17:15 〜 18:30
[STT35-P06] A preliminary attempt for automatic measurement of high frequency seismic signal duration
キーワード：高周波、震動継続時間、自動
Hara (2007, EPS) developed a measurement procedure of high frequency seismic signal duration (HFSD). It consists of baseline correction, high band-pass filtering, squaring of each data point, detection of P arrivals followed by manual checks and corrections, smoothing the time series by the moving window average, and measurements of HFSDs. The end time of HFSD is the time when the amplitude of the smoothed time series becomes smaller than 25 per cent of the maximum amplitude. Recently, Hara (2019, SSJ) showed that it was possible to reduce the scatters of the HFSD measurements to set again the width of the moving window for smoothing time series referring to the HFSD estimate obtained by the above mentioned procedure.
Since the manual check and correction of P arrival times are included in the procedure, it is difficult to make a whole procedure automatic. In the present study, we measured a duration from a time when the amplitude of the smoothed time series exceeds a threshold value to the end time to investigate whether it is possible to use this duration as a proxy for HFSD. We set the threshold value to 5 per cent as a preliminary attempt. We used theoretical P arrival times instead of picked P arrival times for smoothing time series. We analyzed 148 shallow large (moment magnitudes are greater than or equal to 7.2) earthquakes that occurred in the period from 1994 to July, 2020. We retrieved broadband waveform data recorded at the GSN stations in the epicentral distance range between 30 and 85 degrees from the IRIS DMC and used these data for duration measurements.
We compared the measured durations to HFSDs obtained by the procedure of Hara (2007). There is a good correlation between them. This result suggests that it would be possible to use durations measured by the approach of this study as a proxy of the HFSDs, and to make a whole measurement procedure automatic.
Since the manual check and correction of P arrival times are included in the procedure, it is difficult to make a whole procedure automatic. In the present study, we measured a duration from a time when the amplitude of the smoothed time series exceeds a threshold value to the end time to investigate whether it is possible to use this duration as a proxy for HFSD. We set the threshold value to 5 per cent as a preliminary attempt. We used theoretical P arrival times instead of picked P arrival times for smoothing time series. We analyzed 148 shallow large (moment magnitudes are greater than or equal to 7.2) earthquakes that occurred in the period from 1994 to July, 2020. We retrieved broadband waveform data recorded at the GSN stations in the epicentral distance range between 30 and 85 degrees from the IRIS DMC and used these data for duration measurements.
We compared the measured durations to HFSDs obtained by the procedure of Hara (2007). There is a good correlation between them. This result suggests that it would be possible to use durations measured by the approach of this study as a proxy of the HFSDs, and to make a whole measurement procedure automatic.