微動とS波, P波, coda から求められる地盤特性の違いとその理論的解釈

抄録

We examine the differences of empirical site characteristics among S-wave, P-wave, coda, and microtremors and interpret them theoretically. Data used in this study are seismograms of 43 earthquakes and microtremor records observed at 20 strong motion stations within a 20km×30km region in Sendai, Japan. We calculate horizontal-to-vertical spectral ratios (H/V) for microtremors. Earthquake records are divided into P-wave, early P coda, S-wave, early S coda, and late S coda windows. For each window of seismograms we calculate H/V, horizontal spectral ratios (H/H), and vertical spectral ratios (V/V) with respect to a bedrock station where site effects are sufficiently small. The H/V for early P coda rapidly deviates from H/V for P-wave and quickly converges to H/V for microtremors. The H/V for early S coda gradually converges to H/V for microtremors in the frequency range lower than 3HZ as lapse time is longer. At soft sediment stations H/V for early S coda within 15sec after the S-wave arrival becomes identical with H/V for late S coda. On the contrary, at one rock station and several hard soil stations H/V for early S coda agree with H/V for S-wave. The H/H and V/V for early S coda are larger than those for S-wave at soft sediment stations. We also show that H/H for S-wave are different from H/V for microtremors and H/V for S-wave. If we use shallow S-wave velocity structures only, theoretical S-wave site response (H/H) and H/V for the fundamental mode of Rayleigh wave cannot explain H/V for microtremors except for some higher peak frequencies. When we use whole sediment structures on top of the seismological bedrock, theoretical H/V for the fundamental mode of Rayleigh wave agrees with H/V for microtremors and theoretical H/V for an obliquely incident SV-wave agrees well with H/V for S-wave. Theoretical S-wave site response based on the whole sediment structure agrees well with H/H for S-wave, but does not with H/V for microtremors. Our results mean that so-called Nakamura's technique based on H/V for microtremors is not valid in Sendai as a method to estimate empirical S-wave site responses. Only on the condition that the peak amplitudes of H/V for microtremors be greater than 3 and the peak frequencies be lower than 1Hz, the frequencies of the maximum peak in H/V for either microtremors or S-wave are similar to those in H/H for S-wave. Concerning coda we conclude that Rayleigh wave contamination in coda is significant in the frequency range lower than 3Hz at soft sediment stations.