Time-dependent horizontal-to-vertical spectral ratio (HVSR) analysis of ambient seismic noise

Abstract

Existing local research has employed the horizontal-to-vertical spectral ratio (HVSR) for static microzonation and long-term groundwater monitoring. Building on this work, this thesis provides the first high-resolution quantification of how sub-daily meteorological fluctuations, specifically wind-driven noise and atmospheric variables, directly modulate seismic site response parameters across the stations of the Malta Seismic Network in the Maltese archipelago. By establishing these correlations, this study enhances the methodological accuracy of HVSR in the Maltese context, providing a framework to isolate meteorological noise from geological signals and enabling more reliable site response assessments for local engineering and microzonation projects. This study analyses four years (2017-2020) of continuous seismic data recorded by the Malta Seismic Network to investigate temporal variability in HVSR parameters. The research evaluates how HVSR peak frequency and amplitude evolve and examines the potential influence of meteorological variables, including wind speed, precipitation, temperature, and relative humidity. Statistical analysis is used to identify stations that exhibit sensitivity to environmental forcing and to assess the extent to which such factors may influence the stability and interpretation of HVSR measurements in the Maltese Islands. Forward modelling was also used to confirm whether the observed meteorological effects can potentially create subsurface parameter changes, such as variations in soil moisture or density, which could affect the HVSR measurements. This thesis further confirms that HVSR parameters are sensitive to wind strength and reveals that two stations (MELT and CBH9) have a heightened sensitivity to wind speed compared to other stations within the Malta Seismic Network. Hence, caution should be exercised when analysing data from these stations during windy days. Association between HVSR and other meteorological parameters, including precipitation, revealed weak effect sizes, indicating that other short-term meteorological forcing is not registered by HVSR.