11:30 AM - 11:45 AM
[S15-08] Comparative study of temporal variation of coseismic velocity in Korea and Japan
<Abstract>
Post-earthquake temporal variations can provide insightful information about dynamic processes and the state of crust around seismic area. Larger variations of earthquakes were naturally observed in large magnitude, and moderate Korean earthquake after long seismically quiet period showed larger variation in strong Japanease earthquake (Iqbal et al., 2021). For the quantitative study of variation, the technique of Ambient Noise Interferometry (ANI), which consists of cross-correlating continuous noise records from a stations pair acting as doublets, has been widely applied to numerous earthquakes. As done by Iqbal et al (2021), this study using ANI selected earthquakes with three kind magnitude – major, strong, and moderate. Major and strong earthquakes are Kumamoto (ML 7.3) and Tottori earthquake (ML 6.6) occurred in Japan at 2016, respectively, and moderate is Gyeongju (ML 5.8) and Pohang earthquake (ML 5.4) occurred in Korea at 2016 and 2017, respectively. Our study corrected synthetic seasonal velocity changes (Wang et al., 2017), which was not incorporated in the previous studies. The analyzed data ranged continuously from approximately 2 to 9.5 months before and 2 to 4 months after the earthquake. The velocity reduction was observed within the classes of 15 and 30 days before and after the earthquake day, respectively. Additionally, two further classes of 30 and 60 days, i.e., the sum of 15 and 30 days before and after the earthquake, respectively, were included in the analyses. The linear fitting applied to these classes with a regression value of R2 ≥ 0.5 showed the values of velocity drops for the targeted frequencies of 0.1-0.6 Hz, 0.5-0.9 Hz, and 0.1-0.9 Hz, respectively. The largest interstation velocity drop values obtained from the six classes showed that their distribution correlates well within the epicentral proximity of the Tottori and Gyeongju earthquakes. For Kumamoto earthquakes, the highest velocity variations were concentrated near the Aso volcano, while for the Pohang earthquake, these variations were significant near the Gyeongju earthquake epicentre that preceded it a year ago. For the seasonal corrections, two environmental factors: pore pressure change and sea level changes, were considered based on the observations in Korean data. The pore pressure was obtained from precipitation records (Talwani et al., 2007), and along with sea level changes, two factors were filtered with band-pass between 0.0073 and 0.05 cycles. The velocity time series, on the other hand, was smoothened by a low-pass filter at 0.05 cycles/day and correlated with two factors for correlation coefficients. The pore pressure was, additionally, shifted within ten days to obtain the best minima correlation and was combined with sea level changes to produce a synthetic curve. The smoothened velocity drop was considered only for interstation pairs with significant drop values for six classes, and the velocity drop was corrected by subtracting the synthetic curve from the observed drop values. The corrected results naturally showed reduced values of velocity changes compared to the uncorrected values. For the Gyeongju earthquake, high-velocity variations were prominent near the epicentre. On the other hand, noticeable distribution of high values around the hypocentre of the Gyeongju earthquake was observed after the Pohang earthquake. This significant velocity variation around preceded earthquakes may be explained by the fact that cracked crust due to the previous earthquakes was easily influenced by the later earthquakes. A similar phenomenon was observed at repeated seismic regions as greater variations induced by the smaller magnitude earthquakes than those by larger magnitude ones preceded earlier (Brenguier et al., 2008; Yu et al., 2020)
<Acknowledgement>
This work was funded by Research Project in 2021 Sejong University
Post-earthquake temporal variations can provide insightful information about dynamic processes and the state of crust around seismic area. Larger variations of earthquakes were naturally observed in large magnitude, and moderate Korean earthquake after long seismically quiet period showed larger variation in strong Japanease earthquake (Iqbal et al., 2021). For the quantitative study of variation, the technique of Ambient Noise Interferometry (ANI), which consists of cross-correlating continuous noise records from a stations pair acting as doublets, has been widely applied to numerous earthquakes. As done by Iqbal et al (2021), this study using ANI selected earthquakes with three kind magnitude – major, strong, and moderate. Major and strong earthquakes are Kumamoto (ML 7.3) and Tottori earthquake (ML 6.6) occurred in Japan at 2016, respectively, and moderate is Gyeongju (ML 5.8) and Pohang earthquake (ML 5.4) occurred in Korea at 2016 and 2017, respectively. Our study corrected synthetic seasonal velocity changes (Wang et al., 2017), which was not incorporated in the previous studies. The analyzed data ranged continuously from approximately 2 to 9.5 months before and 2 to 4 months after the earthquake. The velocity reduction was observed within the classes of 15 and 30 days before and after the earthquake day, respectively. Additionally, two further classes of 30 and 60 days, i.e., the sum of 15 and 30 days before and after the earthquake, respectively, were included in the analyses. The linear fitting applied to these classes with a regression value of R2 ≥ 0.5 showed the values of velocity drops for the targeted frequencies of 0.1-0.6 Hz, 0.5-0.9 Hz, and 0.1-0.9 Hz, respectively. The largest interstation velocity drop values obtained from the six classes showed that their distribution correlates well within the epicentral proximity of the Tottori and Gyeongju earthquakes. For Kumamoto earthquakes, the highest velocity variations were concentrated near the Aso volcano, while for the Pohang earthquake, these variations were significant near the Gyeongju earthquake epicentre that preceded it a year ago. For the seasonal corrections, two environmental factors: pore pressure change and sea level changes, were considered based on the observations in Korean data. The pore pressure was obtained from precipitation records (Talwani et al., 2007), and along with sea level changes, two factors were filtered with band-pass between 0.0073 and 0.05 cycles. The velocity time series, on the other hand, was smoothened by a low-pass filter at 0.05 cycles/day and correlated with two factors for correlation coefficients. The pore pressure was, additionally, shifted within ten days to obtain the best minima correlation and was combined with sea level changes to produce a synthetic curve. The smoothened velocity drop was considered only for interstation pairs with significant drop values for six classes, and the velocity drop was corrected by subtracting the synthetic curve from the observed drop values. The corrected results naturally showed reduced values of velocity changes compared to the uncorrected values. For the Gyeongju earthquake, high-velocity variations were prominent near the epicentre. On the other hand, noticeable distribution of high values around the hypocentre of the Gyeongju earthquake was observed after the Pohang earthquake. This significant velocity variation around preceded earthquakes may be explained by the fact that cracked crust due to the previous earthquakes was easily influenced by the later earthquakes. A similar phenomenon was observed at repeated seismic regions as greater variations induced by the smaller magnitude earthquakes than those by larger magnitude ones preceded earlier (Brenguier et al., 2008; Yu et al., 2020)
<Acknowledgement>
This work was funded by Research Project in 2021 Sejong University