Abstract
Spectral ratios between soft soil and reference rock sites are often used to predict the sedimentary site response to earthquakes. However, their relationship with the genuine site-specific amplification function is often unclear. We compare the soil-to-rock spectral ratios between the stations that are 3.3 km apart with the “genuine” response given by the ratios between the surface and 17 and 47 m downhole. Data from the SMART1 array in Taiwan are used. The “weak” and “strong” motion records are addressed separately to allow for nonlinear soil response. The soil-to-rock spectral ratios are nearly identical to the “true” amplification at the frequencies from 1 to 10 Hz, if the finite depth of the borehole is taken into account. They correctly capture the strong-motion deamplification effect. However, the soil-to-rock spectral ratios are roughly 1.4 times more uncertain than surface-to-47-m ratios. In summary, the soil-to-rock spectral ratios can be considered as the reliable estimates of the real site response.
Calibration of amplification factors using KiK-net strong-motion records: toward site effective estimation of seismic intensities
