UCLA

Traditional ergodic ground motion models (GMMs), typically conditioned on the time-averaged shear-wave velocity in the upper 30 m of soil (VS30), exhibit high site-to-site variability due to their inability to capture site-specific features such as resonances caused by impedance contrasts. Microtremor-based horizontal-to-vertical spectral ratios (mHVSR), which can contain peaks linked to site resonances, offer a useful supplement to VS30 for site response prediction. We use peak features from mHVSR spectra from 685 sites across California and neighboring regions (e.g., Nevada, Oregon) to develop mHVSR-conditioned site response models, which reduce site-to-site variability compared to a baseline linear site amplification model.