9:45 AM - 10:00 AM
[SSS13-04] Surface wave monitoring using microtremor signal for detecting temporal variation of underground structure
Keywords:SPAC, Surface wave, Array
We identified the subsurface structure and detected the temporal variation of shallow subsurface structure using a seismic array on a landslide body. Using the same array, Oba et al. (2021) analyzed signals from ACROSS, which is a precisely controlled seismic source located 3 km away, to investigate the effects of rainfall on the transfer function from the ACROSS signal. They showed that the arrival time of the later phase significantly varies in response to rainfalls. The temporal variation may reflect the velocity change in the shallow subsurface structure, which undoubtedly can be influenced by an increase/decrease of groundwater level or pore pressure. The variation of subsurface structure can be monitored by the temporal variation of the propagation property of seismic wave including surface wave. Therefore, we investigated the relationship between the precipitation and the temporal variation of subsurface structure inferred from surface wave observation.
We used Rayleigh wave for monitoring because Rayleigh waves are more sensitive than Love waves concerning groundwater.
We used a seismic array that consists of 10 seismometers following the typical array configuration of the SPAC method (a center and three triangles on concentric circles with radii of 4,8,16 meters) at a landslide area in Mori town, in the western part of Shizuoka prefecture. The 10 seismometers are Mark Products L-28-3D (sensitivity 120V/m/s) of three components velocity-type ones with a natural frequency of 4.5Hz. The sampling frequency is 1kHz and the inter-channel skew of sampling is maintained within less than 15 µs. The microtremor signal is recorded continuously with one single AD unit synchronized with the GPS clock to avoid the effect of independent clock drift of multi-data logger observation. We used the data from October 2020 to June 2021.
In this analysis, we directly compared the observed SPAC curve with the theoretical SPAC curve which considering the effect of the higher-mode of Rayleigh wave (Asten,2006; Ikeda,2012) and inverted the subsurface structure for some subarrays with different dimensions. The inverted subsurface structures have a large difference among them, probably due to the lateral inhomogeneities. Therefore, we analyzed the data for smaller arrays (equilateral triangle array with a side length of 6.93meters) in the central part of the site and found the lateral inhomogeneities near the surface. The inversion results were compared with a 10m-depth borehole drilled in the center of the array.
To prevent the trade-off, we fixed the layer depths and discussed the monthly variation of the S wave velocity in the layers. We found a strong negative correlation between the S wave velocity in the near-surface layers and monthly precipitation. i.e., the more precipitation, the lower S-wave velocity we obtained by the inversion.
Considering the unstableness of the inversion, the temporal variation of the SPAC curve is also examined. We extracted the temporal variation of the SPAC curve from daily stacking data, using the average of the nine months' SPAC as the reference. We try to compare the daily precipitation with the temporal variation of the SPAC curve and found that the higher frequency range of the SPAC curve rise in the period when the precipitation was high, showing the decrease of the S wave velocity in near-surface layers. We also found that the temporal variation of the SPAC curve is also affected by human activity depending on days of weak and holidays. This effect should be removed in future studies.
We used Rayleigh wave for monitoring because Rayleigh waves are more sensitive than Love waves concerning groundwater.
We used a seismic array that consists of 10 seismometers following the typical array configuration of the SPAC method (a center and three triangles on concentric circles with radii of 4,8,16 meters) at a landslide area in Mori town, in the western part of Shizuoka prefecture. The 10 seismometers are Mark Products L-28-3D (sensitivity 120V/m/s) of three components velocity-type ones with a natural frequency of 4.5Hz. The sampling frequency is 1kHz and the inter-channel skew of sampling is maintained within less than 15 µs. The microtremor signal is recorded continuously with one single AD unit synchronized with the GPS clock to avoid the effect of independent clock drift of multi-data logger observation. We used the data from October 2020 to June 2021.
In this analysis, we directly compared the observed SPAC curve with the theoretical SPAC curve which considering the effect of the higher-mode of Rayleigh wave (Asten,2006; Ikeda,2012) and inverted the subsurface structure for some subarrays with different dimensions. The inverted subsurface structures have a large difference among them, probably due to the lateral inhomogeneities. Therefore, we analyzed the data for smaller arrays (equilateral triangle array with a side length of 6.93meters) in the central part of the site and found the lateral inhomogeneities near the surface. The inversion results were compared with a 10m-depth borehole drilled in the center of the array.
To prevent the trade-off, we fixed the layer depths and discussed the monthly variation of the S wave velocity in the layers. We found a strong negative correlation between the S wave velocity in the near-surface layers and monthly precipitation. i.e., the more precipitation, the lower S-wave velocity we obtained by the inversion.
Considering the unstableness of the inversion, the temporal variation of the SPAC curve is also examined. We extracted the temporal variation of the SPAC curve from daily stacking data, using the average of the nine months' SPAC as the reference. We try to compare the daily precipitation with the temporal variation of the SPAC curve and found that the higher frequency range of the SPAC curve rise in the period when the precipitation was high, showing the decrease of the S wave velocity in near-surface layers. We also found that the temporal variation of the SPAC curve is also affected by human activity depending on days of weak and holidays. This effect should be removed in future studies.