2:55 PM - 3:15 PM
[O09-04] Strong ground motions and site amplifications in relation to the structural damage ratios during the 2024 Noto Hanto earthquake
★Invited Papers
Keywords:Strong ground motions, horizontal to vertical spectral ratio, basin response, generalized spectral inversion technique
During the 2024 Noto Hanto earthquake, strong-motion records were obtained at more than 20 strong-motion observation sites within the near-fault region. In particular, characteristic waveforms were obtained at strong-motion observation sites located in urban areas in the three municipalities of Suzu, Wajima, and Anamizu, where the rapid survey (Nakazawa et al., 2024) reported that more than 30% of the wooden houses in the surrounding area of these strong-motion sites were completely destroyed. We first illustrate the velocity and displacement waveforms obtained by integrating the observed acceleration waveforms without any filtering, and then describe the characteristics of the waveforms. The most remarkable point was the peak horizontal velocity (PGV) at the strong-motion stations in these severely damaged areas: 112 cm/s at Suzu, 122 cm/s at Wajima, and 159 cm/s at Anamizu. The predominant periods of excitations were all 1 s or longer, and the number of waves was not only one pulse but several pulses.
Next, we collected weak motion records from small to medium earthquakes with peak ground acceleration PGA< 200 cm/s2 before this earthquake and analyzed them using the Generalized Spectral Inversion Technique (GIT) to separate the site amplification characteristics of horizontal motion and vertical motions. The horizontal and vertical site amplification factors are named HSAF and VSAF*, respectively. GIT was performed first on the Fourier spectrum of the S-wave portion (5-15 s) and that of the entire waveform. The reference used was the horizontal spectrum of the S-wave part on the seismological bedrock, which was obtained by extracting the amplification by the weathered layers from the records at YMGH01. The reason why we use VSAF* instead of VSAF is that the same horizontal component of the bedrock spectra was used as a reference.
The figure shows the HSAF of K-NET station ISK002 in Suzu City. The red line is the HSAF of the whole wave (wHSAF), and the black line is that of the S-wave part (sHSAF). The amplification characteristics show that the wHSAF at ISK002 is very large, almost 100 times at 1 Hz, and that the wHSAF is larger than the sHSAF below 1 Hz. Usually wHSAF is larger than sHSAF due to the contribution of basin-induced surface waves. However, ISK002 is located near the coast in a narrow alluvial plain and is not inside a sedimentary basin. The vertical amplification at ISK002 is huge, about 30 times at 1 Hz in wVSAF* and 10 times in sVSAF*, which the usual P-wave amplification characteristics cannot explain.
A similar phenomenon was observed at JMA_E10, a JMA observation station in Wajima City. The amplification factors of wHSAF and sHSAF at 1 Hz are 90 and 40 times, respectively, and those of wVSAF* and sVSAF* are 10 and 6 times. The surface soft layers in Wajima City are only about 20 m deep according to the results of borehole investigations and so the ground amplification by the surface layer cannot explain the observed amplification at all.
As for the building damage, Nakazawa et al. (2024) investigated the observed damage ratios (total destruction rate) of wooden structures in the vicinity of strong-motion observation sites in the Noto Peninsula. In response to the heavy damage in Kobe during the 1995 Hyogo-ken Nanbu earthquake, we have developed a damage prediction model for wooden structures (Nagato & Kawase, 2002; Yoshida & others, 2004). Thus, we applied the model to the observed waveforms with high cut filtering at 1Hz~2Hz to obtain the calculated damage ratios. The results showed that the calculated damage ratios without age classification were 24~29% at ISK002, ISK003, or ISK005, corresponding to the survey results at these sites. In the model with chronological classification, the calculated damage ratio for buildings built before 1950 was 42%, while that for buildings built after 1981 was only 2.3% at ISK003, indicating that newer buildings had a higher probability of survival. Therefore, it is essential to use prediction models that consider the effect of construction age when predicting future earthquake damage.
Next, we collected weak motion records from small to medium earthquakes with peak ground acceleration PGA< 200 cm/s2 before this earthquake and analyzed them using the Generalized Spectral Inversion Technique (GIT) to separate the site amplification characteristics of horizontal motion and vertical motions. The horizontal and vertical site amplification factors are named HSAF and VSAF*, respectively. GIT was performed first on the Fourier spectrum of the S-wave portion (5-15 s) and that of the entire waveform. The reference used was the horizontal spectrum of the S-wave part on the seismological bedrock, which was obtained by extracting the amplification by the weathered layers from the records at YMGH01. The reason why we use VSAF* instead of VSAF is that the same horizontal component of the bedrock spectra was used as a reference.
The figure shows the HSAF of K-NET station ISK002 in Suzu City. The red line is the HSAF of the whole wave (wHSAF), and the black line is that of the S-wave part (sHSAF). The amplification characteristics show that the wHSAF at ISK002 is very large, almost 100 times at 1 Hz, and that the wHSAF is larger than the sHSAF below 1 Hz. Usually wHSAF is larger than sHSAF due to the contribution of basin-induced surface waves. However, ISK002 is located near the coast in a narrow alluvial plain and is not inside a sedimentary basin. The vertical amplification at ISK002 is huge, about 30 times at 1 Hz in wVSAF* and 10 times in sVSAF*, which the usual P-wave amplification characteristics cannot explain.
A similar phenomenon was observed at JMA_E10, a JMA observation station in Wajima City. The amplification factors of wHSAF and sHSAF at 1 Hz are 90 and 40 times, respectively, and those of wVSAF* and sVSAF* are 10 and 6 times. The surface soft layers in Wajima City are only about 20 m deep according to the results of borehole investigations and so the ground amplification by the surface layer cannot explain the observed amplification at all.
As for the building damage, Nakazawa et al. (2024) investigated the observed damage ratios (total destruction rate) of wooden structures in the vicinity of strong-motion observation sites in the Noto Peninsula. In response to the heavy damage in Kobe during the 1995 Hyogo-ken Nanbu earthquake, we have developed a damage prediction model for wooden structures (Nagato & Kawase, 2002; Yoshida & others, 2004). Thus, we applied the model to the observed waveforms with high cut filtering at 1Hz~2Hz to obtain the calculated damage ratios. The results showed that the calculated damage ratios without age classification were 24~29% at ISK002, ISK003, or ISK005, corresponding to the survey results at these sites. In the model with chronological classification, the calculated damage ratio for buildings built before 1950 was 42%, while that for buildings built after 1981 was only 2.3% at ISK003, indicating that newer buildings had a higher probability of survival. Therefore, it is essential to use prediction models that consider the effect of construction age when predicting future earthquake damage.