5:15 PM - 6:45 PM
[STT36-P11] Estimation of shallow structure along the Hinagu fault from the DAS observation
Keywords:Distributed Acoustic sensing (DAS), high density observation, heterogeneous structure, seismic interferometry
Introduction
Imaging the shallow subsurface structure is very important for urban development, such as building planning and use of underground space. Region particularly near faults can be severely damaged by ground rupture or local amplification, and it is significant to investigate and model the shallow subsurface velocity structure to evaluate these ground characteristics. Recently, Distributed Acoustic Sensing (DAS) using fiber optic cables has become popular as a new observation method in geoscience, and by using existing communication cables in urban areas, low-cost and ultra-high density observations have become possible. DAS has been used for various studies such as source determination and site characterization of volcanic earthquakes (Nishimura et al., 2021), shallow imaging and traffic noise detection in urban areas (Song et al., 2021), and fault zone imaging (Yang et al., 2022). In this study, we conducted DAS observations along National Route 3, which runs along the Hinagu Fault in Kumamoto Prefecture, where there is concern for further major seismic activity, to estimate the detailed shallow structure along the fault.
Data and Field
DAS observations were conducted for about one month from February to March 2023 by installing a Silixa Ltd. interrogator (iDAS) at the Kumamoto Maintenance branch office of the Ministry of Land, Infrastructure, Transport and Tourism (MLIT) and connecting it to a 40 km fiber optic cable along National Route 3 in a southerly direction to the Yatsushiro Maintenance branch office of MLIT. Strain rate waveforms were recorded along the cable on a total of 9,984 channels with 4 m channel spacing, 10 m gauge length, and 400 Hz sampling rate. During this period, 22 earthquakes were also observed.
Method
Three hours of daytime ambient noise recordings (between 13:00 and 15:00) were used for analysis along a 1.3 km section of the Hinagu Fault. The 3-hour recordings were divided into 10-minute time windows, band-pass filtered from 1 to 30 Hz, resampled at 100 Hz, waveforms normalized to 1-bit, and spectral whitening performed. Cross-correlation functions (CCFs) were computed for each channel, and finally phase-weighted stacking of the CCFs for each time window was executed. However, since it was difficult to obtain stable CCFs with this conventional seismic interferometric data processing, we applied the Three-station interferometry (TSI) method (Curtis and Halliday 2010, Song et al. 2010, Song et al. 2022). The TSI method uses three stations and applies repeated seismic interferometry to two CCFs to improve the signal-to-noise ratio of the extracted surface waves. The reconstructed CCFs were calculated over 320-channel and it divided into 80-channel sections with a slide of 10 channels, and dispersion curves were obtained for each section by multichannel analysis of surface wave (MASW).
Results
By repeatedly applying the TSI method three times, the reconstructed CCFs had an improved signal-to-noise ratio compared to the original CCFs and was able to extract surface wave propagation to long distances. The reconstructed CCFs produced a dispersion curve that was clearer and extended to higher frequencies. 3-8 Hz bandwidth Rayleigh wave fundamental modes were obtained, and 1D S-wave velocity inversion analysis was performed by reading the phase velocity of the fundamental modes. The shallow S-wave velocity structure was resolved to a depth of 180 m over about 1.3 km. The velocity structure was characterized by a generally soft sedimentary layer near the surface, but there was a sharp increase in the low velocity region with depth from the middle to the latter half of the analysis section.
Discussion and Conclusion
The shallow S-wave structure near the Hinagu fault was estimated from surface waves extracted by applying the TSI method based on seismic interferometry to DAS observations along National Route 3. The change in velocity from the middle to the latter half of the analysis section may be caused by the change in structure from solid volcanic ground to shallow thick sedimentary layers such as fans and deltas in the Yatsushiro Plain due to the gradual separation of a national highway that used to run parallel to the fault. Therefore, if a major earthquake were to occur on the Hinagu Fault, this area could be severely damaged. Detailed shallow structural investigations will continue to be necessary to limit damage.
