10:45 AM - 12:15 PM
[SSS07-P16] Multi-decadal monitoring of seismic velocity changes beneath Izu-Oshima, central Japan, using ambient seismic noise records
Keywords:Seismic interferometry, Temporal variations of seismic velocity, Izu-Oshima
Izu-Oshima is an active volcano located at the northern part of Izu-Bonin arc, central Japan. The latest magmatic eruption occurred at the central cone, Miharayama, in 1986-1987. The observation of crustal deformation revealed the shrinkage of the volcanic edifice from the last eruption until the mid-1990s, following the expansion of the volcanic edifice suggesting the resumption of magma accumulation. In addition to this long-term crustal expansion, periodic expansion and contraction with a time interval of 1 to 2 years are also observed by the GNSS stations. The expansion and contraction sources are estimated at almost the same location, beneath the northern part of Miharayama (Shimamura et al., 2022). Miharayama has experienced the three previous eruptions that occurred at intervals of 36 to 38 years. As 30 years have already passed since the last eruption, suggesting that the next eruption is likely imminent.
Understanding temporal evolutions of crustal deformations in eruption cycles, especially pre-eruption periods, is crucial for improving our ability to perform short-term eruption forecasting. Seismic interferometry with ambient noise is a powerful continuous monitoring tool for identifying temporal variations of seismic velocity associated with subsurface magma movements at active volcanoes (e.g., Brenguier et al., 2008), sudden stress change induced by large earthquakes (e.g., Minato et al., 2012) and hydrological loading as well as fluid diffusion (e.g., Sens-Schönfelder and Wegler, 2006). In the present study, we applied seismic interferometry for continuous seismic recordings in the last 18 years collected at Izu-Oshima.
We used the vertical component of velocity waveforms recorded at the 39 seismic stations during 2003-2020. In calculating the cross-correlation functions (CCFs), we applied bandpass filters of three different frequency ranges (0.1-0.9 Hz, 0.5-2.0 Hz, 1.0-4.0 Hz) to each continuous waveform after spectrum whitening. We calculated the CCFs for all available station pairs and obtained a daily CCF. To obtain a stable record of the daily CCF, we stacked the CCFs using a moving time window of 5 days. We estimated the relative travel-time variations by comparing each daily CCF with reference one, using the moving-window cross-spectrum analysis (Clarke et al., 2011). Then, we estimated fluctuations of the velocity structure, assuming a spatially homogeneous seismic velocity change. To stabilize the temporal variation, we stacked the time series of velocity changes at several station pairs.
The estimated velocity fluctuations show a rapid decrease immediately after the 2011 Tohoku-oki Mw 9.0 Earthquake. The velocity reduction is pronounced in the low-frequency range (0.1-0.9 Hz). Such velocity reductions were also documented in other volcanic regions. Previous works suggest the redistributions of magmatic fluid due to large dynamic stress changes from the Tohoku Earthquake (Brenguier et al., 2014). We also detected periodic fluctuations of the seismic velocity (1 or 2 years) in the frequency range of 0.5-2.0 Hz. These short-term fluctuations of velocity structure are notably observed for the station pairs around the central cone, and are temporally correlated with crustal deformation from recurrent expansion and contraction of pressure sources identified by the geodetic data (Shimamura et al., 2022). The temporal variations may be associated with the changes in velocity structure around the pressure sources due to magmatic fluid intrusion or strain changes accompanied by the source deformations. Alternatively, changes in the station distance due to the periodic deformations could be responsible for short-time velocity changes.
Acknowledgments
We used the seismic data at the stations of the Earthquake Research Institute, Tokyo University, National Research Institute for Earth Science and Disaster Resilience, and Japan Meteorological Agency. We used the Python package of MSNoise (Lecocq et al., 2014).
Understanding temporal evolutions of crustal deformations in eruption cycles, especially pre-eruption periods, is crucial for improving our ability to perform short-term eruption forecasting. Seismic interferometry with ambient noise is a powerful continuous monitoring tool for identifying temporal variations of seismic velocity associated with subsurface magma movements at active volcanoes (e.g., Brenguier et al., 2008), sudden stress change induced by large earthquakes (e.g., Minato et al., 2012) and hydrological loading as well as fluid diffusion (e.g., Sens-Schönfelder and Wegler, 2006). In the present study, we applied seismic interferometry for continuous seismic recordings in the last 18 years collected at Izu-Oshima.
We used the vertical component of velocity waveforms recorded at the 39 seismic stations during 2003-2020. In calculating the cross-correlation functions (CCFs), we applied bandpass filters of three different frequency ranges (0.1-0.9 Hz, 0.5-2.0 Hz, 1.0-4.0 Hz) to each continuous waveform after spectrum whitening. We calculated the CCFs for all available station pairs and obtained a daily CCF. To obtain a stable record of the daily CCF, we stacked the CCFs using a moving time window of 5 days. We estimated the relative travel-time variations by comparing each daily CCF with reference one, using the moving-window cross-spectrum analysis (Clarke et al., 2011). Then, we estimated fluctuations of the velocity structure, assuming a spatially homogeneous seismic velocity change. To stabilize the temporal variation, we stacked the time series of velocity changes at several station pairs.
The estimated velocity fluctuations show a rapid decrease immediately after the 2011 Tohoku-oki Mw 9.0 Earthquake. The velocity reduction is pronounced in the low-frequency range (0.1-0.9 Hz). Such velocity reductions were also documented in other volcanic regions. Previous works suggest the redistributions of magmatic fluid due to large dynamic stress changes from the Tohoku Earthquake (Brenguier et al., 2014). We also detected periodic fluctuations of the seismic velocity (1 or 2 years) in the frequency range of 0.5-2.0 Hz. These short-term fluctuations of velocity structure are notably observed for the station pairs around the central cone, and are temporally correlated with crustal deformation from recurrent expansion and contraction of pressure sources identified by the geodetic data (Shimamura et al., 2022). The temporal variations may be associated with the changes in velocity structure around the pressure sources due to magmatic fluid intrusion or strain changes accompanied by the source deformations. Alternatively, changes in the station distance due to the periodic deformations could be responsible for short-time velocity changes.
Acknowledgments
We used the seismic data at the stations of the Earthquake Research Institute, Tokyo University, National Research Institute for Earth Science and Disaster Resilience, and Japan Meteorological Agency. We used the Python package of MSNoise (Lecocq et al., 2014).