ABSTRACT
The microtremor horizontal‐to‐vertical spectral ratio (MHVSR) has been widely used to evaluate site effects due to its simplicity, high efficiency, and low cost. Commonly, MHVSR is calculated based on the Fourier amplitude spectra (FAS) of microtremor records, which requires a smoothing procedure to clearly observe the overall shape of the curves. However, MHVSR—particularly its amplitude—can be affected by the choice of smoothing technique. A previous study has recommended using response spectra instead of FAS for the MHVSR evaluation to eliminate efforts for smoothing and to obtain unique MHVSR results, but the applicability has not been systematically explored. In this study, a theoretical analysis of the response‐spectra‐based MHVSR (RMHVSR) is first conducted based on random vibration theory. In addition, the RMHVSR is explored based on microtremor records observed at 104 sites in Japan and site borehole data. Specifically, MHVSRs calculated from response spectra are systematically compared with those derived from FAS. It is found that the MHVSR based on response spectra is less sensitive to long‐period signals compared with that based on FAS, making it less susceptible to long‐period noise, such as that caused by wind. When the site fundamental period is short, the shapes of the two MHVSRs are highly similar, with their predominant periods aligning, though the peak amplitude based on response spectra is lower. In this case, RMHVSR is effective. However, when the site fundamental period is longer than around 0.8 s, the peak corresponding to the site fundamental period may be smoothed out when using response spectra, leading to inaccurate evaluations.
