[S09P-12] 四国中央部で記録されたDASデータを用いた応力降下量の正確な推定に向けた試み
The precise estimation of corner frequency and stress drop of earthquakes is fundamentally important in understanding earthquake source physics. Distributed Acoustic Sensing (DAS) technique enables the enhancement of signal-to-noise ratio by stacking procedure due to its high spatial density measurements, resulting in the stable and reliable estimation of corner frequency for collocated earthquakes (e.g., Lior 2024).
In this study, we estimated stress drop using the empirical Green's function (EGF) deconvolution method. We conducted the DAS measurement from the middle of January to the end of March in 2024 using two fiber-optic cables along national road in central Shikoku. The total acquisition length was 53.8 km with a gauge length of 13.47 m, channel pitch of 4.96 m, and sampling frequency of 200 Hz. We applied EGF method to two co-located earthquakes (M2.4 & M0.3) which occurred nearby one of the cables.
We visually picked S-wave arrival times of the two collocated events and chose only less-noisy channels those does not contain large traffic noises. Based on the picked S-arrival time, we cut a 2.56 s time window to calculate the spectral ratio of the target earthquake and EGF event. To obtain more stable spectral ratio, we applied multiple time window approach by shifting the time window by 0.05 s and stacked a total of 10 spectral ratios for each available channel. Then, the spectral ratios were stacked with the available all channels and fitted with Boatwright and Brune source spectral models to estimate the corner frequency of the target event and its uncertainty. The fitting curve of the Brune model showed small uncertainty, and its corner frequency was smaller than that obtained by the Boatwright model. Assuming a circular fault model, we calculated the stress drop using the obtained corner frequency. Furthermore, we compared the stress drop by the DAS data with that estimated using waveform data recorded by regular seismometers of the nation-wide seismic network. The fitting curve of the seismometer data showed a slightly smaller corner frequency, but the spectral ratio at low- frequency range was almost equal to that by the DAS data. The current result demonstrates an effectiveness of DAS data to estimate more precise stress drop.
In this study, we estimated stress drop using the empirical Green's function (EGF) deconvolution method. We conducted the DAS measurement from the middle of January to the end of March in 2024 using two fiber-optic cables along national road in central Shikoku. The total acquisition length was 53.8 km with a gauge length of 13.47 m, channel pitch of 4.96 m, and sampling frequency of 200 Hz. We applied EGF method to two co-located earthquakes (M2.4 & M0.3) which occurred nearby one of the cables.
We visually picked S-wave arrival times of the two collocated events and chose only less-noisy channels those does not contain large traffic noises. Based on the picked S-arrival time, we cut a 2.56 s time window to calculate the spectral ratio of the target earthquake and EGF event. To obtain more stable spectral ratio, we applied multiple time window approach by shifting the time window by 0.05 s and stacked a total of 10 spectral ratios for each available channel. Then, the spectral ratios were stacked with the available all channels and fitted with Boatwright and Brune source spectral models to estimate the corner frequency of the target event and its uncertainty. The fitting curve of the Brune model showed small uncertainty, and its corner frequency was smaller than that obtained by the Boatwright model. Assuming a circular fault model, we calculated the stress drop using the obtained corner frequency. Furthermore, we compared the stress drop by the DAS data with that estimated using waveform data recorded by regular seismometers of the nation-wide seismic network. The fitting curve of the seismometer data showed a slightly smaller corner frequency, but the spectral ratio at low- frequency range was almost equal to that by the DAS data. The current result demonstrates an effectiveness of DAS data to estimate more precise stress drop.