Further investigation over Quaternary silts using the Spatial Autocorrelation (SPAC) and Horizontal to Vertical Spectral Ratio (HVSR) microtremor methods

Abstract

This study extends a previous investigation using the passive seismic Spatial Autocorrelation (SPAC) and Horizontal to Vertical Spectral Ratio (HVSR) methods on microtremor observations in the Port Melbourne area. The area has up to 80 m thickness of Quaternary gravels, sands, and silts, with the near-surface Coode Island Silt being so soft that the base of the unit could not be resolved in a previous study. In order to extend the previous work, further surveys using instrumentation more suited to low-frequency measurements (SPAC and HVSR methods) and a larger array radius (SPAC method) were undertaken.Despite the use of significantly larger array radii and long-period seismometers, coherencies in the sub 1–2 Hz range were poor and allowed only marginally deeper shear-velocity information to be obtained. We have attributed this deficiency in the SPAC spectrum to a lack of energy in the vertical component of Rayleigh wave motion, as indicated by a high-amplitude peak in the HVSR spectrum around 1–2 Hz.Despite the poor low-frequency SPAC results, a combined modelling approach using both HVSR and SPAC results, along with the results from the large array survey, allowed deeper shear-velocity values (to depths of 70–100 m) to be estimated with greater precision than the initial surveys. We note that the HVSR spectrum is more sensitive to variations in shear velocities at these depths than the SPAC spectrum, providing an additional constraint on interpreted shear-velocity profiles from SPAC data.Lack of a secondary HVSR peak below 1 Hz that could be attributed to the sediment-bedrock interface leads us to believe that this expected peak is masked by the high amplitude peak resulting from the large shear-velocity contrast closer to the surface in the sediment profile (a silt–gravel interface).