9:15 AM - 9:30 AM
[STT36-02] Real-time seismic observation around the Tsugaru Strait based on automatic arrival time picking and hypocenter location using distributed acoustic sensing data
Keywords:Distributed acoustic sensing, Tsugaru Strait
Most parts of seismogenic zones are located in the offshore area; therefore, monitoring of seismic activity in offshore areas is important. However, the detectability and the resolution of the hypocenter location of offshore earthquakes are low due to the small number of permanent stations in offshore areas. Recently, distributed acoustic sensing (DAS) measurement has been used for seismic observations. DAS can measure temporal changes of strain or strain rate at intervals of several meters along a fiber-optic cable; therefore, spatially dense observation can be realized by DAS. The installation cost of DAS measurement is relatively low because already installed submarine fiber-optic cable can be used. In addition, since instruments for DAS measurement are installed on land, it is easy to use data in real time. Therefore, DAS is useful for monitoring seismic activity. In this study, to monitor seismic activity in the Tsugaru Strait, we conducted DAS observation (model by AP Sensing) using a submarine fiber-optic cable and constructed a system of automatic P- and S-waves picking and hypocenter location using obtained DAS data.
The DAS observation was conducted from June 26th to October 17th, 2023. To pick P- and S-wave arrivals, we used a deep neural network model, PhaseNet (Zhu and Beroza, 2018). Picking by PhaseNet was conducted in real time after DAS data was acquired. We used a model which was trained by using DAS strain rate waveforms of earthquakes in Tsugaru Strait observed by DAS measurement in 2022. After picking of P- and S-wave arrivals, we located their hypocenters by using a hypocenter location program, hypomh (Hirata and Matsu’ura, 1987). The results of the hypocenter location were transmitted to our institute.
For earthquakes which occurred within a few kilometers from the cable, the difference in epicenters determined using the DAS data and in the catalog by the Japan Meteorological Agency (JMA) is generally less than 5 km. For earthquakes which occurred within 20 kilometers from the cable, the difference is generally less than 10 km. This suggests that we can locate the earthquakes near the cable using only DAS data. The difference in epicenters determined by DAS and in the JMA catalog is larger in an east-west direction, because the cable extends mainly in a north-south direction. For earthquakes observed by DAS and listed in the JMA catalog, we located these earthquakes by both DAS and seismic network data by the Association for the Development of Earthquake Prediction (AS-net, Noguchi et al., 2017). The difference in epicenters determined in this study and in the JMA catalog was reduced to a few kilometers. Therefore, the combination of DAS and inland seismic network data is effective for hypocenter locations on a higher resolution.
In this study, approximately 25 earthquakes which are not listed in the JMA catalog were detected and located. These earthquakes occurred in the Tsugaru Strait with a magnitude of ~0 at a depth of several to 10 km. The signals of these earthquakes were small and observed in a limited area; therefore, it is considered that these earthquakes cannot be detected or located in the JMA catalog. In future studies, we will compare the detectability between this study and catalogs of previous studies and discuss the seismic activity based on the constructed catalog in the Tsugaru Strait on a local scale.
The DAS observation was conducted from June 26th to October 17th, 2023. To pick P- and S-wave arrivals, we used a deep neural network model, PhaseNet (Zhu and Beroza, 2018). Picking by PhaseNet was conducted in real time after DAS data was acquired. We used a model which was trained by using DAS strain rate waveforms of earthquakes in Tsugaru Strait observed by DAS measurement in 2022. After picking of P- and S-wave arrivals, we located their hypocenters by using a hypocenter location program, hypomh (Hirata and Matsu’ura, 1987). The results of the hypocenter location were transmitted to our institute.
For earthquakes which occurred within a few kilometers from the cable, the difference in epicenters determined using the DAS data and in the catalog by the Japan Meteorological Agency (JMA) is generally less than 5 km. For earthquakes which occurred within 20 kilometers from the cable, the difference is generally less than 10 km. This suggests that we can locate the earthquakes near the cable using only DAS data. The difference in epicenters determined by DAS and in the JMA catalog is larger in an east-west direction, because the cable extends mainly in a north-south direction. For earthquakes observed by DAS and listed in the JMA catalog, we located these earthquakes by both DAS and seismic network data by the Association for the Development of Earthquake Prediction (AS-net, Noguchi et al., 2017). The difference in epicenters determined in this study and in the JMA catalog was reduced to a few kilometers. Therefore, the combination of DAS and inland seismic network data is effective for hypocenter locations on a higher resolution.
In this study, approximately 25 earthquakes which are not listed in the JMA catalog were detected and located. These earthquakes occurred in the Tsugaru Strait with a magnitude of ~0 at a depth of several to 10 km. The signals of these earthquakes were small and observed in a limited area; therefore, it is considered that these earthquakes cannot be detected or located in the JMA catalog. In future studies, we will compare the detectability between this study and catalogs of previous studies and discuss the seismic activity based on the constructed catalog in the Tsugaru Strait on a local scale.