11:00 AM - 1:00 PM
[SSS06-P04] Reverse-time seismic reflection imaging of crustal structure in the Kinki region using teleseismic records
Keywords:reverse-time imaging, seismic imaging, crustal structure
With the development of dense seismic stations such as Hi-net, seismic data are being accumulated throughout Japan. These data can be used to estimate the subsurface structures.
Imaging methods using reflected waves based on seismic interferometry (e.g., Wapenaar, 2003, Wapenaar and Fokkema, 2006) can provide structures with higher resolution than seismic tomography and receiver function analysis. However, it is still under development.
Artman (2006) proposed a method for reflection imaging by applying the imaging condition (Claerbout, 1971) to natural seismic records. Shiraishi and Watanabe (2021) applied reverse-time migration (RTM), which is used for active source seismic records, to teleseismic records. In this study, we applied this method to the teleseismic records in the Kinki region. First, we performed numerical simulation using a structural model and then applied this method to the seismic records observed using temporal observations and seismic networks in the Kinki region.
We used a velocity structure model (Nakanishi et al., 2018) with a distance of 400 km and a depth of 60 km around the Kii Peninsula for the numerical simulation. We created teleseismic records using finite difference methods (Virieux, 1986, Levander, 1988) by inputting plane P-waves from the bottom of the model. We applied the imaging method to the records to obtain the reflection images. In comparison with the simulation results of local seismic records by Shiraishi and Watanabe (2021), the result of teleseismic records reduces artifacts in the shallow crust and provides a clearer image of the deep structures. This can be because the teleseismic waves illuminate the structure uniformly over a wide area. Simulations in case of a different alignment and the number of seismic stations were also conducted to intend to use actual observation records. It implied that the use of the observation network alone is not sufficient and that the joint use of temporary observation and the observation networks is necessary to improve the density of observation stations with an average interval of about 5 km.
We applied this method to the crustal structure survey records of the Kinki region conducted from 2005 to 2009 (Shibutani and Ito, 2007). From the continuous waveform data, we extracted 60 seconds wave records that include the P-wave arrival for 841 teleseismic events. After applying seismometer response correction, a band-pass filter (0.2 – 1.5 Hz), and automatic fain control (AGC) with a time window of 2 seconds, acoustic RTM imaging was conducted using the vertical component of the seismic records. The velocity structure by Matsubara et al. (2019) was used for the imaging. The obtained images for each earthquake were stacked, and then a wave-number filter was applied to remove artifacts with a high dip and low-frequency noise and conducted gain adjustment.
In the northern part of the survey line, multiple horizontal discontinuous reflection images were observed. As these correspond to the lower boundary of the upper reflective area, they are interpreted as land Moho and roughly corresponds to the velocity structure of Matsubara et al. (2019) and the Moho from receiver function analysis (Shibutani et al. 2008). Above the Moho, reflection images were observed at depth of 10-25 km, including a nearly horizontal structure that defines the lower limit of the distribution of shallow earthquakes in the crust below the Tamba Mountains. These correspond to the reflectors possibly related to the Arima-Takatsuki Tectonic Line (ATL) and the Median Tectonic Line (MTL) estimated by reflection survey (Ito et al., 2006) and natural earthquakes (Katao, 2007). In the southern part, the oceanic Moho of the subducting PHS was estimated by the receiver function analysis (Shibutani et al., 2008, Shiomi et al., 2008). The imaging results of this study, however, showed segmented reflectors along the plate subduction. They were not recognized as a clear and continuous reflection plane. As a whole, the reflection images were obtained as discontinuous scattered images. The reasons for this may include the effects of inhomogeneities in the crust and the effects of waves from outside the cross-section as well as the effects of the density and arrangement of seismic stations, the effects of velocity structure.
Imaging methods using reflected waves based on seismic interferometry (e.g., Wapenaar, 2003, Wapenaar and Fokkema, 2006) can provide structures with higher resolution than seismic tomography and receiver function analysis. However, it is still under development.
Artman (2006) proposed a method for reflection imaging by applying the imaging condition (Claerbout, 1971) to natural seismic records. Shiraishi and Watanabe (2021) applied reverse-time migration (RTM), which is used for active source seismic records, to teleseismic records. In this study, we applied this method to the teleseismic records in the Kinki region. First, we performed numerical simulation using a structural model and then applied this method to the seismic records observed using temporal observations and seismic networks in the Kinki region.
We used a velocity structure model (Nakanishi et al., 2018) with a distance of 400 km and a depth of 60 km around the Kii Peninsula for the numerical simulation. We created teleseismic records using finite difference methods (Virieux, 1986, Levander, 1988) by inputting plane P-waves from the bottom of the model. We applied the imaging method to the records to obtain the reflection images. In comparison with the simulation results of local seismic records by Shiraishi and Watanabe (2021), the result of teleseismic records reduces artifacts in the shallow crust and provides a clearer image of the deep structures. This can be because the teleseismic waves illuminate the structure uniformly over a wide area. Simulations in case of a different alignment and the number of seismic stations were also conducted to intend to use actual observation records. It implied that the use of the observation network alone is not sufficient and that the joint use of temporary observation and the observation networks is necessary to improve the density of observation stations with an average interval of about 5 km.
We applied this method to the crustal structure survey records of the Kinki region conducted from 2005 to 2009 (Shibutani and Ito, 2007). From the continuous waveform data, we extracted 60 seconds wave records that include the P-wave arrival for 841 teleseismic events. After applying seismometer response correction, a band-pass filter (0.2 – 1.5 Hz), and automatic fain control (AGC) with a time window of 2 seconds, acoustic RTM imaging was conducted using the vertical component of the seismic records. The velocity structure by Matsubara et al. (2019) was used for the imaging. The obtained images for each earthquake were stacked, and then a wave-number filter was applied to remove artifacts with a high dip and low-frequency noise and conducted gain adjustment.
In the northern part of the survey line, multiple horizontal discontinuous reflection images were observed. As these correspond to the lower boundary of the upper reflective area, they are interpreted as land Moho and roughly corresponds to the velocity structure of Matsubara et al. (2019) and the Moho from receiver function analysis (Shibutani et al. 2008). Above the Moho, reflection images were observed at depth of 10-25 km, including a nearly horizontal structure that defines the lower limit of the distribution of shallow earthquakes in the crust below the Tamba Mountains. These correspond to the reflectors possibly related to the Arima-Takatsuki Tectonic Line (ATL) and the Median Tectonic Line (MTL) estimated by reflection survey (Ito et al., 2006) and natural earthquakes (Katao, 2007). In the southern part, the oceanic Moho of the subducting PHS was estimated by the receiver function analysis (Shibutani et al., 2008, Shiomi et al., 2008). The imaging results of this study, however, showed segmented reflectors along the plate subduction. They were not recognized as a clear and continuous reflection plane. As a whole, the reflection images were obtained as discontinuous scattered images. The reasons for this may include the effects of inhomogeneities in the crust and the effects of waves from outside the cross-section as well as the effects of the density and arrangement of seismic stations, the effects of velocity structure.