1:45 PM - 3:15 PM
[SSS09-P08] Reanalysis of the source process of the 2003 Tokachi-oki earthquake - Considering the plate geometry and 3D velocity structure model –
Keywords:The 2003 Tokachi-oki earthquake, Source process, Plate boundary, 3D velocity structure model
- Outline -
A great earthquake with an Mj 8.0 (Mj: JMA magnitude) occurred at 4:50 a.m. on September 26, 2003, the off the Tokachi district of Hokkaido. The earthquake was an event thrusting along the upper boundary of the Pacific slab. Large thrust earthquakes have occurred around these areas and future occurrence is also assumed. However, due to a huge earthquake with Mj 9.0 of the 2011 Tohoku earthquake, M9-class earthquakes are often considered as the largest class earthquake as a plate boundary earthquake for earthquake disaster prevention. Therefore, M8-class earthquakes are one size smaller than the largest earthquakes now, but they are considered to have a higher probability of occurrence than M9-class earthquakes and those are significant as the earthquakes connecting M9-class and M7-class. So, it is important to know the detailed source processes of M8-class earthquakes.
Because the 2003 Tokachi-oki earthquake was an earthquake that occurred after the construction of nationwide strong motion observation networks such as K-NET and KiK-net, many source process analyses using those data, including our result, have been reported (e.g., Koketsu et al., Honda et al., Yamanaka and Kikuchi, Yagi). The main features of the final slip distribution obtained by these analyses are relatively consistent, and it is thought that a large co-seismic slip exists at the deep eastern part of Cape Erimo on the Pacific plate. However, the analyses at that time did not consider the realistic plate shape and three-dimensional velocity structure, and when these subsurface structures are taken into account, it may be possible to know more detailed and reliable source processes. In this paper, we report the trial of the source process inversion analysis considering those models.
- Analysis -
In this analysis, we used 11 KiK-net observation points in Hokkaido, which are used in Koketsu et al., basically. Moreover, a trial study in which the observation stations in the Aomori Prefecture side are added is also performed, because we can calculate the Green's function considering the plate shape effect with the 3D subsurface structure. Similar to the previous analysis, the velocity waveforms obtained by applying a 0.02-0.3Hz band-pass filter to the acceleration waveforms and integrating it are used for the inversion analysis.
The location of the epicenter of the mainshock is based on the JMA unified hypocenter, and the fault plane is set along the top surface of the Pacific plate of JIVSM (Koketsu et al., 2012). For the 2003 Tokachi-oki earthquake, source inversion analysis considering plate shape was also performed by Kobayashi et al. (2012), however, we apply the technique of dividing the surface into many triangular elements, which was already adopted by Hikima and Koketsu (2005) and Koketsu and Hikima (2005).
The Green's functions used for inversion analysis are obtained by the 3D finite difference method using the 3D velocity structure model by JIVSM. At that time, we try to shorten the calculation time by using the reciprocity theorem. The grid spacing of the finite difference method was set to 250 m in the horizontal direction so that effective calculations could be performed up to about 0.5 Hz.
- Preliminary result -
Prior to source process inversion using the 3D Green's function by the finite difference method, we performed the inversion analysis using the 1D Green's functions, which were calculated assuming horizontally stratified velocity models extracted under each observation point from the JIVSM model. As a tentative result, the location of the maximum slip is found at the east of Cape Erimo, and the final slip distribution does not differ significantly from the result using the strong ground motion waveform by Koketsu et al. (2004). On the other hand, the area with a large slip tends to be concentrated in a narrow part, and the source time functions tend to be somewhat short.
In the presentation, we plan to perform analysis using 3D Green's functions and also to examine the largest aftershock (September 26, 2003, 6:08, Mj7.1).
Reference:
Koketsu, Hikima, et al., EPS, 56, 329-334, 2004.
Hikima and Koketsu, monthly Chikyu, 27, 100-105, 2005.
Koketsu and Hikima, Thinking about a huge trench-type earthquake, Symposium Proceedings, 1-6, 2005.
A great earthquake with an Mj 8.0 (Mj: JMA magnitude) occurred at 4:50 a.m. on September 26, 2003, the off the Tokachi district of Hokkaido. The earthquake was an event thrusting along the upper boundary of the Pacific slab. Large thrust earthquakes have occurred around these areas and future occurrence is also assumed. However, due to a huge earthquake with Mj 9.0 of the 2011 Tohoku earthquake, M9-class earthquakes are often considered as the largest class earthquake as a plate boundary earthquake for earthquake disaster prevention. Therefore, M8-class earthquakes are one size smaller than the largest earthquakes now, but they are considered to have a higher probability of occurrence than M9-class earthquakes and those are significant as the earthquakes connecting M9-class and M7-class. So, it is important to know the detailed source processes of M8-class earthquakes.
Because the 2003 Tokachi-oki earthquake was an earthquake that occurred after the construction of nationwide strong motion observation networks such as K-NET and KiK-net, many source process analyses using those data, including our result, have been reported (e.g., Koketsu et al., Honda et al., Yamanaka and Kikuchi, Yagi). The main features of the final slip distribution obtained by these analyses are relatively consistent, and it is thought that a large co-seismic slip exists at the deep eastern part of Cape Erimo on the Pacific plate. However, the analyses at that time did not consider the realistic plate shape and three-dimensional velocity structure, and when these subsurface structures are taken into account, it may be possible to know more detailed and reliable source processes. In this paper, we report the trial of the source process inversion analysis considering those models.
- Analysis -
In this analysis, we used 11 KiK-net observation points in Hokkaido, which are used in Koketsu et al., basically. Moreover, a trial study in which the observation stations in the Aomori Prefecture side are added is also performed, because we can calculate the Green's function considering the plate shape effect with the 3D subsurface structure. Similar to the previous analysis, the velocity waveforms obtained by applying a 0.02-0.3Hz band-pass filter to the acceleration waveforms and integrating it are used for the inversion analysis.
The location of the epicenter of the mainshock is based on the JMA unified hypocenter, and the fault plane is set along the top surface of the Pacific plate of JIVSM (Koketsu et al., 2012). For the 2003 Tokachi-oki earthquake, source inversion analysis considering plate shape was also performed by Kobayashi et al. (2012), however, we apply the technique of dividing the surface into many triangular elements, which was already adopted by Hikima and Koketsu (2005) and Koketsu and Hikima (2005).
The Green's functions used for inversion analysis are obtained by the 3D finite difference method using the 3D velocity structure model by JIVSM. At that time, we try to shorten the calculation time by using the reciprocity theorem. The grid spacing of the finite difference method was set to 250 m in the horizontal direction so that effective calculations could be performed up to about 0.5 Hz.
- Preliminary result -
Prior to source process inversion using the 3D Green's function by the finite difference method, we performed the inversion analysis using the 1D Green's functions, which were calculated assuming horizontally stratified velocity models extracted under each observation point from the JIVSM model. As a tentative result, the location of the maximum slip is found at the east of Cape Erimo, and the final slip distribution does not differ significantly from the result using the strong ground motion waveform by Koketsu et al. (2004). On the other hand, the area with a large slip tends to be concentrated in a narrow part, and the source time functions tend to be somewhat short.
In the presentation, we plan to perform analysis using 3D Green's functions and also to examine the largest aftershock (September 26, 2003, 6:08, Mj7.1).
Reference:
Koketsu, Hikima, et al., EPS, 56, 329-334, 2004.
Hikima and Koketsu, monthly Chikyu, 27, 100-105, 2005.
Koketsu and Hikima, Thinking about a huge trench-type earthquake, Symposium Proceedings, 1-6, 2005.