10:00 AM - 10:15 AM
[MTT39-05] Avalanche infrasounds observed at Mt. Shiribetsu, Hokkaido, during the winter seasons 2023/2024 and 2024/2025
Keywords:Infrasound, Mt.Shiribetsu, Avalanche
There is no avalanches early warning system , such as an earthquake early warning system. Since infrasounds generated by avalanche can be remotely observed, the early warning system for the avalanche can be developed, leading to disaster prevention and mitigation if this characteristic can be utilized. In this presentation, we report on the results of infrasound observations of avalanche conducted at Mt. Shiribetsu in Hokkaido.
An INF04LE (made by SAYA Inc.) was installed in a helicopter hangar approximately 3 km away from Mt. Shiribetsu from early December 2022 to early April 2023 (the first observation). We identified various noise sources distinct from avalanche infrasounds, such as helicopters and wheel loaders. We also detected sounds consistent with avalanche infrasound signatures reported in previous studies at around 9:36, 10 March. Satellite images show the location of the avalanche between 9 and 15 March. The frequency of satellite observations is once a day, and the ground surface is not visible when cloud cover is present. In addition, since no camera for avalanche monitoring was installed in the Mt. Shiribetsu area, it was not possible to limit the time of occurrence any further. Therefore, it was difficult to determine whether the infrasound we found was an avalanche-emitted.
Two INF04LEs and a 6000-16B (made by Paroscientific, hereinafter referred to as Paro), were installed in a triangular formation at intervals of approximately 1 km around Mt. Shiribetsu, so that array observations were conducted to identify avalanches using sensors only from late November 2023 to mid-April 2024 (the second observation). A sound with a peak below 15 Hz and lasting about one minute was found on all three sensors at around 16:17, 4 April. Based on information from a collaborator, the signal was attributed to fighter aircraft. We attempted to estimate the location of the sound with cross-correlation analysis, but it was difficult to find the location of the source due to a problem with the loggers and the inability to synchronize the time of each logger. Furthermore, we detected sounds that were different from local noises, but these sounds were also difficult to estimate their locations with cross-correlation analysis due to the same issues as mentioned above. Since we did not conduct a camera for the avalanche monitoring, we checked the satellite images and found several potential avalanche sites. However, we could not find the avalanche infrasound during that period.
We observed the infrasound with three Paros and a INF04LE from mid-November 2024 (the third observation). We expect to identify the location of the avalanche with the infrasound sensors alone if there are no malfunctions of them. Furthermore, we installed a camera for the monitoring of avalanche. The observation was continued until late April 2025.
In this presentation, we summarize the observations and discuss the results of them. Furthermore, we discuss the development of the avalanche early warning system. We consider a system to detect infrasound from avalanches and send notifications via LINE. The avalanche detection program notifies when signals with high similarity based on cross-correlation analysis match avalanche-specific frequency bands. We also consider using machine learning for identification of avalanches.
An INF04LE (made by SAYA Inc.) was installed in a helicopter hangar approximately 3 km away from Mt. Shiribetsu from early December 2022 to early April 2023 (the first observation). We identified various noise sources distinct from avalanche infrasounds, such as helicopters and wheel loaders. We also detected sounds consistent with avalanche infrasound signatures reported in previous studies at around 9:36, 10 March. Satellite images show the location of the avalanche between 9 and 15 March. The frequency of satellite observations is once a day, and the ground surface is not visible when cloud cover is present. In addition, since no camera for avalanche monitoring was installed in the Mt. Shiribetsu area, it was not possible to limit the time of occurrence any further. Therefore, it was difficult to determine whether the infrasound we found was an avalanche-emitted.
Two INF04LEs and a 6000-16B (made by Paroscientific, hereinafter referred to as Paro), were installed in a triangular formation at intervals of approximately 1 km around Mt. Shiribetsu, so that array observations were conducted to identify avalanches using sensors only from late November 2023 to mid-April 2024 (the second observation). A sound with a peak below 15 Hz and lasting about one minute was found on all three sensors at around 16:17, 4 April. Based on information from a collaborator, the signal was attributed to fighter aircraft. We attempted to estimate the location of the sound with cross-correlation analysis, but it was difficult to find the location of the source due to a problem with the loggers and the inability to synchronize the time of each logger. Furthermore, we detected sounds that were different from local noises, but these sounds were also difficult to estimate their locations with cross-correlation analysis due to the same issues as mentioned above. Since we did not conduct a camera for the avalanche monitoring, we checked the satellite images and found several potential avalanche sites. However, we could not find the avalanche infrasound during that period.
We observed the infrasound with three Paros and a INF04LE from mid-November 2024 (the third observation). We expect to identify the location of the avalanche with the infrasound sensors alone if there are no malfunctions of them. Furthermore, we installed a camera for the monitoring of avalanche. The observation was continued until late April 2025.
In this presentation, we summarize the observations and discuss the results of them. Furthermore, we discuss the development of the avalanche early warning system. We consider a system to detect infrasound from avalanches and send notifications via LINE. The avalanche detection program notifies when signals with high similarity based on cross-correlation analysis match avalanche-specific frequency bands. We also consider using machine learning for identification of avalanches.