4:15 PM - 4:30 PM
[SVC28-16] Long-term base level variations of seismic power at Shinmoedake, revealing fluid migration prior to the 2011 and 2018 eruptions
Keywords:Volcano, Tremor, Monitoring
Shinmoedake is an active cone of the Kirishima Volcano Complex. It had two major magmatic eruptions in January 2011 and March 2018 after 300-year quiescence. Over a year before each event, significant inflation was reported from the GNSS observations, of which source was estimated about 7 km to the northwest of the crater and 8-10 km below the sea level (Nakao et al., 2013). Since the 2011 eruption, researchers have searched for other signs of magma migration before the eruptions. The long-term ambient-noise cross-correlation analyses were performed by Nishida et al. (2020) using seismic stations around the volcano. A gradual drop of seismic velocity was observed for one month before the 2011 eruption, but only at one pair of the stations closest to the crater. No significant velocity change was detected with the 2018 eruption. Kurihara et al. (2019) reported an increase of deep low-frequency earthquakes near Shinmoedake in correlation with the inflation before the 2011 eruption. However, they observed no corresponding sign associated with the 2018 eruption, either. On the other hand, the number of low-frequency earthquakes gradually increased while that of high-frequency events decreased at 10 km depth over half a year before the 2018 eruption (Yamada et al., 2019). It is still unclear when and how magma moved to the shallow depth beneath the Shinmoedake crater.
We investigated the long-term baseline variation of seismic power around Shinmoedake. We used the 1-7 Hz frequency bands, which were the main bands of seismic tremors associated with the eruptions (Ichihara and Matsumoto, 2017). Because the data were contaminated by noise from human activity in the daytime, we used two time-windows of 0:00-6:00 and 18:00-24:00 for each day. We collected the lowest 20 percentile of the power in the individual time windows at each station as the baseline.
The baseline variations generally correlated with the precipitation around the volcano. On the other hand, we found gradual increases of the base levels before the eruptions, which dominated the precipitation noises. The increases preceded four months before the 2011 eruption (since October 2010) and about a year before the 2018 eruption (since April 2017). In addition, the base level was high from July 2013 and January 2014, the beginning of which coincided with the restart of the extensions among the GNSS station pairs. As the base-level increases were significant at the closest stations to the crater, we infer that the source was at the shallow depth beneath it. We regard this seismic baseline analysis to reveal clear precursory signs for both the 2011 and 2018 eruptions. Although the same deep source resumed inflating and recovering right after both eruptions, they accompanied no such increase in the seismic powers. The baseline analysis will provide a simple and easy tool for monitoring fluid movement toward an eruption.
We investigated the long-term baseline variation of seismic power around Shinmoedake. We used the 1-7 Hz frequency bands, which were the main bands of seismic tremors associated with the eruptions (Ichihara and Matsumoto, 2017). Because the data were contaminated by noise from human activity in the daytime, we used two time-windows of 0:00-6:00 and 18:00-24:00 for each day. We collected the lowest 20 percentile of the power in the individual time windows at each station as the baseline.
The baseline variations generally correlated with the precipitation around the volcano. On the other hand, we found gradual increases of the base levels before the eruptions, which dominated the precipitation noises. The increases preceded four months before the 2011 eruption (since October 2010) and about a year before the 2018 eruption (since April 2017). In addition, the base level was high from July 2013 and January 2014, the beginning of which coincided with the restart of the extensions among the GNSS station pairs. As the base-level increases were significant at the closest stations to the crater, we infer that the source was at the shallow depth beneath it. We regard this seismic baseline analysis to reveal clear precursory signs for both the 2011 and 2018 eruptions. Although the same deep source resumed inflating and recovering right after both eruptions, they accompanied no such increase in the seismic powers. The baseline analysis will provide a simple and easy tool for monitoring fluid movement toward an eruption.