The empirical power law relation (PR) between resonance frequency (f0), obtained from H/V spectral ratio analysis of ambient noise, and sediment thickness (h), obtained from boreholes, is frequently used in microzonation studies to predict bedrock depth. In this study, we demonstrate (i) how to optimally construct a PR by including the error on the picked f0 in the regression, and (ii) how to evaluate a regression quality by identifying the under- or overestimation of the sediment thickness prediction. We apply this methodology on f0 data derived from 74 ambient noise recordings acquired above boreholes that reach the Brabant Massif bedrock below Brussels (Belgium). Separating the f0 data into different subset based on the cover geology does not significantly improve the bedrock depth prediction because the cover geology in Brussels has common base layers. In Brussels, the PR relation h = 88.631.f0−1.683 is the best candidate to convert f0 to depth, with a prediction error of 10%. The Brussels PR was subsequently applied on a local survey (404 measurements; 25 km2) in southern Brussels with the aim to study Brussels’ Brabant Massif bedrock paleorelief. By linking the obtained paleorelief, Bouguer gravity data and aeromagnetic data, a NNW-SSE oriented, 20 m-high subsurface ridge could be identified. This ridge stands out because of differential erosion between less-resistant and hard quartzitic rock formations of the Brabant Massif. This subsurface ridge deflects the local radiation of seismic energy resulting in an anomaly in the otherwise regional consistent azimuthal dependency of the resonance frequency. We conclude that adding a polarisation analysis to a microzonation survey analysis allows detecting anomalous features in the paleorelief.
Located in
Library
/
RBINS Staff Publications 2021
