GNGTS 2014 - Atti del 33° Convegno Nazionale

126 GNGTS 2014 S essione 2.2 already evidenced by seismic data recorded in the town (De Luca et al. , 2005; Milana et al. , 2011; Di Giulio et al. , 2014). It is interesting to highlight as the H/V features observed on the new and available noise data show a good correspondence with the H/V ratios computed on earthquakes (Milana et al. , 2011). Systematic differences on f 0 values emerge in terms of frequency, amplitude and shape of the H/V curves moving along each trench. In particular, from the bottom to the top of L’Aquila terrace the low-frequency peak shows an increase of its amplitude and a shift towards lower frequencies (Fig. 2b). Furthermore, moving towards the base of the hill the low-frequency peak ranges in a broad frequency band displaying a bump shape. This variation in the H/V shapes is mainly related to a different behavior of the vertical component. Indeed the vertical recordings show a spectral minimum shifted at higher frequency for measurements nearby the Aterno river (Di Giulio et al. , 2013). These H/V difference around f 0 between the top and the bottom of L’Aquila terrace are dubitatively explained in terms of 1D stratigraphic variation of thickness and seismic velocity between the soft soils and the stiff limestone bedrock. Moreover, the low-frequency resonance shows a significant directionality in the response at about N130°- N150° measured clockwise, as it can be observed through the oriented barrettes in Fig. 1. This polarization is parallel to the mean strike of the Aterno river valley, and perpendicular to the direction of the main elongation of L’Aquila terrace. According to Matsushima et al. (2014), the lateral heterogeneity related to the geomorphologic condition of L’Aquila hill could explain the strong polarization at f 0 on noise data. In terms of amplification, the polarization appears stronger at the top of the terrace and along the scarps than in the valley. In addition, H/V curves sometimes show a secondary resonance frequency f 1 , variable from about 5 to 9 Hz (in Fig. 1 f 1 , when observed, is represented, by circles into squares or barrettes). The high-frequency resonance peak can be related to the thin uppermost soft layer of residual soils which reaches the maximum thickness of 20-30 m �� ��� �������� ���� �� ��� in the southern part of the town and along its scarps ���� ������ (Del Monaco et al. , 2013). The layer characterized by low ����� ���� shear wave velocity ( V S ≈ 250-350 m/s) defines a strong impedance contrast ������� ��� ������� ��� ��� ����� between its surface and the stiff calcareous breccias �( V S ≈ 600-1200 m/s). Fig. 2 illustrates the results on Saint Apollonia trench, from the Aterno River valley to Via XX Settembre. A geolithological cross section, namely A-A’ (Fig. 2a), was constrained using the available geological, geotechnical and geophysical data: two deep boreholes, S3 to 195 m depth with a cross-hole test to 80 m, and S4 to 80 m depth with a seismic dilatometer test to 50 m; two shallow boreholes, S1 and S2; the geogravimetric map of L’Aquila city centre ������ (MS–AQ Working Group, 2010) ��� ��� ����� ������������ ��������� ��� ���� ���� ���� ���� ���� and new noise measurements performed for this work (R8, R13, R15, R16, R16b, R17, R18, see Fig. 2b). The first resonance frequency f 0 was used to confirm the depth of the geological and seismic bedrock (0.5-0.9 Hz), by means the deep impedance contrast between surface soils and stiff limestone. This result confirms the subsurface soil thickness already estimated by the geogravimetric map and found by S3 sounding. Instead, the secondary resonance frequency f 1 , when identified, was introduced to verify the thickness of the red soils (7-12 Hz), partially known from the surrounding boreholes. ��� �������� ��� �������� ���� For example, R13 detected this peak at about 10 Hz, which is related with the 7.5 m of residual deposits ( V S ≈ 300 m/s). The identification of this layer is very important considering the amplification effect due to this upper portion in L’Aquila terrace, as also confirmed by 1D ground response analysis (Amoroso et al. , 2014). Fig. 2b plots the single H/V curves with their standard deviation (dashed line) along Saint Apollonia trench, and an overlay of all these noise measurements. The trend of the results shows the considerations previously introduced, hypothesizing the possible sensitivity to a topographic effect. From Via XX Settembre to the Aterno river valley f 0 shows a decrease of H/ V amplitude (from 6 to 3) and a shift towards higher frequency (from 0.5 to 0.9 Hz). Moreover, moving towards the base of the hill the low-frequency peak ranges in a broad frequency band displaying a bump shape.