketsu et al., 2012).
We analyzed velocity waveforms bandpass-filtered between 0.05 and 0.2 Hz recorded at 22 strong motion stations of K-NET, KiK-net, F-net and JMA. The time window of observed records for inversion analysis was the 60 sec of time histories beginning from 2 sec prior to the direct S-wave onset of the first event because it is difficult to separate between the first and second events.
Results
First, we assumed the rupture process of this earthquake proceed bilaterally from hypocenter along the entire fault plane, then large slips were estimated on the west and east segments. However, the synthesized waveforms at ADM and 47251 stations located at the Sado island near the east segment were not well reproduced.
Next, we analyzed the rupture process models with time lag (10 to 16 seconds) from rupture start time on the east segment. As a result, the residual between the observed and synthesized waveforms were minimized when the rupture on the east segment was delayed by 13 sec.
Large slips were estimated at shallow and deep parts on the west segment and shallow part of the east segment and less slips on the central segment. The total seismic moment is 4.0×1020Nm(Mw7.6), and the average slip over the entire fault plane is 3.0 m. The largest slip is estimated to be 11.0 m at the shallow part of the west segment. This slip distribution may change depending on the assumed velocity structure model. We should be examined validity of this result in the future.
Overall, the source rupture model in this paper reproduces the synthetic waveforms that match the observe ones. However, there are stations where the synthetics do not match always the observations, for example, the ISKH01 (Suzu) near the hypocenter and WJM(Wajima) west of the Noto peninsula.
In order to reproduce the observed waveforms at more stations, more complex fault planes and rupture processes are needed.