5:15 PM - 7:15 PM
[SVC32-P23] Application of the PF method to volcanic earthquakes on Izu-Oshima
Keywords:volcanic earthquake, hypocenter determination
The distribution of hypocenters and their temporal changes in volcanic earthquakes are one of the most important factors in constraining eruption locations and timings. Japan Meteorological Agency (JMA) operates 24-hour monitoring of seismic waveforms and hypocenter determination. However, most of the seismic event detection and phase picking is manually operated, making it challenging to promptly grasp the spatiotemporal distribution of hypocenters during periods of increased volcanic activity. In JMA's seismic event unification process, the automated hypocenter determination using the PF method (Tamaribuchi et al., 2016) has been employed, achieving results that allow for necessary accuracy in hypocenter determination even during times of frequent earthquakes.
In this study, we examined the applicability of the PF method in a volcanic earthquake observation network. We applied the PF method for hypocenter determination using data from the 2021 swarm activity on Izu-Oshima volcano, and compared it with the hypocenters determined manually by JMA.
The PF method for hypocenter calculation involves the following steps: (1) Reading P and S phases using AR-AIC based on the variance ratio and CF method (Allen, 1978) applied to band-pass filtered waveforms (5-10Hz). (2) Grouping the picked P and S phases with hypothetical hypocenters using likelihood calculations. (3) Using the result of the grouping and theoretical travel times to read P and S phases from the raw waveforms, determining the picked times used for hypocenter determination. (4) Determining the hypocenter from the picked P and S time. In this study, we used a modified program from Tamaribuchi et al. (2016) that includes the HYPOMH (Hirata and Matsu'ura, 1987) for travel time calculations and hypocenter determination.
We applied the PF method to seismic data from December 6, 2021, when there was swarm activity on Izu-Oshima. The seismic data used for this period included 14 observation points (5 from the ERI, 4 from NIED, and 5 from JMA). The PF method determined 521 hypocenters over 24 hours. In comparison, JMA's volcanic monitoring recorded 581 events during the same period, with hypocenters determined for 318 of them. Notably, during the hour with 125 registered events, the PF method determined 101 hypocenters, indicating a drop in event detection rates during periods of frequent earthquakes. During this period, earthquakes generally occurred along the western coast of Izu Oshima. Hypocenters determined by the PF method also distributed along the western coast but showed greater scatter compared to manual determinations. Comparing picked values from the PF method and manual determinations revealed higher ratios of events with large time differences at some observation points. These results suggest that the PF method can detect volcanic earthquakes and estimate hypocenters for volcanic activity monitoring.
In this study, we examined the applicability of the PF method in a volcanic earthquake observation network. We applied the PF method for hypocenter determination using data from the 2021 swarm activity on Izu-Oshima volcano, and compared it with the hypocenters determined manually by JMA.
The PF method for hypocenter calculation involves the following steps: (1) Reading P and S phases using AR-AIC based on the variance ratio and CF method (Allen, 1978) applied to band-pass filtered waveforms (5-10Hz). (2) Grouping the picked P and S phases with hypothetical hypocenters using likelihood calculations. (3) Using the result of the grouping and theoretical travel times to read P and S phases from the raw waveforms, determining the picked times used for hypocenter determination. (4) Determining the hypocenter from the picked P and S time. In this study, we used a modified program from Tamaribuchi et al. (2016) that includes the HYPOMH (Hirata and Matsu'ura, 1987) for travel time calculations and hypocenter determination.
We applied the PF method to seismic data from December 6, 2021, when there was swarm activity on Izu-Oshima. The seismic data used for this period included 14 observation points (5 from the ERI, 4 from NIED, and 5 from JMA). The PF method determined 521 hypocenters over 24 hours. In comparison, JMA's volcanic monitoring recorded 581 events during the same period, with hypocenters determined for 318 of them. Notably, during the hour with 125 registered events, the PF method determined 101 hypocenters, indicating a drop in event detection rates during periods of frequent earthquakes. During this period, earthquakes generally occurred along the western coast of Izu Oshima. Hypocenters determined by the PF method also distributed along the western coast but showed greater scatter compared to manual determinations. Comparing picked values from the PF method and manual determinations revealed higher ratios of events with large time differences at some observation points. These results suggest that the PF method can detect volcanic earthquakes and estimate hypocenters for volcanic activity monitoring.