5:15 PM - 6:45 PM
[SVC26-P19] The source process of the explosion earthquake at Sakurajima Volcano of long-period and very-long-period bands
Keywords:Explosion earthquake, Source process, Waveform inversion, Sakurajima volcano
1. Introduction
Explosion earthquakes accompanying Vulcanian eruptions represent phenomena related to the ejection of volcanic rock and ash and the release of volcanic gases. Explosion earthquakes have been observed at volcanoes around the world, and their dynamic processes have been clarified by source mechanism analysis in the very long-period band. Previous studies have revealed volume changes associated with crack opening by gas bubble growth (e.g., Lyons and Waite, 2011; Kim et al., 2014), reaction forces associated with volcanic ejecta, and drag forces associated with magma ascent (Ohminato et al., 2006). The analysis of explosion earthquakes has also been conducted on Sakurajima volcano. The source location of the explosion earthquakes has been estimated to be deeper than that of other volcanoes (Uhira and Takeo, 1994; Tameguri et al., 2002). On the other hand, a recent study has estimated the source location at a shallower location, several hundred meters from the crater, based on the first motion analysis of explosion earthquakes (Nishimura et al., submitted). In this study, we estimate the source time function in the very long-period band to investigate the correspondence with surface phenomena such as pyroclastic ejections associated with eruptions and clarify the eruption mechanism.
2. Data
Broadband seismometers have been operated at five stations located between 2 km and 8 km from the summit. From One broadband seismometer was temporally operated from April 29, 2022 to September 26, 2023. The number of explosions and eruption events during the temporary observation period was 359, according to the list of explosions and eruptions reported by the Kagoshima Meteorological Office. Of these events, two events were identified at all stations and components with a signal-to-noise ratio of 10 or higher in the very long-period band with a period of 0.1-0.5 Hz: the explosion event at 18:52 on July 17, 2022, and the explosion event at 20:05 on July 24, 2022. In this work, we analyzed observational data from two events.
3. Method
Strong amplitude spectra were identified for the two events in the very long-period band (0.01-0.5 Hz) and the long-period band (0.5-5 Hz) (Chouet and Matoza, 2013). A causal bandpass filter of 0.067-1 Hz was applied to the observed waveforms in the two events. The frequency domain waveform inversion analysis of Auger et al. (2006) was then performed on the filtered waveforms. A grid search was conducted to determine the best-fit source location. The grid search area was specified as 2000 m from east to west and north to south in the horizontal plane and from near the surface to 2400 m below sea level in the depth plane, surrounding the eruptive crater based on the distribution of the source location of the explosion earthquake (Nishimura et al. submitted). In our waveform inversion analysis, we consider three source mechanisms: (1) three single forces, (2) six moments, and (3) six moments components and three single forces.
4. Results and discussion
The analysis showed that the source model with six moments and three single forces was optimal for both events based on the residual and AIC values. From the obtained amplitude ratios of the source time functions, July 17 can be approximated by a single sill-like crack and July 24 by a sill-like crack and a dike-like crack. The source location was determined to be 480 m below sea level at an area about 800 m northwest of the crater on July 17, while on July 24, it was determined to be 440 m above sea level near the crater. The sill-like cracks are thought to function as a temporary buffer for the magma in both events, while the dike-like cracks are thought to serve as volcanic conduits. The single force is thought to be caused by the interaction between the source and the surrounding medium, which in the July 17 event is associated with rising magma and gas. The July 24 event may have acted as a reaction to the destruction of the lava plug and the drag from rising magma associated with ejection of pyroclastic materials and upward gas flow.
Acknowledgment: We used broadband seismic data provided by the Osumi office of River and National Highway, Ministry of Land, Infrastructure, Transport and Tourism (MLIT). This work was supported by ERI JURP 2022-M-15 in Earthquake Research Institute, the University of Tokyo and JSPS KAKENHI Grant Number JP22H05304.
Explosion earthquakes accompanying Vulcanian eruptions represent phenomena related to the ejection of volcanic rock and ash and the release of volcanic gases. Explosion earthquakes have been observed at volcanoes around the world, and their dynamic processes have been clarified by source mechanism analysis in the very long-period band. Previous studies have revealed volume changes associated with crack opening by gas bubble growth (e.g., Lyons and Waite, 2011; Kim et al., 2014), reaction forces associated with volcanic ejecta, and drag forces associated with magma ascent (Ohminato et al., 2006). The analysis of explosion earthquakes has also been conducted on Sakurajima volcano. The source location of the explosion earthquakes has been estimated to be deeper than that of other volcanoes (Uhira and Takeo, 1994; Tameguri et al., 2002). On the other hand, a recent study has estimated the source location at a shallower location, several hundred meters from the crater, based on the first motion analysis of explosion earthquakes (Nishimura et al., submitted). In this study, we estimate the source time function in the very long-period band to investigate the correspondence with surface phenomena such as pyroclastic ejections associated with eruptions and clarify the eruption mechanism.
2. Data
Broadband seismometers have been operated at five stations located between 2 km and 8 km from the summit. From One broadband seismometer was temporally operated from April 29, 2022 to September 26, 2023. The number of explosions and eruption events during the temporary observation period was 359, according to the list of explosions and eruptions reported by the Kagoshima Meteorological Office. Of these events, two events were identified at all stations and components with a signal-to-noise ratio of 10 or higher in the very long-period band with a period of 0.1-0.5 Hz: the explosion event at 18:52 on July 17, 2022, and the explosion event at 20:05 on July 24, 2022. In this work, we analyzed observational data from two events.
3. Method
Strong amplitude spectra were identified for the two events in the very long-period band (0.01-0.5 Hz) and the long-period band (0.5-5 Hz) (Chouet and Matoza, 2013). A causal bandpass filter of 0.067-1 Hz was applied to the observed waveforms in the two events. The frequency domain waveform inversion analysis of Auger et al. (2006) was then performed on the filtered waveforms. A grid search was conducted to determine the best-fit source location. The grid search area was specified as 2000 m from east to west and north to south in the horizontal plane and from near the surface to 2400 m below sea level in the depth plane, surrounding the eruptive crater based on the distribution of the source location of the explosion earthquake (Nishimura et al. submitted). In our waveform inversion analysis, we consider three source mechanisms: (1) three single forces, (2) six moments, and (3) six moments components and three single forces.
4. Results and discussion
The analysis showed that the source model with six moments and three single forces was optimal for both events based on the residual and AIC values. From the obtained amplitude ratios of the source time functions, July 17 can be approximated by a single sill-like crack and July 24 by a sill-like crack and a dike-like crack. The source location was determined to be 480 m below sea level at an area about 800 m northwest of the crater on July 17, while on July 24, it was determined to be 440 m above sea level near the crater. The sill-like cracks are thought to function as a temporary buffer for the magma in both events, while the dike-like cracks are thought to serve as volcanic conduits. The single force is thought to be caused by the interaction between the source and the surrounding medium, which in the July 17 event is associated with rising magma and gas. The July 24 event may have acted as a reaction to the destruction of the lava plug and the drag from rising magma associated with ejection of pyroclastic materials and upward gas flow.
Acknowledgment: We used broadband seismic data provided by the Osumi office of River and National Highway, Ministry of Land, Infrastructure, Transport and Tourism (MLIT). This work was supported by ERI JURP 2022-M-15 in Earthquake Research Institute, the University of Tokyo and JSPS KAKENHI Grant Number JP22H05304.