5:15 PM - 6:45 PM
[SCG40-P21] Increased activity of very low-frequency earthquakes in the southwest Ryukyu
Keywords:very low frequency earthquake, Ryukyu Trench, interplate coupling, crustal movement
Very low-frequency earthquakes (VLFEs) occur episodically in the southwestern Ryukyu Trench at intervals of several months (Nakamura and Sunagawa, 2015). Detailed spaciotemporal change in VLFE activity is essential for understanding the state of friction in the plate interface. However, Nakamura and Sunagawa’s (2015) method have poor accuracy in determining the epicenter. Therefore, long-term activity was investigated after improving the accuracy of VLFE epicenter determination, and the state of interplate coupling in the southwest Ryukyu Trench was clarified.
Broadband seismic waveform records from F-net and BATS were used to detect VLFEs; when both F-net and BATS were used, the analysis period was 2000-2022. On the other hand, if only the BATS is used, the analysis period is from 1998 to 2022. Since the F-net stations in the Ryukyu Islands were established between 2000 and 2001, the detection capability is expected to have changed significantly between 2000 and 2001. If only BATS is used, the activity since 1998 can be ascertained with the same detection capability. However, the accuracy of epicenter determination was lower than that of the F-net combination.
A template-matching method was used for VLFE detection and epicenter determination. Three-component waveforms at each station were bandpass filtered at 0.02-0.05 Hz and cross-correlation with the template earthquake was calculated. Thrust-type earthquakes and VLFEs have been used as template earthquakes. The locations and CMT solutions of the template VLFEs were determined using the grid search method. After removing earthquakes from the JMA catalogue and teleseismic earthquakes, events with a cross-correlation greater than 0.5 were determined to be VLFEs. The magnitude completeness of the F-net and BATS and the BATS only was determined to be 3.7 and 3.9, respectively.
The VLFEs obtained by F-net and BATS were distributed along the trench axis at slab depths of 10-15 km, with VLFEs concentrated at slab depths of 15 km from 123.0°E to 124.25°E. In contrast, from 124.25°E to 125.25°E, the VLFEs were distributed almost along the trench axis at a slab depth of 10 km. VLFE activity was absent in the west of 123.0°E. The VLFE activity began to increase around 2001, and in 2006 and 2015-2016. The 2001 activation was observed in the catalogue detected by BATS only.
The seismic activity and the occurrence rate of slow-slip events around Iriomote Island increased from to 2002-2006 and after 2013 (Nakamura and Kinjo, 2018; Tu and Heki, 2017). These were triggered by the 2002 afterslip and 2013 dyke intrusion in the Okinawa Trough. However, their timing did not correspond to the periods of VLFE activation in 2006 and 2015-2016. There were no significant intra-slab earthquakes in the area of VLFE occurrence during these two periods.
The 2001 VLFE activation started roughly in the second half of 2001., This period coincides with a period of frequent Mw 6-7 class earthquakes in the southwestern Ryukyu Trench. First, a Mw 5.9 thrust-type earthquake occurred near the trench axis on November 9, 2000. Another Mw 6.5 normal fault earthquake occurred off the southwest coast of Yonaguni Island on December 16, 2001. At that time, VLFE activity was triggered in the 123.0-124.0°E VLFE-occurring area; when a Mw6.5 thrust-type earthquake of 26 March 2002 occurred near the trench axis, a VLFE occurred hours later at 23.7°N, 124.5°E.
Thus, the VLFE activation observed in 2006 and 2015-2016 did not show the seismic activity that appeared to have triggered it; in the case of VLFE activation from late 2001, an interplate earthquake also occurred near the trench axis at the same time as the VLFE activation. This suggests that VLFE activation was influenced by factors other than those that activated tectonic activity in the Ryukyu Arc. This suggests that changes in interplate coupling in the VLFE region and long-term slow slip events might occasionally occur near the trench axis.
Broadband seismic waveform records from F-net and BATS were used to detect VLFEs; when both F-net and BATS were used, the analysis period was 2000-2022. On the other hand, if only the BATS is used, the analysis period is from 1998 to 2022. Since the F-net stations in the Ryukyu Islands were established between 2000 and 2001, the detection capability is expected to have changed significantly between 2000 and 2001. If only BATS is used, the activity since 1998 can be ascertained with the same detection capability. However, the accuracy of epicenter determination was lower than that of the F-net combination.
A template-matching method was used for VLFE detection and epicenter determination. Three-component waveforms at each station were bandpass filtered at 0.02-0.05 Hz and cross-correlation with the template earthquake was calculated. Thrust-type earthquakes and VLFEs have been used as template earthquakes. The locations and CMT solutions of the template VLFEs were determined using the grid search method. After removing earthquakes from the JMA catalogue and teleseismic earthquakes, events with a cross-correlation greater than 0.5 were determined to be VLFEs. The magnitude completeness of the F-net and BATS and the BATS only was determined to be 3.7 and 3.9, respectively.
The VLFEs obtained by F-net and BATS were distributed along the trench axis at slab depths of 10-15 km, with VLFEs concentrated at slab depths of 15 km from 123.0°E to 124.25°E. In contrast, from 124.25°E to 125.25°E, the VLFEs were distributed almost along the trench axis at a slab depth of 10 km. VLFE activity was absent in the west of 123.0°E. The VLFE activity began to increase around 2001, and in 2006 and 2015-2016. The 2001 activation was observed in the catalogue detected by BATS only.
The seismic activity and the occurrence rate of slow-slip events around Iriomote Island increased from to 2002-2006 and after 2013 (Nakamura and Kinjo, 2018; Tu and Heki, 2017). These were triggered by the 2002 afterslip and 2013 dyke intrusion in the Okinawa Trough. However, their timing did not correspond to the periods of VLFE activation in 2006 and 2015-2016. There were no significant intra-slab earthquakes in the area of VLFE occurrence during these two periods.
The 2001 VLFE activation started roughly in the second half of 2001., This period coincides with a period of frequent Mw 6-7 class earthquakes in the southwestern Ryukyu Trench. First, a Mw 5.9 thrust-type earthquake occurred near the trench axis on November 9, 2000. Another Mw 6.5 normal fault earthquake occurred off the southwest coast of Yonaguni Island on December 16, 2001. At that time, VLFE activity was triggered in the 123.0-124.0°E VLFE-occurring area; when a Mw6.5 thrust-type earthquake of 26 March 2002 occurred near the trench axis, a VLFE occurred hours later at 23.7°N, 124.5°E.
Thus, the VLFE activation observed in 2006 and 2015-2016 did not show the seismic activity that appeared to have triggered it; in the case of VLFE activation from late 2001, an interplate earthquake also occurred near the trench axis at the same time as the VLFE activation. This suggests that VLFE activation was influenced by factors other than those that activated tectonic activity in the Ryukyu Arc. This suggests that changes in interplate coupling in the VLFE region and long-term slow slip events might occasionally occur near the trench axis.