S054-0013
Stress Drops of Inland Crustal Earthquake Sequences in Japan based on the S-wave Coda Spectral Ratio Method

Tuesday, 15 December 2020
Poster
Kazuhiro Somei1, Kimiyuki Asano2 and Tomotaka Iwata2, (1)Graduate School of Science, Kyoto University, Kyoto, Japan, (2)Disaster Prevention Research Institute, Kyoto University, Uji, Japan
Abstract:
To understand the difference of source physics between the mainshock and aftershock as well as what makes variation in stress drops, we have estimated corner frequency and stress drop of 369 events (Mw: 3.1-7.1) in 19 inland crustal earthquake sequences in Japan from 1995 to 2018. The observed source spectral ratio between large and small events was estimated from ground motion records of KiK-net (downhole data) and F-net (strong motion and broadband seismographs) by using the S-wave coda spectral ratio method (Somei et al., 2014). The commonality of decay curves of S-wave coda in each event pair was guaranteed by carefully examining the decay ratios of RMS envelopes in analyzed time windows of six narrowband frequency ranges. The source spectral ratios derived from S-wave coda are much more stable by station to station than those derived from direct S-wave, because the direct S-wave has the azimuth dependences on observed source spectra due to the rupture directivity effect. The observed source spectral ratios were well explained by the ω-2 source spectral model, and we estimated the corner frequency of each event by a grid search method.

The Brune-type stress drop was estimated from the seismic moment and the source radius (Eshelby, 1957), which was related to the corner frequency (Brune, 1970, 1971). The estimated stress drops of mainshocks and aftershocks varied in 0.8-18.6 and 0.01-25.8 MPa, respectively. We found the stress drops of most aftershocks do not exceed that of mainshock in each sequence. For the mainshocks, we could not see both the regional and mechanism dependences of stress drop. For the aftershocks, the dependence of stress drop on faulting style was not seen for both between and within the earthquake sequences. The stress drops of events excluding the mainshocks in our dataset increase with their source depths. We will also discuss the source dimensions obtained from the source spectrum and heterogeneous slip model.

Acknowledgements: Ground motion data from KiK-net and F-net were provided by the National Research Institute for Earth Science and Disaster Resilience. We used the moment tensor solutions and their seismic moments provided by F-net.