4:45 PM - 5:00 PM
[U16-06] Features of nonlinear site responses at strong-motion sites during the 2024 Noto Peninsula Earthquake
Keywords:The 2024 Noto Peninsula Earthquake, Nonlinear site response, Strong ground motions, Degree of nonlinearity
The rupture heterogeneities and local site conditions strongly influence ground motions from the source faults at short distances. Nonlinear site response (NLSR) can significantly change the resonance frequency and amplification characteristics of the soli sites. Typically, a reduction of higher frequency components and a shift of predominant frequency to a lower one are well known. In this study, a preliminary analysis was carried out to identify and extract features of nonlinear site responses at more than 50 sites combined from KiK-net and K-NET, where peak ground velocities (PGVs) of 10 cm/s or over were recorded during the 2024, Mw 7.5, Noto Peninsula Earthquake (abbreviated as the 2024 mainshock).
A comparison of the surface-to-borehole S-wave spectral ratios (S/B) for the horizontal components between the 2024 mainshock and other weak motion records was made at the KiK-net sites, where records were available at the surface and borehole. At the K-NET stations, where records are available only at the surface, a horizontal-to-vertical spectral ratios (H/V) technique was adopted. In this study, weak motion records are defined as those with peak ground accelerations (PGAs) smaller than 30 cm/s/s. Time windows of 20 seconds were used to compute the S-wave Fourier spectra. The degree of nonlinearity (DNL) was calculated using the method of Noguchi and Sasatani (2008). A DNL value of 2.5 and 4 or larger obtained using the S/B and H/V methods may indicate that the site experienced NLSR.
Our analysis based on the S/B method indicated that KiK-net sites with PGAs larger than about 130 cm/s/s may have experienced NLSR. The DNL values generally increased with the PGAs, and shifts in predominant frequencies were recognized. The percentage shifts in peak frequencies in relation to the weak motions were greater than 40, reaching approximately 70 at sites where DNL values were 6 or larger and the PGVs were greater than 60 cm/s. A few sites showed gains in amplifications at lower frequencies in relation to the weak motions. Still, most sites showed only a decrease in amplifications at frequencies mostly higher than 1 Hz in the present analysis. The results from the S/B and H/V methods were generally comparable at the KiK-net sites, showing the applicability of the H/V method at the K-NET sites. The Vs30 values at the KiK-net sites with the top 5 DNL values (6~11) (they were also the top 5 PGA sites among the KiK-net stations located within 20 km of the source fault plane) were between about 300 and 700 m/s. The site conditions may be considered dense or stiff soil sites according to NEHRP site classifications. These results may imply that larger input motions can induce stronger NLSR even in dense soils. The number of K-NET sites analyzed in this study was about three times the number of KiK-net sites analyzed due to the difference in the station spacing between the two networks. A few K-NET stations with dense soil recorded atypical waveforms containing intermittent spikes with higher PGAs, most probably resulting from cyclic mobility. Pervasive local ground deformations from liquefaction have been reported in the Noto peninsula and beyond, resulting from earthquake shaking. The H/V analysis at the sites with high PGA spikes indicated noticeable reductions in higher-frequency spectral ratios and shifts in peak frequency. We also found that the ISK006 K-NET site with the largest PGA of approximately 2g (0.1-10 Hz) experienced relatively weaker NLSR despite its location near the fault plane and comparatively similar Vs10 value to other sites, requiring further analysis. A few stations with relatively lower PGAs during the 2024 mainshock gave larger DNL values using the H/V method, but it was found that the DNL values were not consistent with other similar PGA records at the sites or with the SB method. We discuss these various results in the present paper.
A comparison of the surface-to-borehole S-wave spectral ratios (S/B) for the horizontal components between the 2024 mainshock and other weak motion records was made at the KiK-net sites, where records were available at the surface and borehole. At the K-NET stations, where records are available only at the surface, a horizontal-to-vertical spectral ratios (H/V) technique was adopted. In this study, weak motion records are defined as those with peak ground accelerations (PGAs) smaller than 30 cm/s/s. Time windows of 20 seconds were used to compute the S-wave Fourier spectra. The degree of nonlinearity (DNL) was calculated using the method of Noguchi and Sasatani (2008). A DNL value of 2.5 and 4 or larger obtained using the S/B and H/V methods may indicate that the site experienced NLSR.
Our analysis based on the S/B method indicated that KiK-net sites with PGAs larger than about 130 cm/s/s may have experienced NLSR. The DNL values generally increased with the PGAs, and shifts in predominant frequencies were recognized. The percentage shifts in peak frequencies in relation to the weak motions were greater than 40, reaching approximately 70 at sites where DNL values were 6 or larger and the PGVs were greater than 60 cm/s. A few sites showed gains in amplifications at lower frequencies in relation to the weak motions. Still, most sites showed only a decrease in amplifications at frequencies mostly higher than 1 Hz in the present analysis. The results from the S/B and H/V methods were generally comparable at the KiK-net sites, showing the applicability of the H/V method at the K-NET sites. The Vs30 values at the KiK-net sites with the top 5 DNL values (6~11) (they were also the top 5 PGA sites among the KiK-net stations located within 20 km of the source fault plane) were between about 300 and 700 m/s. The site conditions may be considered dense or stiff soil sites according to NEHRP site classifications. These results may imply that larger input motions can induce stronger NLSR even in dense soils. The number of K-NET sites analyzed in this study was about three times the number of KiK-net sites analyzed due to the difference in the station spacing between the two networks. A few K-NET stations with dense soil recorded atypical waveforms containing intermittent spikes with higher PGAs, most probably resulting from cyclic mobility. Pervasive local ground deformations from liquefaction have been reported in the Noto peninsula and beyond, resulting from earthquake shaking. The H/V analysis at the sites with high PGA spikes indicated noticeable reductions in higher-frequency spectral ratios and shifts in peak frequency. We also found that the ISK006 K-NET site with the largest PGA of approximately 2g (0.1-10 Hz) experienced relatively weaker NLSR despite its location near the fault plane and comparatively similar Vs10 value to other sites, requiring further analysis. A few stations with relatively lower PGAs during the 2024 mainshock gave larger DNL values using the H/V method, but it was found that the DNL values were not consistent with other similar PGA records at the sites or with the SB method. We discuss these various results in the present paper.