5:15 PM - 7:15 PM
[SCG63-P05] Development of a real-time epicenter estimation system using array analyses with a permanent seismic network
Keywords:array analysis, real-time analysis
Array analyses are a powerful technique to investigate the properties of wave propagation. Because the analyses require many instruments in a narrow area, array analyses have been conducted with a dense temporal observation network. However, the recent development of a permanent observation network has enabled us to conduct array analyses with these permanent stations. The advantages of the use of a permanent seismic network in array analyses are its applicability in real-time and its availability of multiple subarrays.
Seismic waves are excited by not only usual earthquakes but also fluid movements, landslides, explosions, etc. Some of these events are important for disaster mitigation or national security. Hence, we start to develop a real-time epicenter estimation system using array analyses with a permanent seismic network. Real-time analyses of multiple subarrays enable us to monitor the propagation of seismic waves and to estimate epicenters.
We refer the Automated Event Location Using Mesh of Arrays method (AELUMA method, de Groot-Hedlin et al., 2018; Fan et al., 2019) to implement the system, but we modify a little bit to the framework of the AELUMA method, e.g., the system reads waveforms via standard output of “shmdump” command implemented in the WIN system, and the use of a particle filter method instead of a grid search to find epicenters. Further, we make monitors of waveforms and epicenters to visualize the analysis processes. At this time, event discrimination processes are not implemented.
We run the system with the waveforms of the nuclear tests conducted by North Korea and tsunamigenic event near the Sofu-gan volcano on October 9, 2023, recorded by the F-net broadband network. We confirmed that when the signal-to-noise ratio of waveforms was sufficiently high, the system could locate epicenters at the expected locations, i.e., the nuclear test site or the location of the Sofu sea mount. Computation time was less than 1 s when we used the waveforms of the F-net. Future work is to implement the event discrimination scheme from the results of array analyses which are obtained every 1 s.
Acknowledgment: We used the waveforms of the F-net broadband seismic network (https://doi.org/10.17598/NIED.0005) deployed by the National Research Institute for Earth Science and Disaster Resilience.
Seismic waves are excited by not only usual earthquakes but also fluid movements, landslides, explosions, etc. Some of these events are important for disaster mitigation or national security. Hence, we start to develop a real-time epicenter estimation system using array analyses with a permanent seismic network. Real-time analyses of multiple subarrays enable us to monitor the propagation of seismic waves and to estimate epicenters.
We refer the Automated Event Location Using Mesh of Arrays method (AELUMA method, de Groot-Hedlin et al., 2018; Fan et al., 2019) to implement the system, but we modify a little bit to the framework of the AELUMA method, e.g., the system reads waveforms via standard output of “shmdump” command implemented in the WIN system, and the use of a particle filter method instead of a grid search to find epicenters. Further, we make monitors of waveforms and epicenters to visualize the analysis processes. At this time, event discrimination processes are not implemented.
We run the system with the waveforms of the nuclear tests conducted by North Korea and tsunamigenic event near the Sofu-gan volcano on October 9, 2023, recorded by the F-net broadband network. We confirmed that when the signal-to-noise ratio of waveforms was sufficiently high, the system could locate epicenters at the expected locations, i.e., the nuclear test site or the location of the Sofu sea mount. Computation time was less than 1 s when we used the waveforms of the F-net. Future work is to implement the event discrimination scheme from the results of array analyses which are obtained every 1 s.
Acknowledgment: We used the waveforms of the F-net broadband seismic network (https://doi.org/10.17598/NIED.0005) deployed by the National Research Institute for Earth Science and Disaster Resilience.