10:00 〜 10:15
[SVC31-05] クラックモデルと地震モーメントから推定される火山性低周波地震の震源特性とスケーリング関係
Long-period (LP) seismic events are thought to arise from the oscillations of a fluid-filled resonator. The frequencies and Q factors of observed LP waveforms vary depending on the resonator geometry and fluid properties, and the amplitudes vary depending on the fluid pressure change in the resonator. So, analyzing LP waveforms is important for improved monitoring of fluid states under volcanoes. The resonator geometry and fluid properties of LP sources have been estimated by comparing observed frequencies and Q factors to those calculated from a crack model. Taguchi et al. (JGR, 2018; GJI, 2021) developed a method to conduct the comparison using formulas for frequencies and Q factors of crack resonances. They applied the method to LP events at Kusatsu-Shirane volcano (Japan) between 1992 and 1993, then crack geometry and fluid properties can be estimated simultaneously. However, waveform inversions for the LP events revealed that crack aperture is smaller than its change estimated based on seismic moments for some events (Nakano & Kumagai, Fall Meeting of the Volcanological Society of Japan, 2021). To investigate this issue, we applied the analytical method of Taguchi et al. (2018, 2021) to more LP events during the same period.
In the method to estimate crack geometry and fluid properties, we compared the frequencies and Q factors of crack resonances to observed ones assuming the resonance mode for the lowest peak frequency. We applied this method to LP events at Kusatsu-Shirane in 1992-1993 assuming misty gas as the fluid and the mode with wavelength 2L/3 (L: crack length) for the lowest peak frequency. Then some events with large Q values could not be explained by any fluids, and our estimated crack aperture (d) are smaller than the aperture change (Δd) estimated from seismic moments for other many events, especially for those with large Q values. So, we assumed the mode with wavelength 2L/5 for the lowest peak frequencies, then all the events can be explained by the crack containing misty gas and d becomes much larger than Δd.
Then we found the following two features in our estimates of crack geometry: (1) the crack volume was roughly proportional to the seismic moment, and (2) Δd was constant at around 10-4 m regardless of the seismic moment. The features (1) and (2) show that we can estimate the product of crack length and width from seismic moments assuming Δd is 10-4 m. Furthermore, the feature (3) implies that the pressure in the crack is almost constant regardless of the seismic moment since Δd is considered to be proportional to the pressure (e.g. Maeda & Kumagai, GRL, 2013).
Our results imply that we can obtain more accurate source properties by incorporating seismic moments into the analytical method proposed by Taguchi et al. (2018, 2021). In addition, we can obtain the scaling relations for the LP events at Kusatsu-Shirane in 1992-1993. In the future, the analytical method incorporating seismic moments should be applied to LP events at other volcanoes to investigate the universality of our estimated scaling relations.
In the method to estimate crack geometry and fluid properties, we compared the frequencies and Q factors of crack resonances to observed ones assuming the resonance mode for the lowest peak frequency. We applied this method to LP events at Kusatsu-Shirane in 1992-1993 assuming misty gas as the fluid and the mode with wavelength 2L/3 (L: crack length) for the lowest peak frequency. Then some events with large Q values could not be explained by any fluids, and our estimated crack aperture (d) are smaller than the aperture change (Δd) estimated from seismic moments for other many events, especially for those with large Q values. So, we assumed the mode with wavelength 2L/5 for the lowest peak frequencies, then all the events can be explained by the crack containing misty gas and d becomes much larger than Δd.
Then we found the following two features in our estimates of crack geometry: (1) the crack volume was roughly proportional to the seismic moment, and (2) Δd was constant at around 10-4 m regardless of the seismic moment. The features (1) and (2) show that we can estimate the product of crack length and width from seismic moments assuming Δd is 10-4 m. Furthermore, the feature (3) implies that the pressure in the crack is almost constant regardless of the seismic moment since Δd is considered to be proportional to the pressure (e.g. Maeda & Kumagai, GRL, 2013).
Our results imply that we can obtain more accurate source properties by incorporating seismic moments into the analytical method proposed by Taguchi et al. (2018, 2021). In addition, we can obtain the scaling relations for the LP events at Kusatsu-Shirane in 1992-1993. In the future, the analytical method incorporating seismic moments should be applied to LP events at other volcanoes to investigate the universality of our estimated scaling relations.