2:45 PM - 3:00 PM
[SSS07-10] Spatial distribution of the attenuation parameters around Hakone volcanic area estimated with MLTW method
Keywords:scattering attenuation, intrinsic attenuation, MLTW
Introduction
Attenuation structure in Japan has been studied to understand the detailed seismic wave propagation process. Precise attenuation structure improves the quality of ground motion prediction in future destructive earthquakes. In addition, we are interested in utilizing the attenuation structure for finding the reservoir for natural resource mining. Since geothermal power generation requires to find a reservoir consisting of gathering cracks that locate 3–5 km underground, better resolution of attenuation map will contribute it. In this study, we analyze microearthquakes of relatively smaller source distances than previous studies (20 km < ) to obtain a better spatial resolution of the scattering and intrinsic attenuation structure.
Methods
Multiple Lapse Time Window (MLTW) is a method to separate the scattering and intrinsic attenuation (Hoshiba 1991; Fehler et al. 1992). The seismic envelope after the S arrival is affected by both scattering and intrinsic attenuation, while the later S coda is mainly affected by scattering attenuation. Using this effect, MLTW method separately estimates the two types of attenuation effect with seismic envelopes over multiple time windows. We compare the observed envelope with an approximate analytical solution based on radiative transfer theory (Paasschens, 1997) and obtained the best attenuation factors through grid search. Three different frequency ranges, 2-4 Hz, 4-8 Hz, and 8-12 Hz, are analyzed.
Data
We use 4373 precisely relocated (Yukutake et al. 2017) earthquakes that occurred from April 2015 to November 2015 that moment magnitude smaller than three. Earthquakes were recorded at 15 stations maintained by Hot Springs Research Institute. We choose 17339 velocity waveforms in the east-west component with an SN ratio higher than four. Most hypocenters are located inside the caldera region, and the stations are inside and outside the caldera. Sampling frequency depends on the station, 100 Hz and 200 Hz.
Results and Discussions
We estimate one pair of scattering and intrinsic Q values for the source and the station pair, and mapped them on the path with Q tomography. The results show strong scattering attenuation (Qs = 100 – 200) inside the caldera region and weak attenuation outside the caldera region (Qs = 900 – 1000). Scattering attenuation show frequency dependency; the 2-4 Hz of the attenuation map show stronger attenuation in the entire analysis area, while 4-8 Hz and 8-16 Hz show weaker attenuation in some areas. This frequency dependency of the attenuation factor is consistent with previously reported S coda attenuation characteristics (Sato and Fehler, 1998). The deeper part (2-4 km) showed relatively high attenuation than the shallower part (0-2 km).
The intrinsic attenuation pattern is different from the scattering attenuation pattern. Some areas in the attenuation map show weak attenuation in all frequency ranges. The deeper part (2-4 km) shows high intrinsic attenuation, similar to the scattering attenuation.
The attenuation structure obtained in this study shows consistency with previous structural research with other methods in this region (e.g., Yoshimura et al., 2018). Mapped Q values show several km resolutions. We conclude MLTW with a dense seismic network and events achieve a fine attenuation structure imaging than before, which could help to explore the reservoir in combination with other exploration methods.
Attenuation structure in Japan has been studied to understand the detailed seismic wave propagation process. Precise attenuation structure improves the quality of ground motion prediction in future destructive earthquakes. In addition, we are interested in utilizing the attenuation structure for finding the reservoir for natural resource mining. Since geothermal power generation requires to find a reservoir consisting of gathering cracks that locate 3–5 km underground, better resolution of attenuation map will contribute it. In this study, we analyze microearthquakes of relatively smaller source distances than previous studies (20 km < ) to obtain a better spatial resolution of the scattering and intrinsic attenuation structure.
Methods
Multiple Lapse Time Window (MLTW) is a method to separate the scattering and intrinsic attenuation (Hoshiba 1991; Fehler et al. 1992). The seismic envelope after the S arrival is affected by both scattering and intrinsic attenuation, while the later S coda is mainly affected by scattering attenuation. Using this effect, MLTW method separately estimates the two types of attenuation effect with seismic envelopes over multiple time windows. We compare the observed envelope with an approximate analytical solution based on radiative transfer theory (Paasschens, 1997) and obtained the best attenuation factors through grid search. Three different frequency ranges, 2-4 Hz, 4-8 Hz, and 8-12 Hz, are analyzed.
Data
We use 4373 precisely relocated (Yukutake et al. 2017) earthquakes that occurred from April 2015 to November 2015 that moment magnitude smaller than three. Earthquakes were recorded at 15 stations maintained by Hot Springs Research Institute. We choose 17339 velocity waveforms in the east-west component with an SN ratio higher than four. Most hypocenters are located inside the caldera region, and the stations are inside and outside the caldera. Sampling frequency depends on the station, 100 Hz and 200 Hz.
Results and Discussions
We estimate one pair of scattering and intrinsic Q values for the source and the station pair, and mapped them on the path with Q tomography. The results show strong scattering attenuation (Qs = 100 – 200) inside the caldera region and weak attenuation outside the caldera region (Qs = 900 – 1000). Scattering attenuation show frequency dependency; the 2-4 Hz of the attenuation map show stronger attenuation in the entire analysis area, while 4-8 Hz and 8-16 Hz show weaker attenuation in some areas. This frequency dependency of the attenuation factor is consistent with previously reported S coda attenuation characteristics (Sato and Fehler, 1998). The deeper part (2-4 km) showed relatively high attenuation than the shallower part (0-2 km).
The intrinsic attenuation pattern is different from the scattering attenuation pattern. Some areas in the attenuation map show weak attenuation in all frequency ranges. The deeper part (2-4 km) shows high intrinsic attenuation, similar to the scattering attenuation.
The attenuation structure obtained in this study shows consistency with previous structural research with other methods in this region (e.g., Yoshimura et al., 2018). Mapped Q values show several km resolutions. We conclude MLTW with a dense seismic network and events achieve a fine attenuation structure imaging than before, which could help to explore the reservoir in combination with other exploration methods.