Continuous H/V Spectral Ratio Analysis of Ambient Noise Recorded by Stationary Seismic Stations to Improve Microzonation Results Obtained by Mobile Stations

Estimating the resonance frequency and amplification factor of unconsolidated sediments by H/V spectral ratio (HVSR) analysis of seismic ambient noise has been widely used since Nakamura's proposal in 1989. The fundamental frequency (f0) usually correlates well with the thickness of unconsolidated sediments above the bedrock. To measure f0 properly, Nakamura suggested to perform microzonation surveys at night when the artificial microtremor is small and does not fully disrupt the ambient seismic noise. As nightly fieldwork is not always a reasonable demand, we propose an alternative workflow of Nakamura's technique to improve the quality of HVSR results obtained by ambient noise measurements of mobile stations during the day. This new workflow includes the automated H/V calculation of one year of continuous seismic data of a stationary/permanent station located nearby the sites selected for microzonation. By means of an automated python script, the daily, weekly, monthly and seasonally variations of the fundamental frequency and the H/V amplitude at the site where the stationary station is installed are evaluated. Continuous HVSR analysis of sites with constant bedrock depth shows that the changes in the determined f0 and H/V amplitude are dominantly caused by the human behaviour which is stored in the ambient seismic noise (e.g. later onset of traffic in a weekend, quiet Sundays, differences between daily/nightly activity,…). This continuous analysis allows the characterisation of the deviation of the measured f0 to the true f0 throughout the whole year! Consequently, as mobile stations are affected by the same variation of the ambient noise, a correction factor can be applied on the calculated f0 of individual measurements during the microzonation survey and a proper Vs can be estimated. In this presentation, we apply this workflow to two different case studies; i.e. a rural site with a shallow bedrock depth of 30 m and an urban site (Brussels, capital of Belgium, bedrock depth of 110 m) where human activity is continuous 24h/day. Our results show that microzonation with mobile stations should always be accompanied by a stationary seismic station to control the error.