10:45 〜 12:15
[SSS07-P15] S-netで記録された常時微動の自己相関解析による，海洋潮汐に伴う地震波速度変化
キーワード：常時微動、地震波速度変化、海洋潮汐
Microcracks or micropores in geomaterials cause variations in the elastic moduli under applied strain. The temporal variations in the elastic moduli of rocks alter the seismic wave velocity, which can be monitored to provide information on the strain applied to the crust. Previous studies have reported the temporal changes in the seismic wave velocity associated with the static strain changes induced by, for instance, large earthquakes and volcanic activities. As we can precisely compute the static strain caused by a solid earth tide, the seismic velocity changes related to the tidal strain provide information on the strain-velocity relationships on Earth. Recently, Takano and Nishida (2023) estimated the tidal response of the velocity variations in the shallow crust in Japan with a state-space model and proposed the criteria for detecting the tidal responses. In this work, we estimate the seismic velocity changes in response to ocean tides with a state-space model using autocorrelation functions of ambient noise recorded at S-net stations.
We determine the seismic velocity variations in response to ocean tides (M2 and K1 tide) from the hourly stacked ambient noise autocorrelations with an extended Kalman filter and Maximum Likelihood method based on a state-space model (Durbin and Koopman, 2012). We analyzed the ambient noise recorded at 150 three-component accelerometers of S-net. After the correction of the sensor orientations (Takagi et al., 2019), we computed three components (north-north, east-east, and upward-upward) of autocorrelation and six components (east-north, east-upward, north-east, north-upward, upward-east, and upward-north) of a single station cross-correlation (Hobiger et al., 2014). Before computing the noise correlations, we discarded the data, including transient amplitude anomalies by the root mean square value of the time segment. The correlation functions are filtered at frequency bands of 0.2-0.5 Hz and 0.8-1.6 Hz, respectively.
We compared the AIC between the hyperparameters, including the tidal response of the velocity variations. Among all seismic stations, 65% of the stations show a reliable tidal response to velocity variations. The velocity variations in response to the ocean tide are estimated to be up to about 0.6 %. We observe large velocity variations in response to tides at stations located in shallow water. This suggests that the response of the structure to applied static strain is different at shallow and deep water depths.
We determine the seismic velocity variations in response to ocean tides (M2 and K1 tide) from the hourly stacked ambient noise autocorrelations with an extended Kalman filter and Maximum Likelihood method based on a state-space model (Durbin and Koopman, 2012). We analyzed the ambient noise recorded at 150 three-component accelerometers of S-net. After the correction of the sensor orientations (Takagi et al., 2019), we computed three components (north-north, east-east, and upward-upward) of autocorrelation and six components (east-north, east-upward, north-east, north-upward, upward-east, and upward-north) of a single station cross-correlation (Hobiger et al., 2014). Before computing the noise correlations, we discarded the data, including transient amplitude anomalies by the root mean square value of the time segment. The correlation functions are filtered at frequency bands of 0.2-0.5 Hz and 0.8-1.6 Hz, respectively.
We compared the AIC between the hyperparameters, including the tidal response of the velocity variations. Among all seismic stations, 65% of the stations show a reliable tidal response to velocity variations. The velocity variations in response to the ocean tide are estimated to be up to about 0.6 %. We observe large velocity variations in response to tides at stations located in shallow water. This suggests that the response of the structure to applied static strain is different at shallow and deep water depths.