5:15 PM - 7:15 PM
[SVC32-P32] Monitoring activities of Izu volcanic islands based on DAS observations using seafloor fiber-optic cables: preliminary results
Keywords:Distributed acoustic sensing, Izu Oshima, Miyake-jima, volcanic activity
Introduction:
Izu-Oshima and Miyake-jima are active volcanic islands that have periodically erupted in recent years, with the last eruptions occurring in 1986 for Izu-Oshima and 2000 for Miyake-jima. Monitoring seismic activities is crucial for assessing volcanic activities. However, as most of these volcanoes are located underwater, developing ocean floor observation systems is essential to enhance our monitoring capabilities.
Our goal is to develop monitoring systems of volcanic activities in the Izu islands, using the distributing acoustic sensing (DAS) with the seafloor cables of the Tokyo Metropolitan Government. In this study, we conducted preliminary DAS observations to evaluate its capability in monitoring seismic activities.
Observations:
We conducted DAS observations by installing equipment at the landing stations of Izu-Oshima and Miyake-jima from January to March 2024, using an AP Sensing N5226B R120 DAS interrogator. The observation period was about two weeks for each station. However, due to system difficulties, we only obtained four days of data from Miyake-jima cable. We could not receive GNSS radio signals at the stations, preventing timestamp corrections during the observations.
We first investigated the background noise levels, which significantly decreased at depths greater than 100 m, making the noise levels sufficiently low for seismic observations in most parts of the cable routes, where the water depths exceed 200 m.
Next, we detected earthquakes using data from the Izu-Oshima cable, employing the method of Baba et al. (2024AGU). We utilized PhaseNet (Zhu and Beroza, 2019GJI) to detect earthquake signals and pick P and S arrivals, using waveform traces obtained every 1 km from thousands of channels. We used 2-minute data with half-overlapping time windows for picking P and S arrival times. Grouping the picks with a simple algorithm, we determined hypocenters using the method of Hirata and Matsu’ura (1987, hypomh). We used the original PhaseNet model that was trained for land observation data. Despite the lower accuracy of phase pickings compared to that for land observation data due to of the difference of the waveform nature of seafloor DAS data, it was sufficient for this preliminary study.
We determined more than 600 hypocenters from the two-week data, mostly located beneath Izu-Oshima. Note that this result includes some duplicate events due to the use of half-overlapping time windows for picking arrival times. About 110 events corresponded to those listed in the JMA catalog, while the rest were newly detected by DAS observations. This result indicates that DAS observations improve our earthquake detection capabilities in this region, although some signals might originate from artificial activities.
Future plans:
Our results indicate that DAS seismic observations using seafloor cables will enhance our monitoring capabilities for volcanic activities in the Izu islands. In our observations, the timestamp of the observation system was not adjusted due to the inability to receive GNSS radio signals. The accuracy of phase picking for hypocenter determinations was not satisfactory. We plan to resolve these issues in future studies and observations, and to carry out further observations in the near future to develop real-time monitoring systems for the Izu volcanic islands using seafloor cables.
Acknowledgments:
We are grateful to the Tokyo Metropolitan Government and NTT East Corporation for permissions the use of seafloor cables and stations in our DAS experiments. This work was supported by JSPS KAKENHI Grant Numbers JP22H00251 and JP21H05205.
Izu-Oshima and Miyake-jima are active volcanic islands that have periodically erupted in recent years, with the last eruptions occurring in 1986 for Izu-Oshima and 2000 for Miyake-jima. Monitoring seismic activities is crucial for assessing volcanic activities. However, as most of these volcanoes are located underwater, developing ocean floor observation systems is essential to enhance our monitoring capabilities.
Our goal is to develop monitoring systems of volcanic activities in the Izu islands, using the distributing acoustic sensing (DAS) with the seafloor cables of the Tokyo Metropolitan Government. In this study, we conducted preliminary DAS observations to evaluate its capability in monitoring seismic activities.
Observations:
We conducted DAS observations by installing equipment at the landing stations of Izu-Oshima and Miyake-jima from January to March 2024, using an AP Sensing N5226B R120 DAS interrogator. The observation period was about two weeks for each station. However, due to system difficulties, we only obtained four days of data from Miyake-jima cable. We could not receive GNSS radio signals at the stations, preventing timestamp corrections during the observations.
We first investigated the background noise levels, which significantly decreased at depths greater than 100 m, making the noise levels sufficiently low for seismic observations in most parts of the cable routes, where the water depths exceed 200 m.
Next, we detected earthquakes using data from the Izu-Oshima cable, employing the method of Baba et al. (2024AGU). We utilized PhaseNet (Zhu and Beroza, 2019GJI) to detect earthquake signals and pick P and S arrivals, using waveform traces obtained every 1 km from thousands of channels. We used 2-minute data with half-overlapping time windows for picking P and S arrival times. Grouping the picks with a simple algorithm, we determined hypocenters using the method of Hirata and Matsu’ura (1987, hypomh). We used the original PhaseNet model that was trained for land observation data. Despite the lower accuracy of phase pickings compared to that for land observation data due to of the difference of the waveform nature of seafloor DAS data, it was sufficient for this preliminary study.
We determined more than 600 hypocenters from the two-week data, mostly located beneath Izu-Oshima. Note that this result includes some duplicate events due to the use of half-overlapping time windows for picking arrival times. About 110 events corresponded to those listed in the JMA catalog, while the rest were newly detected by DAS observations. This result indicates that DAS observations improve our earthquake detection capabilities in this region, although some signals might originate from artificial activities.
Future plans:
Our results indicate that DAS seismic observations using seafloor cables will enhance our monitoring capabilities for volcanic activities in the Izu islands. In our observations, the timestamp of the observation system was not adjusted due to the inability to receive GNSS radio signals. The accuracy of phase picking for hypocenter determinations was not satisfactory. We plan to resolve these issues in future studies and observations, and to carry out further observations in the near future to develop real-time monitoring systems for the Izu volcanic islands using seafloor cables.
Acknowledgments:
We are grateful to the Tokyo Metropolitan Government and NTT East Corporation for permissions the use of seafloor cables and stations in our DAS experiments. This work was supported by JSPS KAKENHI Grant Numbers JP22H00251 and JP21H05205.