5:15 PM - 7:15 PM
[SSS09-P08] Stability of estimation of the S-wave envelopes broadening in the backarc side of northeast Japan
Keywords:scattering, inhomogeneity, envelope broadening, backarc side, northeast Japan
High-frequency near-field seismic waves does not always consist of impulsive direct waves. They may exhibit spindle-shaped seismic waves with collapsed pulses. This is mainly due to seismic wave scattering caused by the short-wavelength inhomogeneous structure of the Earth’s interior. Recent studies have shown that envelope broadening in the S-wave region (hereafter referred to as S-wave envelope broadening) characterizes short-wavelength inhomogeneous structures. Previous analyses of S-wave envelope broadening in northeastern Japan have focused on earthquakes that occurred in mainly the subduction zone of the Pacific Plate, but such a spindle-shaped waveforms are also observed for earthquakes occurring in the shallow continental crust beneath the Sea of Japan in northeastern Japan. Kusumi et al. (2024) investigated the regional characteristics of S-wave envelope broadening using the time lag (peak delay time) between the arrival time of S-waves and the peak time in the S-wave region of the envelope for such earthquakes occurring on beneath the Sea of Japan in northeastern Japan. However, we recently found that the peak delay time estimation is subject to large uncertainties associated with the choice of the phase-picking method of the travel time of S wave. This study re-analyzed the envelope by introducing a different index based on the time integration of the envelope for better stability of the estimation.
We used the earthquake catalog of the Japan Meteorological Agency and velocity wave data from the Hi-net seismic network of the National Research Institute for Earth Science and Disaster Prevention. We analyzed seismograms of magnitudes between 3.0 and 5.5 that occurred in the shallow region beneath the Sea of Japan in northeastern Japan at hypocentral distances of 100–250 km. Then, we processed seismograms to obtain smoothed envelopes for four frequency bands: 2–4, 4–8, 8–16, and 16–32 Hz. After integrating the envelope with respect to time, we measured the time difference between 5% and 95% of the maximum of the integrated trace, which represents 90% of the signal duration. Following previous studies, we adopted this duration time as an indicator of S-wave envelope broadening. After the measurement of the duration time for many event and station pairs, we calculated the regression line against the hypocentral distance and obtained its logarithmic deviation. Finally, we divided the region of interest into a 10 km square grid and assigned the logarithmic deviation to each wave path. We investigated the regional characteristics of S-wave envelope broadening by calculating the first quartile within each grid cell and reflecting it in a two-dimensional spatial distribution.
The analysis confirmed that the 90% signal duration depends on the hypocentral distance. The variation in the logarithmic deviation from the regression line was evaluated using the standard deviation, which was significantly smaller than that obtained when applying the peak delay time across all frequency bands. This suggests that using the 90% signal duration is becoming a more stable estimation method.
The values in the two-dimensional spatial distribution of logarithmic deviations were relatively large near the volcanoes distributed in the Ou Mountains and on its forearc side, with particularly strong deviations on the high-frequency side. This trend was also confirmed by the peak delay time analysis. In these areas, short-wavelength inhomogeneous structures that cause seismic wave scattering are presumed to be present. On the other hand, relatively small values were observed off the Sea of Japan of Akita and Yamagata prefectures, as well as in northern Aomori prefecture. In particular, the former region is different from previous trends. Several earthquakes that occurred in this region exhibited fluctuations in the detection of direct wave travel time. The introduction of a new indicator may have contributed to resolving this uncertainty.
We used the earthquake catalog of the Japan Meteorological Agency and velocity wave data from the Hi-net seismic network of the National Research Institute for Earth Science and Disaster Prevention. We analyzed seismograms of magnitudes between 3.0 and 5.5 that occurred in the shallow region beneath the Sea of Japan in northeastern Japan at hypocentral distances of 100–250 km. Then, we processed seismograms to obtain smoothed envelopes for four frequency bands: 2–4, 4–8, 8–16, and 16–32 Hz. After integrating the envelope with respect to time, we measured the time difference between 5% and 95% of the maximum of the integrated trace, which represents 90% of the signal duration. Following previous studies, we adopted this duration time as an indicator of S-wave envelope broadening. After the measurement of the duration time for many event and station pairs, we calculated the regression line against the hypocentral distance and obtained its logarithmic deviation. Finally, we divided the region of interest into a 10 km square grid and assigned the logarithmic deviation to each wave path. We investigated the regional characteristics of S-wave envelope broadening by calculating the first quartile within each grid cell and reflecting it in a two-dimensional spatial distribution.
The analysis confirmed that the 90% signal duration depends on the hypocentral distance. The variation in the logarithmic deviation from the regression line was evaluated using the standard deviation, which was significantly smaller than that obtained when applying the peak delay time across all frequency bands. This suggests that using the 90% signal duration is becoming a more stable estimation method.
The values in the two-dimensional spatial distribution of logarithmic deviations were relatively large near the volcanoes distributed in the Ou Mountains and on its forearc side, with particularly strong deviations on the high-frequency side. This trend was also confirmed by the peak delay time analysis. In these areas, short-wavelength inhomogeneous structures that cause seismic wave scattering are presumed to be present. On the other hand, relatively small values were observed off the Sea of Japan of Akita and Yamagata prefectures, as well as in northern Aomori prefecture. In particular, the former region is different from previous trends. Several earthquakes that occurred in this region exhibited fluctuations in the detection of direct wave travel time. The introduction of a new indicator may have contributed to resolving this uncertainty.