5:15 PM - 7:15 PM
[HDS07-P12] Temporal change of seismic wave characteristics associated with debris and slush flows at Mt. Fuji
Keywords:Mount Fuji, slush avalanche, lahar, seismic particle motion, seismic waveform data, spectral analysis
Debris flows are one of the primary volcanic hazards and their monitoring is crucial for disaster mitigation. At Mt. Fuji, debris flows occur due to snowmelt and heavy rainfall. In early spring and winter, slush flows occur when warm rain melts snow. To detect them, video cameras and wire sensors are commonly used, but seismic data allows for monitoring in poor visibility. Previous research estimated the locations of slush flows using seismic amplitude (Pérez-Guillén, 2019). However, the seismic amplitude for specific time periods (less than a few minutes) was only analyzed. To monitor the sediment and water movement for more than several hours, we analyzed the RMS amplitude, particle motion direction, and spectrum of seismic data on days when debris or slush flows occurred at Mt. Fuji.
We used three seismic stations at different elevations along the Osawa Valley on the western flank of Mt. Fuji: V.FUJ2 (3772 m height) and V.FUJD (1761 m height) operated by the Japan Meteorological Agency (JMA), and N.FJHV (571 m height) by the National Research Institute for Earth Science and Disaster Resilience. Meteorological data from JMA (AMeDAS stations) and the Yamanashi Prefecture Road Corporation were also used. We targeted five events (two debris flows in 2021 and three slush flows in 2018, 2021, and 2024) reported by the Mt. Fuji Sabo Office. The analysis focused on the 1–10 Hz range, where seismic energy was concentrated. The RMS amplitude was calculated from the vertical component, and the median was computed every five minutes to capture long-scale variations. To determine the particle motion direction, we calculated the covariance matrix of the two horizontal components every second and used its largest principal eigenvector as the direction. The median direction was then calculated every five minutes. Spectral analysis was performed using five-minute time windows with 50% overlap.
For all events, increases in amplitude, changes in particle motion direction, and increases in high-frequency energy were observed several hours before and after the debris flows identified by video camera along the Osawa Valley. In the slush flow on April 9, 2024, the RMS amplitude at the three stations increased around 1:00 AM, coinciding with the onset of increasing precipitation. The peak amplitude aligned with the period (8:20–10:10 AM) when multiple debris flows were detected by video camera and wire sensors. The particle motion direction at N.FJHV, initially oriented east-west at 6:00 AM, changed to a direction about 20° counterclockwise from north. At 7:30 AM, the stability in direction decreased at V.FUJD and the spectral peak changed from 1–2 Hz to 4–6 Hz. When the first slush flow was observed on camera at 8:20 AM, the stability in particle motion direction at N.FJHV increased. Additionally, an increase in energy at 2–3 Hz was observed at V.FUJ2, and at 2 Hz and 4 Hz at V.FUJD. After the precipitation decreased, the particle motion direction at N.FJHV showed decreased stability around 11:40 AM and returned to its original east-west orientation around 1:00 PM. During these changes, no sudden changes in wind speed were observed, and similar patterns of particle motion direction did not appear on days with heavy rainfall when no debris flows were reported. Therefore, these seismic characteristics should be associated with changes in sediment and water flow. In other regions, Ohsumi et al. (2008) reported that muddy water flowing ahead of debris flows generated seismic waves. This suggests that the seismic characteristics before debris flows at Mt. Fuji may be due to the excitation of seismic waves by muddy water.
These results indicate that the seismic parameters in this study have the potential to monitor changes in sediment and water flow for more than several hours at Mt. Fuji. Comparing these parameters with video data will enhance our understanding of their seismic characteristics.
We used three seismic stations at different elevations along the Osawa Valley on the western flank of Mt. Fuji: V.FUJ2 (3772 m height) and V.FUJD (1761 m height) operated by the Japan Meteorological Agency (JMA), and N.FJHV (571 m height) by the National Research Institute for Earth Science and Disaster Resilience. Meteorological data from JMA (AMeDAS stations) and the Yamanashi Prefecture Road Corporation were also used. We targeted five events (two debris flows in 2021 and three slush flows in 2018, 2021, and 2024) reported by the Mt. Fuji Sabo Office. The analysis focused on the 1–10 Hz range, where seismic energy was concentrated. The RMS amplitude was calculated from the vertical component, and the median was computed every five minutes to capture long-scale variations. To determine the particle motion direction, we calculated the covariance matrix of the two horizontal components every second and used its largest principal eigenvector as the direction. The median direction was then calculated every five minutes. Spectral analysis was performed using five-minute time windows with 50% overlap.
For all events, increases in amplitude, changes in particle motion direction, and increases in high-frequency energy were observed several hours before and after the debris flows identified by video camera along the Osawa Valley. In the slush flow on April 9, 2024, the RMS amplitude at the three stations increased around 1:00 AM, coinciding with the onset of increasing precipitation. The peak amplitude aligned with the period (8:20–10:10 AM) when multiple debris flows were detected by video camera and wire sensors. The particle motion direction at N.FJHV, initially oriented east-west at 6:00 AM, changed to a direction about 20° counterclockwise from north. At 7:30 AM, the stability in direction decreased at V.FUJD and the spectral peak changed from 1–2 Hz to 4–6 Hz. When the first slush flow was observed on camera at 8:20 AM, the stability in particle motion direction at N.FJHV increased. Additionally, an increase in energy at 2–3 Hz was observed at V.FUJ2, and at 2 Hz and 4 Hz at V.FUJD. After the precipitation decreased, the particle motion direction at N.FJHV showed decreased stability around 11:40 AM and returned to its original east-west orientation around 1:00 PM. During these changes, no sudden changes in wind speed were observed, and similar patterns of particle motion direction did not appear on days with heavy rainfall when no debris flows were reported. Therefore, these seismic characteristics should be associated with changes in sediment and water flow. In other regions, Ohsumi et al. (2008) reported that muddy water flowing ahead of debris flows generated seismic waves. This suggests that the seismic characteristics before debris flows at Mt. Fuji may be due to the excitation of seismic waves by muddy water.
These results indicate that the seismic parameters in this study have the potential to monitor changes in sediment and water flow for more than several hours at Mt. Fuji. Comparing these parameters with video data will enhance our understanding of their seismic characteristics.