(Acknowledgment) We used optical fiber cables owned by the Kumamoto Office of Rivers and National Highways, Kyushu Regional Development Bureau, Ministry of Land, Infrastructure, Transport and Tourism.
Imaging the shallow subsurface structure is very important for urban development, such as building planning and use of underground space. Region particularly near faults can be severely damaged by ground rupture or local amplification, and it is significant to investigate and model the shallow subsurface velocity structure to evaluate these ground characteristics. Recently, Distributed Acoustic Sensing (DAS) using fiber optic cables has become popular as a new observation method in geoscience, and by using existing communication cables in urban areas, low-cost and ultra-high density observations have become possible. DAS has been used for various studies such as source determination and site characterization of volcanic earthquakes (Nishimura et al., 2021), shallow imaging and traffic noise detection in urban areas (Song et al., 2021), and fault zone imaging (Yang et al., 2022). In this study, we conducted DAS observations along National Route 3, which runs along the Hinagu Fault in Kumamoto Prefecture, where there is concern for further major seismic activity, to estimate the detailed shallow structure along the fault.
Data and Field
DAS observations were conducted for about one month from February to March 2023 by installing a Silixa Ltd. interrogator (iDAS) at the Kumamoto Maintenance branch office of the Ministry of Land, Infrastructure, Transport and Tourism (MLIT) and connecting it to a 40 km fiber optic cable along National Route 3 in a southerly direction to the Yatsushiro Maintenance branch office of MLIT. Strain rate waveforms were recorded along the cable on a total of 9,984 channels with 4 m channel spacing, 10 m gauge length, and 400 Hz sampling rate. During this period, 22 earthquakes were also observed.
Method
Three hours of daytime ambient noise recordings (between 13:00 and 15:00) were used for analysis along a 1.3 km section of the Hinagu Fault. The 3-hour recordings were divided into 10-minute time windows, band-pass filtered from 1 to 30 Hz, resampled at 100 Hz, waveforms normalized to 1-bit, and spectral whitening performed. Cross-correlation functions (CCFs) were computed for each channel, and finally phase-weighted stacking of the CCFs for each time window was executed. However, since it was difficult to obtain stable CCFs with this conventional seismic interferometric data processing, we applied the Three-station interferometry (TSI) method (Curtis and Halliday 2010, Song et al. 2010, Song et al. 2022). The TSI method uses three stations and applies repeated seismic interferometry to two CCFs to improve the signal-to-noise ratio of the extracted surface waves. The reconstructed CCFs were calculated over 320-channel and it divided into 80-channel sections with a slide of 10 channels, and dispersion curves were obtained for each section by multichannel analysis of surface wave (MASW).
Results
By repeatedly applying the TSI method three times, the reconstructed CCFs had an improved signal-to-noise ratio compared to the original CCFs and was able to extract surface wave propagation to long distances. The reconstructed CCFs produced a dispersion curve that was clearer and extended to higher frequencies. 3-8 Hz bandwidth Rayleigh wave fundamental modes were obtained, and 1D S-wave velocity inversion analysis was performed by reading the phase velocity of the fundamental modes. The shallow S-wave velocity structure was resolved to a depth of 180 m over about 1.3 km. The velocity structure was characterized by a generally soft sedimentary layer near the surface, but there was a sharp increase in the low velocity region with depth from the middle to the latter half of the analysis section.
Discussion and Conclusion
The shallow S-wave structure near the Hinagu fault was estimated from surface waves extracted by applying the TSI method based on seismic interferometry to DAS observations along National Route 3. The change in velocity from the middle to the latter half of the analysis section may be caused by the change in structure from solid volcanic ground to shallow thick sedimentary layers such as fans and deltas in the Yatsushiro Plain due to the gradual separation of a national highway that used to run parallel to the fault. Therefore, if a major earthquake were to occur on the Hinagu Fault, this area could be severely damaged. Detailed shallow structural investigations will continue to be necessary to limit damage.
(Acknowledgment) We used optical fiber cables owned by the Kumamoto Office of Rivers and National Highways, Kyushu Regional Development Bureau, Ministry of Land, Infrastructure, Transport and Tourism.