5:15 PM - 6:45 PM
[SVC26-P10] Estimation of Dispersion Curves by the Frequency-Bessel Transform Method at Mt. Ontake, Central Japan
Keywords:Mt. Ontake, crustal structure, seismic interferometry, frequency-Bessel transform method, seismic ambient noise, dispersion curve
1. OBJECTIVE
Mt. Ontake, central Japan, is a stratovolcano and is characterized by hydrothermal activity and swarm earthquakes. It has also experienced several phreatic eruptions in the past; in particular, the eruption on 2014 caused one of the largest volcanic disasters after WWII. Studies of the subsurface structure of Mt. Ontake are limited, although they are useful for understanding its activity. In this study, we analyzed seismic ambient noise from seismometers around Mt. Ontake and estimated the dispersion curves, which are important for examining subsurface structures.
2. METHOD
The seismic ambient noise can be treated as a stochastic wavefield. Therefore, based on seismic interferometry, the cross-correlation function (CCF) of ambient noise recorded at two points gives the Green function assuming that one point is a virtual source, and the other is a station. Hence, by computing the CCFs of ambient noise, we can derive the numerical information of the subsurface structure. In this study, we calculated the CCFs of ambient noise obtained from 43 seismic stations located around Mt. Ontake. The maximum interstation distance was approximately 42 km, and the data period analyzed was three years (01/01/2019–31/12/2021). The CCFs were then applied to the frequency-Bessel transform method (F-J method) to obtain the frequency-Bessel spectrogram (F-J spectrogram) which tends to infinity when the dispersion equation is satisfied. Then, from the singularity distribution of F-J spectrogram, the frequency-velocity distribution of the ambient noise satisfying the dispersion equation, that is, the dispersion curves, were estimated.
3. RESULT
The CCFs are shown in Fig. 1. It indicated that the longer the interstation distance, the longer the travel time between stations. The strong contrast between the noise (yellow) and signal (red and blue) bands indicates a high signal-to-noise ratio. These results demonstrate the success of extracting the CCF of ambient noise around Mt. Ontake.
The F-J spectrograms and their peaks are shown in Fig. 2. The spectrogram is slightly broader than the ideal dispersion curve, the narrower one is necessary to estimate the more accurate dispersion curves. The peak distribution shows the fundamental mode and the first higher mode; however, other modes are unclear. When performing an inverse analysis using dispersion curves, by referring not only to a fundamental mode but also to higher modes, the more accurate velocity structure is estimated. Therefore, the F-J spectrogram that more clearly reflects the higher modes is necessary.
4. DISCUSSION
When applying the F-J method with stations located on a heterogeneous structure, the F-J spectrogram reflects all dispersion curves for each locally stratified homogeneous structure contained within overall structure. Therefore, division of the study area is effective for improving the broader F-J spectrogram. The appearance of unclear higher modes is caused by the approximation of discrete sums in the calculation of the F-J spectrogram. Thus, they can be obtained more clearly by modifying the calculation process of it.
In the future, the 3-D subsurface velocity structure of Mt. Ontake will be estimated by inverse analysis using the modified dispersion curves. In this process, the study area will be divided, and the 1-D velocity structure will be estimated in each sub area. However, the maximum interstation distances in each sub area should be approximately 10–25 km. By this, a velocity model with a maximum depth of approximately 1–5 km below the surface of Mt. Ontake would be estimated.
5. ACKNOWLEDGEMENT
This work was supported by JSPS KAKENHI Grant Number JP19K04016. Seismic stations we used are maintained by Nagoya University, Japan Meteorological Agency, Gifu and Nagano Prefecture, and National Research Institute for Earth Science and Disaster Resilience.
Mt. Ontake, central Japan, is a stratovolcano and is characterized by hydrothermal activity and swarm earthquakes. It has also experienced several phreatic eruptions in the past; in particular, the eruption on 2014 caused one of the largest volcanic disasters after WWII. Studies of the subsurface structure of Mt. Ontake are limited, although they are useful for understanding its activity. In this study, we analyzed seismic ambient noise from seismometers around Mt. Ontake and estimated the dispersion curves, which are important for examining subsurface structures.
2. METHOD
The seismic ambient noise can be treated as a stochastic wavefield. Therefore, based on seismic interferometry, the cross-correlation function (CCF) of ambient noise recorded at two points gives the Green function assuming that one point is a virtual source, and the other is a station. Hence, by computing the CCFs of ambient noise, we can derive the numerical information of the subsurface structure. In this study, we calculated the CCFs of ambient noise obtained from 43 seismic stations located around Mt. Ontake. The maximum interstation distance was approximately 42 km, and the data period analyzed was three years (01/01/2019–31/12/2021). The CCFs were then applied to the frequency-Bessel transform method (F-J method) to obtain the frequency-Bessel spectrogram (F-J spectrogram) which tends to infinity when the dispersion equation is satisfied. Then, from the singularity distribution of F-J spectrogram, the frequency-velocity distribution of the ambient noise satisfying the dispersion equation, that is, the dispersion curves, were estimated.
3. RESULT
The CCFs are shown in Fig. 1. It indicated that the longer the interstation distance, the longer the travel time between stations. The strong contrast between the noise (yellow) and signal (red and blue) bands indicates a high signal-to-noise ratio. These results demonstrate the success of extracting the CCF of ambient noise around Mt. Ontake.
The F-J spectrograms and their peaks are shown in Fig. 2. The spectrogram is slightly broader than the ideal dispersion curve, the narrower one is necessary to estimate the more accurate dispersion curves. The peak distribution shows the fundamental mode and the first higher mode; however, other modes are unclear. When performing an inverse analysis using dispersion curves, by referring not only to a fundamental mode but also to higher modes, the more accurate velocity structure is estimated. Therefore, the F-J spectrogram that more clearly reflects the higher modes is necessary.
4. DISCUSSION
When applying the F-J method with stations located on a heterogeneous structure, the F-J spectrogram reflects all dispersion curves for each locally stratified homogeneous structure contained within overall structure. Therefore, division of the study area is effective for improving the broader F-J spectrogram. The appearance of unclear higher modes is caused by the approximation of discrete sums in the calculation of the F-J spectrogram. Thus, they can be obtained more clearly by modifying the calculation process of it.
In the future, the 3-D subsurface velocity structure of Mt. Ontake will be estimated by inverse analysis using the modified dispersion curves. In this process, the study area will be divided, and the 1-D velocity structure will be estimated in each sub area. However, the maximum interstation distances in each sub area should be approximately 10–25 km. By this, a velocity model with a maximum depth of approximately 1–5 km below the surface of Mt. Ontake would be estimated.
5. ACKNOWLEDGEMENT
This work was supported by JSPS KAKENHI Grant Number JP19K04016. Seismic stations we used are maintained by Nagoya University, Japan Meteorological Agency, Gifu and Nagano Prefecture, and National Research Institute for Earth Science and Disaster Resilience.