10:45 AM - 12:15 PM
[SSS10-P13] Quantitative evaluation of inland earthquakes after the 2011 Tohoku earthquake derived from the refined earthquake catalog
Keywords:Earthquake catalog, The 2011 Tohoku earthquake, ETAS model
After the 2011 Tohoku earthquake, seismic activity increased not only near the epicenter but also inland throughout Japan (e.g., Hirose et al., 2011). On the other hand, due to the large number of earthquakes, microearthquakes (inland M<2, offshore M<3) were not always catalogued, making it difficult to comprehensively quantify the changes in seismic activity due to the 2011 Tohoku earthquake. Tamaribuchi and Kudo (2022, SSJ Fall Meeting) developed an automatic hypocenter determination method using machine learning and detected "hidden" earthquakes that were three times larger than those in the conventional earthquake catalog from March 2011 to February 2012. In this study, we extracted inland seismic activity throughout Japan from the obtained merged catalog of the automatically determined catalog and the JMA unified catalog, and quantified seismic activity after the 2011 Tohoku earthquake based on the ETAS model.
First, ETAS model parameters were obtained for each 0.4-degree grid using the JMA unified catalog (inland, M>=1, depth<=30 km) from January 2000 to February 2011. Then, using the obtained ETAS parameters, the stochastic declustering of Zhuang et al. (2002) was applied to the merged catalog from March 2011 to February 2012. The change in background seismicity μ was estimated from the slope of the cumulative mainshock probability obtained by the declustering. The other parameters of the ETAS model (K, α, c, p) are assumed to be time-invariant in each grid.
The obtained relative changes in seismic activity μ show that, on average, the seismicity is 2-3 times higher than before the Tohoku earthquake, although the variation from grid to grid is large. The background seismicity rate in April 2011 was estimated to be about 2.5 times (1.6-4.5 times) higher than before within 400 km of the Tohoku earthquake epicenter, and about 1.7 times (1.2-2.5 times) higher than before between 400 km and 800 km from the epicenter. The background seismicity rate gradually returned to its original level over time, and in February 2012, the background seismicity within 400 km of the epicenter was about 1.3 times higher than before.
In addition, comparing the background seismicity with the maximum shear strain rate estimated for each grid based on the one-month difference in GNSS data, a positive correlation was found: the background seismicity was about 1.7 times higher at a strain rate of 1.5×103 nstrain/yr, and about 3 times higher at 6×103 nstrain/yr. These results may be useful for quantifying the increase in inland seismicity after future large earthquakes.
Acknowledgments
We used the waveforms from JMA, NIED, universities, and institutions and the JMA unified earthquake catalog. We used GEONET (F5) from GSI. We used etas_solve from Kasahara et al. (2016) to estimate ETAS parameters, and Shen et al. (1996, 2015) to calculate strain rates.
First, ETAS model parameters were obtained for each 0.4-degree grid using the JMA unified catalog (inland, M>=1, depth<=30 km) from January 2000 to February 2011. Then, using the obtained ETAS parameters, the stochastic declustering of Zhuang et al. (2002) was applied to the merged catalog from March 2011 to February 2012. The change in background seismicity μ was estimated from the slope of the cumulative mainshock probability obtained by the declustering. The other parameters of the ETAS model (K, α, c, p) are assumed to be time-invariant in each grid.
The obtained relative changes in seismic activity μ show that, on average, the seismicity is 2-3 times higher than before the Tohoku earthquake, although the variation from grid to grid is large. The background seismicity rate in April 2011 was estimated to be about 2.5 times (1.6-4.5 times) higher than before within 400 km of the Tohoku earthquake epicenter, and about 1.7 times (1.2-2.5 times) higher than before between 400 km and 800 km from the epicenter. The background seismicity rate gradually returned to its original level over time, and in February 2012, the background seismicity within 400 km of the epicenter was about 1.3 times higher than before.
In addition, comparing the background seismicity with the maximum shear strain rate estimated for each grid based on the one-month difference in GNSS data, a positive correlation was found: the background seismicity was about 1.7 times higher at a strain rate of 1.5×103 nstrain/yr, and about 3 times higher at 6×103 nstrain/yr. These results may be useful for quantifying the increase in inland seismicity after future large earthquakes.
Acknowledgments
We used the waveforms from JMA, NIED, universities, and institutions and the JMA unified earthquake catalog. We used GEONET (F5) from GSI. We used etas_solve from Kasahara et al. (2016) to estimate ETAS parameters, and Shen et al. (1996, 2015) to calculate strain rates.