10:45 AM - 11:00 AM
[SSS07-01] Examination of applicability of seismic interferometry in exploration of underground structure under a volcanic body
Keywords:exploration of underground structure, seismic interferometry, microtremor, tomography, Daisen volcano
Reflection seismic survey using a large vibrator is difficult to apply to underground structure exploration of rugged mountain bodies. In addition, it is hard to conduct exploration by blasting because it is practically tough to secure a sufficient number of blasting points. Therefore, we tried to apply the seismic interferometry using microtremors as a part of the study of the underground structure exploration of mountain body. Daisen volcano in Tottori Prefecture, which is known to have had several eruptions, was selected for our analysis.
The data used in this analysis is a continuous velocity records observed by NIED (Hi-net) and Kyoto University, at a total of 16 stations. These stations are located within a distance of approximately 50 km from the summit of Daisen Volcano. The data were collected for one year from January 1, 2019. Missing periods of microtremor recordings were visually searched from the raw data and removed before the analysis.
The analysis method is described. First, processing to reduce large-amplitude events due to natural earthquakes or noise was conducted with the spectral whitening method (Bensen et al., 2007), and the records were bandpass-filtered in the frequency range of 0.05 Hz and 2Hz. Then, data with a length of 409.6 seconds was cut out while moving the time window every 450 seconds, and the cross-correlation function was estimated for each time window. Rayleigh wave and Love wave extracted by interfering the UD component and the transverse component, respectively. In the case of there is no data loss, the maximum number of stacks is 70,080. Cross-correlation function obtained by interfering UD components is shown in Fig. 1. Group velocity between the two observation points was obtained by reading the peak of the nonstationary spectrum which calculated by the multiple filter method for data that folded the cross-correlation function with a delay time of 0. Phase velocity was converted from the group velocity using the phase shift calculated in the previous process. The frequency band to obtain a plausible the group velocity was manually determined. The plot of the obtained the phase velocity with results of Yamada (2018, 2019) is shown in Fig. 2. Finally, 2D tomography was performed using the FMST routine (Rawlinson and Sambridge, 2005) to obtain the spatial distribution of the group velocity and the phase velocity for Rayleigh and Love waves, respectively. The settings are as follows: grid width is 0.05 degrees, the initial value of the model is the average of all ray paths, and iterative calculation was performed 7 times. Fig. 3 shows the spatial distribution of the phase velocity of the Rayleigh wave at 0.4 Hz.
The results are shown below. Most of the cross-correlation functions are asymmetric. In addition, the seasonal change in the ratio of the maximum amplitude for each of the positive and negative delay times of the cross-correlation functions was remarkable in some pair of stations located in the north-south direction. This ratio is highest in January and lowest in June. Therefore, it tends to behave as if there was a wave source at the north of the group of stations (Sea of Japan) in winter. The region where the phase velocity of Rayleigh wave is slow in Fig. 3 is almost coincident with the region where negative gravity anomalies are observed in Fig. 4. Because it reflects a thick sedimentary layer of about 1.5 km (Nishida, 2003), propagation velocity is slow at Yumigahama where negative gravity anomalies are most prominent within the range shown in Fig. 3. In areas where the north-eastern side of the tectonic line from Mt. Koreisan to Mt. Daisen (Ota, 1962) and around of Hiruzen volcanic group, negative gravity anomalies are observed (Komuro et al., 1997), and propagation velocity of surface wave is slow compared to the surrounding area. This tendency was remarkable at low frequencies that are affected by deeper layers. A similar tendency appeared in the group velocity of Rayleigh waves and the phase velocity and the group velocity of Love waves. Since these volcanoes are composed of relatively low-density volcanic rocks, they are considered to form contrast with the surroundings where the base rock (plutonic rock) is exposed as a region where surface wave propagation velocity is slow or negative gravity anomaly is observed.
Acknowledgment: we used continuous velocity records observed by NIED (Hi-net) and Kyoto University.
The data used in this analysis is a continuous velocity records observed by NIED (Hi-net) and Kyoto University, at a total of 16 stations. These stations are located within a distance of approximately 50 km from the summit of Daisen Volcano. The data were collected for one year from January 1, 2019. Missing periods of microtremor recordings were visually searched from the raw data and removed before the analysis.
The analysis method is described. First, processing to reduce large-amplitude events due to natural earthquakes or noise was conducted with the spectral whitening method (Bensen et al., 2007), and the records were bandpass-filtered in the frequency range of 0.05 Hz and 2Hz. Then, data with a length of 409.6 seconds was cut out while moving the time window every 450 seconds, and the cross-correlation function was estimated for each time window. Rayleigh wave and Love wave extracted by interfering the UD component and the transverse component, respectively. In the case of there is no data loss, the maximum number of stacks is 70,080. Cross-correlation function obtained by interfering UD components is shown in Fig. 1. Group velocity between the two observation points was obtained by reading the peak of the nonstationary spectrum which calculated by the multiple filter method for data that folded the cross-correlation function with a delay time of 0. Phase velocity was converted from the group velocity using the phase shift calculated in the previous process. The frequency band to obtain a plausible the group velocity was manually determined. The plot of the obtained the phase velocity with results of Yamada (2018, 2019) is shown in Fig. 2. Finally, 2D tomography was performed using the FMST routine (Rawlinson and Sambridge, 2005) to obtain the spatial distribution of the group velocity and the phase velocity for Rayleigh and Love waves, respectively. The settings are as follows: grid width is 0.05 degrees, the initial value of the model is the average of all ray paths, and iterative calculation was performed 7 times. Fig. 3 shows the spatial distribution of the phase velocity of the Rayleigh wave at 0.4 Hz.
The results are shown below. Most of the cross-correlation functions are asymmetric. In addition, the seasonal change in the ratio of the maximum amplitude for each of the positive and negative delay times of the cross-correlation functions was remarkable in some pair of stations located in the north-south direction. This ratio is highest in January and lowest in June. Therefore, it tends to behave as if there was a wave source at the north of the group of stations (Sea of Japan) in winter. The region where the phase velocity of Rayleigh wave is slow in Fig. 3 is almost coincident with the region where negative gravity anomalies are observed in Fig. 4. Because it reflects a thick sedimentary layer of about 1.5 km (Nishida, 2003), propagation velocity is slow at Yumigahama where negative gravity anomalies are most prominent within the range shown in Fig. 3. In areas where the north-eastern side of the tectonic line from Mt. Koreisan to Mt. Daisen (Ota, 1962) and around of Hiruzen volcanic group, negative gravity anomalies are observed (Komuro et al., 1997), and propagation velocity of surface wave is slow compared to the surrounding area. This tendency was remarkable at low frequencies that are affected by deeper layers. A similar tendency appeared in the group velocity of Rayleigh waves and the phase velocity and the group velocity of Love waves. Since these volcanoes are composed of relatively low-density volcanic rocks, they are considered to form contrast with the surroundings where the base rock (plutonic rock) is exposed as a region where surface wave propagation velocity is slow or negative gravity anomaly is observed.
Acknowledgment: we used continuous velocity records observed by NIED (Hi-net) and Kyoto University.