SITE CLASSIFICATION BASED ON SPECTRAL AMPLIFICATIONPATTERNS FOR MICROTREMOR H/V RATIOS

摘要

The authors apply soil response estimation techniques employing accelerograms for fifteen earthquakesrecorded at the Yokohama Strong Motion Network and its vertical array of nine sites, plus microtremordata recorded at all 150 sites. Assessment of the reliability of surface to a reference site- and horizontal tovertical- spectral ratios of S-waves, Coda, and microtremors, relies on cross-validation with both, surfaceto borehole spectral ratios, and theoretical soil response functions for vertically incident SH plane wavescalculated from the vertical array and logging data, respectively. The results indicate that the H/V ofmicrotremors provides a better match of soil response estimations when considering no more than ±30%mismatch in the determination of the predominant periods. The application of pattern recognition revealsa category structure that fits the mutual relationship of ground motion- and microtremor-basedestimations. The calculation of surface to borehole spectral ratios for the vertical components of groundmotion and microtremors, as well as the computation of theoretical H/V spectral ratios for SV plane wavesconfirms the robustness of the grouping scheme. The nine spectral amplification patterns composing thecategory structure thus validated serve as the cornerstone of our site classification method. The methodconstitutes an attempt at expanding the applicability of microtremor-based site effects estimations throughthe improvement of their reliability. To achieve it, we exploit the nexus between the nine amplificationpatterns and the corresponding local geological conditions revealed in the course of our study. Theprospect of obtaining more reliable site effects parameters from a relatively simple analysis at acomparatively low economic cost, combined with the possibility of recording without strict spatial or timerestrictions turns microtremors into a particularly appealing approach.