GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 2.2 341 The slope affected by the landslide shows an articulated morphology. Examining Google Earth’s images of the area covering the last 10 years, one can note an upslope and lateral enlargement of the detachment niche. This suggests that the landslide dam deposit likely results from the overlapping of materials produced during different slope failure events. Three noise measurement campaigns were carried out (October 2017, May and October 2018), using a set of 3 tromographs. In each campaign, data were acquired keeping one sensor at a reference station, in continuous recording, and moving the others to 8 different stations for 30-40 minute recording sessions. The presence of a reference station made it possible to distinguish between signal differences that are due to site conditions at different stations, from variations that may reflect changes in the properties of the noise wave field sources at the time of the signal recording. The data were analyzed using the traditional Nakamura’s tecnhnique, HVNR (Nakamura, 1989), and the innovative technique HVIP (Del Gaudio, 2017). While the HVNR technique calculates the average spectral ratios between the horizontal and vertical component of noise recording, the new HVIP technique, by means of an instantaneous polarization analysis, identifies and extracts from the recording of ambient noise only Rayleigh wave packets. Thus, the ratios H/V between the amplitude of horizontal and vertical components of ground motion can be more reliably interpreted in terms of ellipticity of Rayleigh waves. Their variations as a function of frequency can then provide information on properties of subsoil material i.e.: i) layers thickness and velocity from the resonance frequencies, ii) impedance contrast between surface layer and bedrock from the H/V peak amplitudes, and iii) deposit mechanical anisotropy from directional variation of H/V amplitude. In comparison to the HVNR, the new technique proved to provide more stable results in the determination of H/V peak amplitudes and more details on site resonance frequencies. The main results of ambient noise analysis are summarized in Fig. 1. Starting from the north-eastern end of the profile, i.e. the reference station named YJG0, analysis of the data from the two 2018 campaigns reveals the presence of two main peaks at 3 or 4 Hz. In terms of amplitude, at the fundamental resonance frequency, the outcome of the HVIP analysis confirmed to be characterized by a greater stability than the HVNR method, which showed a much larger scattering of H/V peak values. This variability is likely related to the fact that the HVNR method analyses the entire acquired noise signal, which can be constituted, in variable proportions, by different types of waves differently excited, depending on changing environmental conditions in noise wavefield generation. Thus, while the smaller amplitude variations in HVIP results can be interpreted in terms of variations in site response properties (e.g. related to seasonal factor), the larger HVNR variations depend also on the variable proportion of different types of waves composing the recorded noise. The HVIP analysis revealed also the presence of a secondary peak at around 10 Hz, which can be related to the most superficial layer of the deposit. Analysing the data acquired in the following stations (named YJG1-8), the two adjoining peaks, already revealed in reference station, are found at frequencies that vary only slightly between 2 and 4 Hz, while the secondary peak at higher frequency shows larger variations between 6 and 13 Hz. Directivity at 45° seems to prevail in each station with the appearance, starting from the station YJG3, of a transverse directivity at about 135°. Going from the stationYJG1 to the stationYJG8, both peaks at lower (2-4 Hz) and higher (6- 13Hz) frequencies, which reflect the response of deeper andmore superficial layers, respectively, show changes in their directional properties, with local rotations of the direction of maximum H/V amplitude or the appearance of a more isotropic response (Fig. 1). Directional variations of ground motion amplification are often due to anisotropy of soil mechanical properties, which can induce larger amplifications along minor stiffness directions. Thus, the observed changes in peak directional characteristics can be related to the presence, within the different layers of the deposits, of fissuring systems whose direction is different as effect of deposit adaptation to topographic irregularities derived from the accumulation of previous landslides.