1:45 PM - 3:15 PM
[SSS09-P09] Re-examination of broadband source model of the 2007 Niigata-ken Chuetsu-oki earthquake (MW 6.6)
Keywords:Large slip area, Strong-motion generation area, Source inversion, 3D Green’s functions, Empirical Green’s functions, Source scaling relationship
Strong ground motions with the maximum JMA Intensity Scale of 6+ were observed during the Niigata-ken Chuetsu-oki earthquake (MW 6.6) that occurred on July 16, 2007. The strong-motion waveforms with three pulses were recorded at the Kashiwazaki-Kariwa Nuclear Power Plant, which is located near the source region. Many previous studies using source inversion (e.g., Aoi et al., 2008; Miyake et al., 2010) and empirical Green's function (EGF) method (e.g., Kawabe and Kamae, 2008; Yamamoto and Takenaka, 2009) have investigated the three asperities corresponding to the three pulses. However, Miyakoshi et al. (2008) pointed out that there was still considerable difference in the location and/or scale of the asperities between studies. This fact leads us to re-examine the source characteristic of this earthquake. In this study, using strong-motion waveforms, we perform a source inversion (0.05-0.5 Hz) and a forward simulation with EGF method (0.2-10 Hz) to investigate the broadband source characteristics of this earthquake. Our inversion and simulation assumed an identical fault plane.
First, using the arrival time data picked by JMA, we relocated the hypocenter of the mainshock and aftershocks by the programs of Hirata and Matsu'ura (1987) and Waldhauser and Ellsworth (2000). Based on the relocated hypocenter, we assumed a southeastward-dipping fault plane with a length of 28 km and a width of 22 km. The source inversion was based on the multi-time-window linear waveform-inversion method (Hartzell and Heaton 1983). The advantage of our inversion is taking into account the effects of the thick 3D sedimentary basin in the Niigata area on seismic wave propagation. There are four current 3D velocity structure models (JNES, 2008; Sekiguchi et al., 2009; Fujiwara et al., 2012, Koketsu et al. 2012). Using 3D ground-motion simulations (Graves, 1996; Pitarka, 1999) against the waveforms recorded during aftershocks, we tested the performance of these velocity structure models and then calibrated the thicknesses of the shallow sedimentary layers in the model with the best performance (Koketsu et al. 2012) by trial and error. The Green's functions for source inversion were calculated using the calibrated 3D velocity structure model. The forward simulation using the EGF method was based on Irikura (1986). To consider the appropriate propagation path effects, our simulation used the aftershocks close to strong-motion generation areas (SMGAs) as the element events.
Our inversion estimated the seismic moment to be 7.8 × 1018 Nm (MW 6.53). Compared to the 2nd stage of the three-stage source scaling relationship (Irikura and Miyake, 2011; Miyakoshi et al., 2018), the estimated average slip was comparable, and the effective fault rupture area was slightly larger. Applying the criterion of Somerville et al. (1999) to the slip distribution, we obtained three asperities (large slip areas) having the significant contribution to the observed waveforms, whose total dimension was 17% of the rupture area. The southernmost asperity was located shallower and had a larger average slip than the other two. The southernmost SMGA estimated by the EGF method was also located in the shallow part of the fault. We clearly showed that the locations of the three SMGAs were in good agreement with those of the three asperities obtained by source inversion. The short-period spectral level generated by these SMGAs was 1.6 × 1019 Nm/s2, which was 1.4-1.5 times as large as the scaling relationship of Dan et al. (2001).
Acknowledgements: This study was based on the 2022 research project “the study on the characterized source model for inland crustal earthquakes” by the Secretariat of the Nuclear Regulation Authority, Japan. The strong-motion waveform data were obtained from the Tokyo Electric Power Company and the National Research Institute for Earth Science and Disaster Resilience.
First, using the arrival time data picked by JMA, we relocated the hypocenter of the mainshock and aftershocks by the programs of Hirata and Matsu'ura (1987) and Waldhauser and Ellsworth (2000). Based on the relocated hypocenter, we assumed a southeastward-dipping fault plane with a length of 28 km and a width of 22 km. The source inversion was based on the multi-time-window linear waveform-inversion method (Hartzell and Heaton 1983). The advantage of our inversion is taking into account the effects of the thick 3D sedimentary basin in the Niigata area on seismic wave propagation. There are four current 3D velocity structure models (JNES, 2008; Sekiguchi et al., 2009; Fujiwara et al., 2012, Koketsu et al. 2012). Using 3D ground-motion simulations (Graves, 1996; Pitarka, 1999) against the waveforms recorded during aftershocks, we tested the performance of these velocity structure models and then calibrated the thicknesses of the shallow sedimentary layers in the model with the best performance (Koketsu et al. 2012) by trial and error. The Green's functions for source inversion were calculated using the calibrated 3D velocity structure model. The forward simulation using the EGF method was based on Irikura (1986). To consider the appropriate propagation path effects, our simulation used the aftershocks close to strong-motion generation areas (SMGAs) as the element events.
Our inversion estimated the seismic moment to be 7.8 × 1018 Nm (MW 6.53). Compared to the 2nd stage of the three-stage source scaling relationship (Irikura and Miyake, 2011; Miyakoshi et al., 2018), the estimated average slip was comparable, and the effective fault rupture area was slightly larger. Applying the criterion of Somerville et al. (1999) to the slip distribution, we obtained three asperities (large slip areas) having the significant contribution to the observed waveforms, whose total dimension was 17% of the rupture area. The southernmost asperity was located shallower and had a larger average slip than the other two. The southernmost SMGA estimated by the EGF method was also located in the shallow part of the fault. We clearly showed that the locations of the three SMGAs were in good agreement with those of the three asperities obtained by source inversion. The short-period spectral level generated by these SMGAs was 1.6 × 1019 Nm/s2, which was 1.4-1.5 times as large as the scaling relationship of Dan et al. (2001).
Acknowledgements: This study was based on the 2022 research project “the study on the characterized source model for inland crustal earthquakes” by the Secretariat of the Nuclear Regulation Authority, Japan. The strong-motion waveform data were obtained from the Tokyo Electric Power Company and the National Research Institute for Earth Science and Disaster Resilience.