5:15 PM - 6:45 PM
[U15-P25] Source Process of the 2024 Noto Peninsula Earthquake Estimated Using the Back Projection Method
Keywords:the 2024 Noto Peninsula Earthquake, Source process, Back Projection method
The M7.6 earthquake that occurred at 16:10 on Jan. 1, 2024 (the 2024 Noto Peninsula earthquake) has a complex aftershock distribution and aftershocks with various mechanisms, which suggests that the fault plane geometry is complex. The geometry of the fault plane assumed when estimating the rupture process of an earthquake can have a significant effect on the analytical results. In this study, we used the back projection method, which does not require the assumption of a fault plane and can estimate the source of the wave group from the coherency obtained by array analysis, to estimate how the rupture progressed. We used data recorded by 7 stations of K-NET and KiK-net of NIED and 16 stations in Nagano Prefecture. The observed acceleration waveforms were band-pass filtered and resampled at 10 Hz after integrated.
The method assumes the source of the wave packet and the seismic velocity structure to calculate the travel time. The semblance value can be obtained by using the travel time difference between the reference point and the other stations to shift the waveform. If the travel time difference from the correct source is used in the calculation, a large semblance value can be obtained. By calculating the semblance value assuming multiple sources, and then restoring the value to the time at each source using the travel times to the reference site, we can estimate when the target wave was propagated from which source (Honda et al., 2009). We choose NGN007 of K-NET as the reference site and performed the analysis for 70 s, including the S-wave portion. Two types of analyses were performed: one assuming fault planes and the other using a 3-D grid without a fault plane.
For the analysis using fault planes, we introduced three faults referring to the F42 and F43 proposed by MLIT and the NT9 fault by MEXT respectively and they were divided into small faults at 5 km intervals. F43 and N9 are SE-dipping faults, and F42, the northernmost fault, is NW-dipping. We assumed the hypocenter at 137.238E, 37.5N, 12.1 km depth, determined by NIED Hi-net. A one-dimensional velocity structure model for each observation site was constructed by combining J-SHIS and Japan Integrated Velocity Structure Model. In this analysis, band-pass filtered velocity waveforms of 0.03-0.5 Hz were used to calculate semblance in a 10sec window. The time variation of the semblance values plotted on the fault plane indicates that the rupture propagated bilaterally from the hypocenter. In the first 35 seconds based on 16:10:00, the central fault was ruptured and then rupture propagated to the southwest and northeastern fault. The southwestern fault released a large amount of energy 45 to 55 seconds after the rupture starting, while the rupture on the east-northeast side lasted until about 50 seconds.
Next, we performed analysis with 5-km 3-D grid without assuming a fault plane. The depth of the grid from 2.1km to 22.1km. In the horizontal direction, 40 grids were placed in the WSW-ENE direction and 13 in the NNW-SSE direction in a rectangular area that includes all of F42, F43, and NT9. In this analysis, velocity waveforms band-passed in the range of 0.05-2.0 Hz were used. The travel time for each pair of grid and observation site was calculated using the 3-D velocity structure of Matsubara et al. (2022). Other parameters are the same as in the previous analysis. The results show the feature of bilaterally propagating rupture from the hypocenter, that is same as those in the previous analysis. Relatively large semblance values are obtained in the deepest grid to the north-northwest at WSW off shore of Noto Peninsula. On the other hand, at ENE off shore of the Noto Peninsula, relatively large semblance values are obtained in the shallow grid to the north-northwest. Such differences in the distribution of semblance may reflect the difference of the fault geometry.
We thank Nagano prefectural government and Seismic Kanto Research Project for providing observed waveforms.
The method assumes the source of the wave packet and the seismic velocity structure to calculate the travel time. The semblance value can be obtained by using the travel time difference between the reference point and the other stations to shift the waveform. If the travel time difference from the correct source is used in the calculation, a large semblance value can be obtained. By calculating the semblance value assuming multiple sources, and then restoring the value to the time at each source using the travel times to the reference site, we can estimate when the target wave was propagated from which source (Honda et al., 2009). We choose NGN007 of K-NET as the reference site and performed the analysis for 70 s, including the S-wave portion. Two types of analyses were performed: one assuming fault planes and the other using a 3-D grid without a fault plane.
For the analysis using fault planes, we introduced three faults referring to the F42 and F43 proposed by MLIT and the NT9 fault by MEXT respectively and they were divided into small faults at 5 km intervals. F43 and N9 are SE-dipping faults, and F42, the northernmost fault, is NW-dipping. We assumed the hypocenter at 137.238E, 37.5N, 12.1 km depth, determined by NIED Hi-net. A one-dimensional velocity structure model for each observation site was constructed by combining J-SHIS and Japan Integrated Velocity Structure Model. In this analysis, band-pass filtered velocity waveforms of 0.03-0.5 Hz were used to calculate semblance in a 10sec window. The time variation of the semblance values plotted on the fault plane indicates that the rupture propagated bilaterally from the hypocenter. In the first 35 seconds based on 16:10:00, the central fault was ruptured and then rupture propagated to the southwest and northeastern fault. The southwestern fault released a large amount of energy 45 to 55 seconds after the rupture starting, while the rupture on the east-northeast side lasted until about 50 seconds.
Next, we performed analysis with 5-km 3-D grid without assuming a fault plane. The depth of the grid from 2.1km to 22.1km. In the horizontal direction, 40 grids were placed in the WSW-ENE direction and 13 in the NNW-SSE direction in a rectangular area that includes all of F42, F43, and NT9. In this analysis, velocity waveforms band-passed in the range of 0.05-2.0 Hz were used. The travel time for each pair of grid and observation site was calculated using the 3-D velocity structure of Matsubara et al. (2022). Other parameters are the same as in the previous analysis. The results show the feature of bilaterally propagating rupture from the hypocenter, that is same as those in the previous analysis. Relatively large semblance values are obtained in the deepest grid to the north-northwest at WSW off shore of Noto Peninsula. On the other hand, at ENE off shore of the Noto Peninsula, relatively large semblance values are obtained in the shallow grid to the north-northwest. Such differences in the distribution of semblance may reflect the difference of the fault geometry.
We thank Nagano prefectural government and Seismic Kanto Research Project for providing observed waveforms.