Application of continuous wavelet transform to optimize H/V ratio in seismic microzonation of Dargahan region (Qeshm, Iran)

Neotectonic and Seismotectonic activities have caused major seismicity in Iran. Dargahan city, located in the southeast of Iran, runs a high risk of earthquakes, hence its selection for the research. A large earthquake (Mw = 6.5 28Jun2006) was chosen for the study due to its provision of a suitable geological and velocity structure in this region. To calculate the amplification factor of the soil, it is indispensable to know its elastic properties, especially above the bedrock. One of the best ways to achieve this goal is to implement ambient noise. The spectral ratio between the horizontal and vertical components (H/V ratio) of microtremors measured at the ground surface was employed to estimate the resonance frequency of the site. The purpose of this paper was to retrieve the Rayleigh wave so as to estimate resonance frequency with higher accuracy and offer velocity structure for a single station record. The main problem with the ellipticity curve of Rayleigh wave retrieved from H/V ratio is the existence of the body and love waves in the waveform of the microtremor. The P-SV part of a waveform contribution is only in the vertical component while the SH part is in the horizontal component of the motion. If the SH-part is removed, the H/V ratios better indicate the ellipticity of the fundamental mode of Rayleigh wave, because the interest frequency band of the P-SV is dominated by the fandamental mode. In this paper, the SH removal is done by time-frequency analysis of all three components of the ambient noise. Continuous wavelet transform (CWT) was employed for this task by using the modified Morlet wavelet. The higher the value of the Morlet wavelet parameter (m) is, the narrower the wavelet in the spectral domain, and the better the frequency resolution will be, which leads to more accuracy of the H/V technique. In this article, first-time location of the most energetic time-frequency coefficient of the vertical component was identified and the corresponding local frequency spectrum was obtained. Then, a local horizontal spectrum at the same time location was calculated from an average time-frequency representation of the two horizontal components. Finally, the H/V ratio is calculated by dividing these two local spectra. We selected 12 stations in Dargahan which data (ambient noise) were recorded by CM6TD seismometers over a period of 30 minutes. The frequency-time analysis was done for all three components in the 12 stations. Primarily, Morlet parameter (m) was changed in the range (0.5-16), and the (H/V) curve was subsequently obtained according to this parameter, where the best value for the Morlet parameter was found to be 8. The results obtained for resonance frequency were compared with those of classic methods, showing a shift toward lower frequency in the proposed method. It was also shown that there existed two peaks in the reliable resonance frequency of the M03 station. So as to address the issue, the theoretical ellipticity curve of the Rayleigh wave was obtained from borehole data by the spectral ratio between horizontal and vertical components of its fundamental mode. The theoretical ellipticity curve of the Rayleigh wave and the log (H/V) curve were compared, upon which comparison we found out that the second peak was related to the resonance frequency in site M03. The values of the elliptic curve for near borehole stations which is in good agreement with the H/V curve, are suggested as the shear wave velocity structure for each single stations. By this method, one can estimate the shear wave structure for 8 stations close to the borehole.