Near-surface cavity detection using ambient noise-based microseismic analysis in highly urbanized historical settings

This paper presents new results from the analysis of ambient microseismic data for the detection of near-surface cavities in highly urbanized historical settings. Specifically, the paper presents observations and analyses of the distribution of frequencies and amplitudes of standing waves generated by microtremors in the space between the ground surface and the underground cavity, obtained by averaging the amplitude spectra of microseismic records. This study clearly shows that the frequency of the fundamental resonance peak in the vertical component, if it corresponds to the frequency of a moderate- to high-amplitude (2.1-12) peak in the horizontal-to-vertical (H/V) ratio, is a good indicator of the presence of cavities. If the compressional wave velocity of the materials overlying the cavity is known, this frequency can be used to estimate the depth of such cavities. The opportunity to cross-validate microseismic data with ground-truth measurements allowed to determine an approximately 80 % success rate in identifying the cavities, a 1.37 % deviation in estimating the cavity depth, and a detectability depth ranging between 1.5 and 3.8 times the horizontal size of the cavity. Inaccurate knowledge of compressional wave velocity results in inaccurate definition of cavity depth. This study represents one of the first attempts to use ambient microseismic data to detect subsurface cavities. The results have clear implications for the use of passive seismic noise techniques in the detection and exploration of shallow subsurface cavities under non-planar layered models and non-ideal seismic noise source conditions.