12:00 〜 12:15
[SSS04-12] Kinematic earthquake source modeling with ocean bottom seismometer data and empirical Green’s function approach
★Invited Papers
キーワード：震源過程、動力学的震源モデリング、海底地震計、経験的グリーン関数、地震震源物理、中規模地震
Spatiotemporal slip distribution during an earthquake, the so-called source process, provides knowledge regarding the physical mechanism and background stress of the earthquake. The source process is estimated by solving the inverse problem so that the observed waveforms are reproduced (e.g. Olson and Apsel, 1982; Hartzell and Heaton, 1983). To constrain the rupture propagation direction in the source process, better station coverage with various azimuths is preferable. However, the station coverage is sometimes problematic for offshore earthquakes, as onshore stations only exist in a certain direction from the source fault beneath a seafloor. Recently, the seafloor observation networks including the S-net, DONET (Aoi et al., 2020), and N-net (Aoi et al., 2023) have been constructed in the offshore region of the Pacific Sea among Japan island arc. We can improve the station coverage for the source model estimation of offshore earthquakes by utilizing such ocean bottom seismograms (OBS).
The use of the OBS data is still challenging as we need to assume the appropriate path effect derived from the attenuation structure and the appropriate site effect caused by the sedimentary layer. In the waveform inversion, we usually use waveforms with a dominant frequency. Specifically, the waveforms within a frequency of 0.1–1.0 Hz are often used for the source process estimation of Mw-6 class earthquakes, and we need to use higher-frequency waveforms for earthquakes with Mw < 6. Therefore, a high-resolution velocity structure is required to correctly assume path and site effects. However, such a high-resolution velocity structure has not been estimated well in an offshore region, which makes it difficult to explain the observed waveforms with a theoretical approach, such as the wavenumber integral method (e.g. Bouchon, 1981) or the finite difference method (e.g. Maeda et al., 2017).
Empirical Green’ s function (EGF) approach is an effective way in a case where we cannot assume a velocity structure with enough resolution. This method regards a waveform from a nearby source fault with a similar focal mechanism as a waveform from a location on a source fault of the target earthquake (Hartzell, 1978), and therefore, the path and site effects of the EGF waveform are approximately the same as those of the target waveforms. Shibata and Aso (2025) developed a new method to estimate source processes including slip directions with a single EGF event by correcting the radiation pattern between waveforms of an EGF event and each location on the source fault.
In this study, we exemplified the source model estimation with a better station coverage including the seafloor seismic observatories than the conventional study using the new EGF approach of Shibata and Aso (2025). Specifically, we utilized the accelerogram of the KiK-net (Aoi et al., 2011) and S-net within a frequency of 0.05–0.5 Hz for the 2016 Mw 7.1 Off-Fukushima earthquake. In addition, we used the velocity seismogram of Hi-net and S-net with a frequency of 0.2–2.0 Hz for the 2017 Mw 5.1 Off-Sanriku earthquake. In each analysis, we obtained the source processes with fine quality of the waveform reproductivity both for the onshore and offshore seismograms, indicating the applicability of the OBS data for the waveform inversion. Such an expansion of the applicability of the waveform inversion can lead to a further understanding of the source kinematics in offshore regions.
The use of the OBS data is still challenging as we need to assume the appropriate path effect derived from the attenuation structure and the appropriate site effect caused by the sedimentary layer. In the waveform inversion, we usually use waveforms with a dominant frequency. Specifically, the waveforms within a frequency of 0.1–1.0 Hz are often used for the source process estimation of Mw-6 class earthquakes, and we need to use higher-frequency waveforms for earthquakes with Mw < 6. Therefore, a high-resolution velocity structure is required to correctly assume path and site effects. However, such a high-resolution velocity structure has not been estimated well in an offshore region, which makes it difficult to explain the observed waveforms with a theoretical approach, such as the wavenumber integral method (e.g. Bouchon, 1981) or the finite difference method (e.g. Maeda et al., 2017).
Empirical Green’ s function (EGF) approach is an effective way in a case where we cannot assume a velocity structure with enough resolution. This method regards a waveform from a nearby source fault with a similar focal mechanism as a waveform from a location on a source fault of the target earthquake (Hartzell, 1978), and therefore, the path and site effects of the EGF waveform are approximately the same as those of the target waveforms. Shibata and Aso (2025) developed a new method to estimate source processes including slip directions with a single EGF event by correcting the radiation pattern between waveforms of an EGF event and each location on the source fault.
In this study, we exemplified the source model estimation with a better station coverage including the seafloor seismic observatories than the conventional study using the new EGF approach of Shibata and Aso (2025). Specifically, we utilized the accelerogram of the KiK-net (Aoi et al., 2011) and S-net within a frequency of 0.05–0.5 Hz for the 2016 Mw 7.1 Off-Fukushima earthquake. In addition, we used the velocity seismogram of Hi-net and S-net with a frequency of 0.2–2.0 Hz for the 2017 Mw 5.1 Off-Sanriku earthquake. In each analysis, we obtained the source processes with fine quality of the waveform reproductivity both for the onshore and offshore seismograms, indicating the applicability of the OBS data for the waveform inversion. Such an expansion of the applicability of the waveform inversion can lead to a further understanding of the source kinematics in offshore regions.