5:15 PM - 7:15 PM
[STT41-P01] Comparison of peak ground velocities recorded by Hi-net and co-located KiK-net
Keywords:Hi-net, Peak velocity, KiK-net
The High Sensitivity Seismograph Network Japan (Hi-net), operated by NIED, is a nationwide high-density seismic observation network in Japan with an average station spacing of 20 km. Hi-net employs a high-sensitivity seismometer installed at the bottom of boreholes at depths of 100-3500 m for continuous observation of weak ground shaking from micro-earthquakes. Recently, studies to assess the seismic hazard caused by aftershocks following a large earthquake have progressed based on Hi-net data (Sawazaki et al. 2016; Sawazaki 2021). These studies forecasted the probability of subsurface shaking at the Hi-net seismometer location based on continuous records from Hi-net borehole seismometers. However, for seismic hazard assessment, it is essential to convert the probability of subsurface shaking to that at the surface. To achieve this, it is necessary to investigate the range of ground motions that Hi-net’s high-sensitivity seismometers can reliably observe, as well as their relationship to surface shaking. In this study, we compared the peak velocity recorded by Hi-net seismometers during earthquakes with those recorded by surface and borehole seismometers of KiK-net, which is co-located with Hi-net.
We selected approximately 40 Hi-net observation stations around the source region of the 2016 Kumamoto earthquake and the 2024 Noto Peninsula earthquake, giving priority to stations that had sufficiently recorded both the mainshock and immediate aftershocks. We then calculated peak velocity from the three-component continuous waveforms of Hi-net borehole seismometers during earthquake events. For common events, we also calculated peak velocity, peak acceleration, and peak of real-time seismic intensity from three-component triggered waveforms recorded by KiK-net borehole and surface seismometers. Here, we applied the instrument correction method (Maeda et al. 2011) to the Hi-net continuous waveform records to enhance the amplitude in the low-frequency range below 1 Hz compared to the raw records.
The peak velocity recorded by Hi-net and KiK-net borehole seismometers generally agreed in the range of 0.1 cm/s to several cm/s. When the peak velocity exceeded several cm/s, discrepancies between their values became apparent. These results suggest that Hi-net borehole seismometers with instrument correction can reliably record peak velocities in the range of 0.1 cm/s to several cm/s. Shiomi et al. (2005), who investigated the saturation of Hi-net seismometers, pointed out that saturation can occur even at 1 cm/s depending on the dominant period of seismic waves. Our results in which Hi-net seismometers can record peak velocity up to several cm/s can be attributed to the following factors: (1) Instrument response correction enabled the detection of low-frequency ground motion components. (2) Saturation of Hi-net seismometer is more likely to be caused by low-frequency shaking (Shiomi et al., 2005), and medium inland earthquakes, which are less likely to excite low-frequency shaking, were the main target of this analysis. (3) Even when seismometer saturation occurs, minor saturation does not significantly affect peak velocity measurements. Note that because of the characteristics of Hi-net instruments, the effect of seismometer saturation increases for ground motions larger than several cm/s, and the peak velocity recorded is likely to be underestimated, so care should be taken when handling them.
The peak velocity recorded by Hi-net borehole seismometers is linearly correlated with that of KiK-net surface seismometers with a slope close to 1 and a variation within a factor of two. This implies that by evaluating the amplification effects of subsurface structures, it is possible to convert the peak velocity recorded by Hi-net borehole seismometers into the surface peak velocity.
We selected approximately 40 Hi-net observation stations around the source region of the 2016 Kumamoto earthquake and the 2024 Noto Peninsula earthquake, giving priority to stations that had sufficiently recorded both the mainshock and immediate aftershocks. We then calculated peak velocity from the three-component continuous waveforms of Hi-net borehole seismometers during earthquake events. For common events, we also calculated peak velocity, peak acceleration, and peak of real-time seismic intensity from three-component triggered waveforms recorded by KiK-net borehole and surface seismometers. Here, we applied the instrument correction method (Maeda et al. 2011) to the Hi-net continuous waveform records to enhance the amplitude in the low-frequency range below 1 Hz compared to the raw records.
The peak velocity recorded by Hi-net and KiK-net borehole seismometers generally agreed in the range of 0.1 cm/s to several cm/s. When the peak velocity exceeded several cm/s, discrepancies between their values became apparent. These results suggest that Hi-net borehole seismometers with instrument correction can reliably record peak velocities in the range of 0.1 cm/s to several cm/s. Shiomi et al. (2005), who investigated the saturation of Hi-net seismometers, pointed out that saturation can occur even at 1 cm/s depending on the dominant period of seismic waves. Our results in which Hi-net seismometers can record peak velocity up to several cm/s can be attributed to the following factors: (1) Instrument response correction enabled the detection of low-frequency ground motion components. (2) Saturation of Hi-net seismometer is more likely to be caused by low-frequency shaking (Shiomi et al., 2005), and medium inland earthquakes, which are less likely to excite low-frequency shaking, were the main target of this analysis. (3) Even when seismometer saturation occurs, minor saturation does not significantly affect peak velocity measurements. Note that because of the characteristics of Hi-net instruments, the effect of seismometer saturation increases for ground motions larger than several cm/s, and the peak velocity recorded is likely to be underestimated, so care should be taken when handling them.
The peak velocity recorded by Hi-net borehole seismometers is linearly correlated with that of KiK-net surface seismometers with a slope close to 1 and a variation within a factor of two. This implies that by evaluating the amplification effects of subsurface structures, it is possible to convert the peak velocity recorded by Hi-net borehole seismometers into the surface peak velocity.