S34B-07
Depth-dependence of time-lapse seismic velocity change detected by a joint interferometric analysis of vertical array data

Wednesday, 16 December 2015: 17:30
307 (Moscone South)
Kaoru Sawazaki1, Tatsuhiko Saito1, Tomotake Ueno1,2 and Katsuhiko Shiomi1, (1)National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan, (2)NIED National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan
Abstract:
In this study, utilizing depth-sensitivity of interferometric waveforms recorded by co-located Hi-net and KiK-net sensors, we separate the responsible depth of seismic velocity change associated with the M6.3 earthquake occurred on November 22, 2014, in central Japan. The Hi-net station N.MKGH is located about 20 km northeast from the epicenter, where the seismometer is installed at the 150 m depth. At the same site, the KiK-net has two strong motion seismometers installed at the depths of 0 and 150 m. To estimate average velocity change around the N.MKGH station, we apply the stretching technique to auto-correlation function (ACF) of ambient noise recorded by the Hi-net sensor. To evaluate sensitivity of the Hi-net ACF to velocity change above and below the 150 m depth, we perform a numerical wave propagation simulation using 2-D FDM. To obtain velocity change above the 150 m depth, we measure response waveform from the depths of 150 m to 0 m by computing deconvolution function (DCF) of earthquake records obtained by the two KiK-net vertical array sensors. The background annual velocity variation is subtracted from the detected velocity change. From the KiK-net DCF records, the velocity reduction ratio above the 150 m depth is estimated to be 4.2 % and 3.1 % in the periods of 1-7 days and 7 days - 4 months after the mainshock, respectively. From the Hi-net ACF records, the velocity reduction ratio is estimated to be 2.2 % and 1.8 % in the same time periods, respectively. This difference in the estimated velocity reduction ratio is attributed to depth-dependence of the velocity change. By using the depth sensitivity obtained from the numerical simulation, we estimate the velocity reduction ratio below the 150 m depth to be lower than 1.0 % for both time periods. Thus the significant velocity reduction and recovery are observed above the 150 m depth only, which may be caused by strong ground motion of the mainshock and following healing in the shallow ground.