GNGTS 2021 - Atti del 39° Convegno Nazionale

251 GNGTS 2021 S essione 2.2 The three sites described in this paper, CHC2 (lat=38.6817,  lon=16.4106), MNGA (lat=38.5174, lon=16.3222) and SCAI (lat=38.5343, lon=16.5177), are different among them with regards to the local geology, topographic features, and the resonance frequency of the HVSR peak. HVSR curve of CHC2 site is characterized by a well-defined peak at 4.0 Hz with average amplitude of 6 during the day that decreases to about 3 during night hours (Figs. 1a and 3a). All earthquakes analyzed at this site show a clear resonance peak at the same frequency but sharper and higher (always >7) than that observed for noise recordings (Fig. 3d). Similar results have been obtained at MNGA site. This site is characterized by a well-defined resonance peak at a frequency of 4.8 Hz reaching an amplitude greater than 5 during diurnal hours, but almost unidentifiable during the night (Figs. 1b and 3b). This is one of the most interesting results of our study. HVSR computed on local earthquakes shows a well defined peak at the same frequency and with a similar amplitude compared with that computed on diurnal recordings of seismic noise (Fig. 3e). Finally, results for SCAI site slightly differ from the previous two. In fact, here HVSR curves are characterized by values greater than 1 in a large frequency band with a broad peak at about 2.1±0.3 Hz reaching an amplitude of 3 during day hours and 2 during the night (Figs. 1c, 2e and 3c). Although the cyclic variation in the spectral ratio is not so remarkable as seen for MNGA and CHC2, we consider this result worth of attention because very stable in all time periods analyzed (many months of continuous signals). The results of amplitude and polarization analysis for 10 days of signals at this site are shown in Figure 2. The higher values of rectilinearity (plot b) are observed during day hours, in correspondence of higher RMS (plot a) and higher H/V (plot e). These features of the ground motion suggest a greater amount of body waves in the background noise during daily hours. About the analysis of earthquakes at this site, we see HVSR curves with main peak and average values very similar to those obtained from the analysis of diurnal seismic noise (Fig. 3f). Discussion and Conclusions For the selected sites we observe clear variations in HVSR curves well correlated with the day- night cycle. This is particularly evident for MNGA and SCAI sites, where the amplitude of the main HVSR peak is often smaller than 2 during night hours (Fig. 3b and 3c). This means that at these sites the estimation of site effects based on the HVSR analysis applied to a short recording of night hours seismic noise could lead to a misleading result and wrong conclusion. Our results suggest that those variations depend on the contribution of body waves on wavefield composition. This hypothesis is supported by the higher rectilinearity of the seismic noise observed during day Fig. 3 - Day and night HVSR (top plots) and earthquake HVSRs for three sites.