Seasonal variations in amplitudes and resonance frequencies of the HVSR amplification peaks linked to groundwater

Abstract

SUMMARY Following the installation of a temporary seismological network in western Greece north of the Gulf of Patras, we determined the quality of the sites of each of the 10 stations in the network. For this, we used the horizontal-to-vertical spectral ratio (HVSR) method and calculated an average curve over randomly selected days between 0 and 10 Hz. The daily HVSR curve is determined by the HVSR 12-hr calculation (1 hr every two) without distinction between seismic ambient noise and earthquake signal. The HVSR curves obtained can be classified in three categories: flat curves without amplification, curves with a amplification peaks covering a large frequency range, and curves with one or more narrow peaks. In this third category C3, one station has one peak, two have two and one has three. On the contrary of what it is commonly assumed, the amplitudes and the resonance frequencies of these narrow peaks are not stable over time in C3. We determined the maximum of the amplitude of each peak with the corresponding central frequency for each day during 2.5 yr. Except for the station with three peaks, which finally appears stable within the uncertainties, the principal peak exhibits a seasonal variation, with a maximum in winter and a minimum in summer, the observations being more dispersed during winter. The second peak, when it exists, varies in the same way except at one station where it varies oppositely. These variations are clearly correlated with the loading and unloading cycle of the underlying aquifers as shown by the comparison with water level and yield measurements from wells located close to the stations. Moreover, they are also correlated with the vertical surface displacements observed at continuously recording GPS stations. The dispersion of the observed maximum amplitude in winter is probably related to the rainfall and the soil moisture modifying the S-wave velocity as revealed by other studies. From this study, we would like to emphasize that the use the HVSR method to constrain the S-wave velocity and the thickness of the sediment layer over the bedrock in the basin, has to be done with caution. Upon further confirmation of its robustness, the HVSR methodology presented here could be a good and easy-to-use tool for a qualitative survey of the aquifer backdrop and its seasonal behaviour, and of the soil moisture conditions.