10:45 AM - 11:00 AM
[SSS09-07] Effects of subsurface heterogeneities and surface topography on long-period surface wave monitoring
Keywords:long-period surface wave, earthquake, landslides, ground motion simulation, topography, subsurface heterogeneity
Dense broadband seismic networks allow us to detect various phenomena, which radiate seismic waves. Especially because the effects of microseisms are weak in periods of 20-50s, this period band has been widely used in monitoring regular earthquakes, shallow very low frequency earthquakes, volcanic eruption, landslides, stormquakes…etc. (e.g., Tsuruoka et al., 2009; Matsuzawa et al., 2012; Fan et al., 2019; Takemura et al., 2019; Okuwaki et al., 2021). In this study, to achieve precise modeling of physical sources of these phenomena, we investigated the effects of subsurface heterogeneities and irregular surface topography on generation and propagation of long-period (20-50 s) surface waves using the numerical simulations via OpenSWPC (Maeda et al. 2017).
We simulated ground motions of a crustal earthquake that occurred at Awaji Island on 13 April 2013. The source parameters of this earthquake were referred from Takemura et al. (2020). We also simulated ground motions caused by a landslide that occurred at Shizuoka Prefecture. By using the estimated model parameters of Okuwaki et al. (2021), this source could be modeled by body forces on the surface. We used three heterogeneous models: (1) the Japan Integrated Velocity Structure Model (JIVSM; Koketsu et al. 2012) with ETOPO1 (Amante and Eakins, 2009), (2) the JIVSM with a flat surface, and (3) the F-net 1D crustal model (Kubo et al., 2002).
To evaluate quantities differences in simulation results, we calculated variance reductions (VRs) of simulated waveforms between the JIVSM with ETOPO1 and simplified models. In VR calculations, we did not permit any time shift, which adjusts travel time differences between synthetic and target waveforms. In the case of a crustal earthquake, high (> 90%) VRs in the F-net 1D model only appeared at distances shorter than 100 km due to non-time shift VR evaluation. In the JIVSM with a flat surface, VRs at almost all stations were greater than 90 %. This result indicates that the surface topography has limited effects on long-period surface waves from regular earthquakes. Similar characteristics were also found in the simulation results of the landslide source, but we found the directivity of VR decrease, especially in the transverse component. The VR decrease was localized around the Chugoku region. This directivity may be related to the topography around the landslide source. To achieve precise modeling of phenomena on the earth’s surface, Green’s functions, including effects of irregular topography, should be applied.
Acknowledgments: We used NIED Hi-net and F-net data https://doi.org/10.17598/NIED.0003 https://doi.org/10.17598/NIED.0005. Numerical simulations were conducted on the Fujitsu PRIMERGY CX600M1/CX1640M1 (Oakforest-PAC) in the Information Technology Center, University of Tokyo. This study was also supported by the ERI JURP 2020-S-04
We simulated ground motions of a crustal earthquake that occurred at Awaji Island on 13 April 2013. The source parameters of this earthquake were referred from Takemura et al. (2020). We also simulated ground motions caused by a landslide that occurred at Shizuoka Prefecture. By using the estimated model parameters of Okuwaki et al. (2021), this source could be modeled by body forces on the surface. We used three heterogeneous models: (1) the Japan Integrated Velocity Structure Model (JIVSM; Koketsu et al. 2012) with ETOPO1 (Amante and Eakins, 2009), (2) the JIVSM with a flat surface, and (3) the F-net 1D crustal model (Kubo et al., 2002).
To evaluate quantities differences in simulation results, we calculated variance reductions (VRs) of simulated waveforms between the JIVSM with ETOPO1 and simplified models. In VR calculations, we did not permit any time shift, which adjusts travel time differences between synthetic and target waveforms. In the case of a crustal earthquake, high (> 90%) VRs in the F-net 1D model only appeared at distances shorter than 100 km due to non-time shift VR evaluation. In the JIVSM with a flat surface, VRs at almost all stations were greater than 90 %. This result indicates that the surface topography has limited effects on long-period surface waves from regular earthquakes. Similar characteristics were also found in the simulation results of the landslide source, but we found the directivity of VR decrease, especially in the transverse component. The VR decrease was localized around the Chugoku region. This directivity may be related to the topography around the landslide source. To achieve precise modeling of phenomena on the earth’s surface, Green’s functions, including effects of irregular topography, should be applied.
Acknowledgments: We used NIED Hi-net and F-net data https://doi.org/10.17598/NIED.0003 https://doi.org/10.17598/NIED.0005. Numerical simulations were conducted on the Fujitsu PRIMERGY CX600M1/CX1640M1 (Oakforest-PAC) in the Information Technology Center, University of Tokyo. This study was also supported by the ERI JURP 2020-S-04