3:30 PM - 5:00 PM
[SCG60-P04] Toward semi-realtime scenario fault estimation judged by ground motion time histories
After an earthquake, many studies have been conducted on methods for estimating faults from seismic observation records in real-time seismic information. In addition to single indices such as maximum acceleration, maximum velocity, and seismic intensity, estimation using envelope waveforms is also performed. For tsunami studies, potential source regions are estimated by comparing observed waveform records with pre-calculated waveform records.
Many seismic hazard assessments in the Seismic Hazard Maps (Scenario Seismic Hazard Maps) identify the seismic source faults of the National Seismic Hazard Maps 2020 published by the Headquarters for Earthquake Research Promotion, targeting active faults in Japan. Therefore, it is expected that pre-calculated ground motion time histories will be useful for earthquake evaluation to grasp which scenario earthquakes are similar to each other within a few hours to half a day when an earthquake occurs. Therefore, by referring to the calculated waveforms from the NIED J-SHIS database, a trial calculation was performed to determine which assumed fault the observed waveforms are close to in semi-realtime. Scenario seismic hazard maps are calculated for a total of 4 cases, a case in which two asperities are exchanged for each fault segment and a case in which the rupture starting points are exchanged. Also, if there are multiple fault segments, the number of combinations will increase.
In this trial calculation, the observed velocity waveform was obtained by integrating the observed acceleration waveform of NIED K-NET/KiK-net sampled at 100 Hz or 200 Hz, and the calculated velocity waveform of the scenario seismic hazard map sampled at 120 Hz, with a 0.1-1 Hz band-pass filter was applied, cross-correlation coefficients were obtained, and scenario fault candidates were ranked.
For the borehole and surface waveforms observed at the KiK-net Mashiki station (KMMH16) for the foreshocks of the 2016 Kumamoto earthquake, the Takano-Shirahata section of the Hinagu fault zone and the Uto section of the Futagawa fault zone ranked high as scenario fault candidates, indicating the effectiveness of the trial method. On the other hand, for the borehole and surface waveforms observed at the KiK-net Mashiki station for the mainshock of the 2016 Kumamoto earthquake, the north coast section of the Uto Peninsula for the Futagawa fault zone was extracted at the top. The Futagawa section of the Futagawa fault zone could not be extracted at the higher rank. The reason for this is that when the dominant periods of the observed waveform and the calculated waveform differ greatly, it is difficult to extract the fault only by cross-correlation regardless of the normalization of the amplitude.
Many seismic hazard assessments in the Seismic Hazard Maps (Scenario Seismic Hazard Maps) identify the seismic source faults of the National Seismic Hazard Maps 2020 published by the Headquarters for Earthquake Research Promotion, targeting active faults in Japan. Therefore, it is expected that pre-calculated ground motion time histories will be useful for earthquake evaluation to grasp which scenario earthquakes are similar to each other within a few hours to half a day when an earthquake occurs. Therefore, by referring to the calculated waveforms from the NIED J-SHIS database, a trial calculation was performed to determine which assumed fault the observed waveforms are close to in semi-realtime. Scenario seismic hazard maps are calculated for a total of 4 cases, a case in which two asperities are exchanged for each fault segment and a case in which the rupture starting points are exchanged. Also, if there are multiple fault segments, the number of combinations will increase.
In this trial calculation, the observed velocity waveform was obtained by integrating the observed acceleration waveform of NIED K-NET/KiK-net sampled at 100 Hz or 200 Hz, and the calculated velocity waveform of the scenario seismic hazard map sampled at 120 Hz, with a 0.1-1 Hz band-pass filter was applied, cross-correlation coefficients were obtained, and scenario fault candidates were ranked.
For the borehole and surface waveforms observed at the KiK-net Mashiki station (KMMH16) for the foreshocks of the 2016 Kumamoto earthquake, the Takano-Shirahata section of the Hinagu fault zone and the Uto section of the Futagawa fault zone ranked high as scenario fault candidates, indicating the effectiveness of the trial method. On the other hand, for the borehole and surface waveforms observed at the KiK-net Mashiki station for the mainshock of the 2016 Kumamoto earthquake, the north coast section of the Uto Peninsula for the Futagawa fault zone was extracted at the top. The Futagawa section of the Futagawa fault zone could not be extracted at the higher rank. The reason for this is that when the dominant periods of the observed waveform and the calculated waveform differ greatly, it is difficult to extract the fault only by cross-correlation regardless of the normalization of the amplitude.