5:15 PM - 6:30 PM
[SSS09-P06] Scattering properties in Japanese volcanoes as inferred from the peak ratio analysis of teleseismic P waves using the JMA seismometer network data
Keywords:Heterogeneity, Seismic scattering, Volcanoes, JMA seismometers, Hi-net, Surface wave conversion
Stochastic approaches to study the small-scale medium heterogeneities are based on seismic scattered waves invoked by them. One approach is to use the partition of teleseismic P wave energy into the transverse component seismogram. The parameter used to measure the strength of this seismic wave scattering is the peak energy ratio, which is defined as the ratio of the maximum P wave energy in the transverse component seismogram envelope, to that of the sum of the transverse, radial, and vertical components. According to Sato (2006), the peak ratio can be approximated as 1.81ε2h/a where ε, h, and a are fractional fluctuation in seismic velocity, thickness of the heterogeneous medium, and correlation distance respectively.
The quantification of the small-scale velocity heterogeneities through the computation of the peak ratio has been done by Kubanza et al., (2007) on a global scale. Nishimura (2012) estimated the peak ratio values at the Hi-net seismometer network in Japan. There are two significant observations that can be made from these two studies. One is that the peak ratio increases with increasing frequency. The second is that the more active regions, such as those in the proximity of quaternary volcanoes, yield higher peak ratios. The first observation has been explored by Ganefianto et al. (2020 SSJ meeting) through our estimation of the frequency gradient which characterises the power spectral density function of the small-scale heterogeneity. In this study, we would like to further examine the second observation, by calculating the peak ratios at JMA stations in many volcanoes of Japan. Through this approach, we would like to test the conjecture that the peak ratio gets higher at stations closer to a volcanic centre, in line with the increase in geological activity.
Data from 99 teleseismic events, recorded between 2010 and 2019 were considered. These earthquakes have depths of >300 km, magnitudes of roughly 5 to 6.5, and range between 0o to 60o epicentral distance from Japan. Analyses were performed under four separate frequency bands: 0.5-1, 1-2, 2-4, and 4-8 Hz. Source-receiver pairs with the signal-to-noise ratio of <10 were discarded. We computed seismogram envelopes of the remaining pairs. Finally, for individual stations that record >10 earthquakes, we stacked their envelopes together to obtain the mean-squared envelopes.
Our results show that the peak ratio of the JMA stations mostly range from roughly 0.16-0.44, 0.22-0.52, 0.24-0.54, and 0.30-0.60 for the 0.5-1, 1-2, 2-4, and 4-8 Hz band, respectively. The average and median value is increasing with the frequency as well. These ranges are on the high-end scale of what was found for the corresponding Hi-net peak ratios (e.g. Nishimura, 2012). As a result, on average, the peak ratio of the JMA network is larger than the Hi-net network in all frequency bands. This is also true when comparing only the Hi-net stations that are near quaternary volcanoes. We differentiate the JMA stations in our study according to the volcano they monitor. We see that most JMA stations indicate larger peak ratios than what was estimated for the quaternary volcano stations in the previous study. Furthermore, we differentiate the JMA stations into borehole and surface stations. And we observed that the peak ratios at the surface stations are higher than those at the borehole stations. All of these observations are verified to be statistically significant from the t test. In other words, this suggests that it is highly unlikely that the patterns that we see are simply due to chances.
We speculate that there are two interdependent causes for such observations. One is because the JMA stations are positioned much closer to the volcanic centre which exposes them to richer heterogeneous elements created by volcanism. The second is due to the topography where it may convert P wave energies into surface waves, which can also excite the station’s transverse component. This second cause seems to be much more prevalent at the surface stations compared to the borehole stations, owing to the penetration depth of surface waves.
Acknowledgement: We would like to thank the National Research Institute for Earth Science and Disaster Resilience (NIED) and the Japan Meteorological Agency (JMA) for the access to the data of the high sensitivity seismograph network (Hi-net) and the JMA volcanic seismometer network.
The quantification of the small-scale velocity heterogeneities through the computation of the peak ratio has been done by Kubanza et al., (2007) on a global scale. Nishimura (2012) estimated the peak ratio values at the Hi-net seismometer network in Japan. There are two significant observations that can be made from these two studies. One is that the peak ratio increases with increasing frequency. The second is that the more active regions, such as those in the proximity of quaternary volcanoes, yield higher peak ratios. The first observation has been explored by Ganefianto et al. (2020 SSJ meeting) through our estimation of the frequency gradient which characterises the power spectral density function of the small-scale heterogeneity. In this study, we would like to further examine the second observation, by calculating the peak ratios at JMA stations in many volcanoes of Japan. Through this approach, we would like to test the conjecture that the peak ratio gets higher at stations closer to a volcanic centre, in line with the increase in geological activity.
Data from 99 teleseismic events, recorded between 2010 and 2019 were considered. These earthquakes have depths of >300 km, magnitudes of roughly 5 to 6.5, and range between 0o to 60o epicentral distance from Japan. Analyses were performed under four separate frequency bands: 0.5-1, 1-2, 2-4, and 4-8 Hz. Source-receiver pairs with the signal-to-noise ratio of <10 were discarded. We computed seismogram envelopes of the remaining pairs. Finally, for individual stations that record >10 earthquakes, we stacked their envelopes together to obtain the mean-squared envelopes.
Our results show that the peak ratio of the JMA stations mostly range from roughly 0.16-0.44, 0.22-0.52, 0.24-0.54, and 0.30-0.60 for the 0.5-1, 1-2, 2-4, and 4-8 Hz band, respectively. The average and median value is increasing with the frequency as well. These ranges are on the high-end scale of what was found for the corresponding Hi-net peak ratios (e.g. Nishimura, 2012). As a result, on average, the peak ratio of the JMA network is larger than the Hi-net network in all frequency bands. This is also true when comparing only the Hi-net stations that are near quaternary volcanoes. We differentiate the JMA stations in our study according to the volcano they monitor. We see that most JMA stations indicate larger peak ratios than what was estimated for the quaternary volcano stations in the previous study. Furthermore, we differentiate the JMA stations into borehole and surface stations. And we observed that the peak ratios at the surface stations are higher than those at the borehole stations. All of these observations are verified to be statistically significant from the t test. In other words, this suggests that it is highly unlikely that the patterns that we see are simply due to chances.
We speculate that there are two interdependent causes for such observations. One is because the JMA stations are positioned much closer to the volcanic centre which exposes them to richer heterogeneous elements created by volcanism. The second is due to the topography where it may convert P wave energies into surface waves, which can also excite the station’s transverse component. This second cause seems to be much more prevalent at the surface stations compared to the borehole stations, owing to the penetration depth of surface waves.
Acknowledgement: We would like to thank the National Research Institute for Earth Science and Disaster Resilience (NIED) and the Japan Meteorological Agency (JMA) for the access to the data of the high sensitivity seismograph network (Hi-net) and the JMA volcanic seismometer network.