GNGTS 2017 - 36° Convegno Nazionale

692 GNGTS 2017 S essione 3.3 position, obtained values were averaged and normalized to the related standard deviation (α RP and α RN for positive and negative offsets respectively). This normalization enabled to compare attenuation coefficients obtained for different frequencies and to identify abrupt variations. To obtain a single imaging for both positive and negative offsets, the absolute values of α RP and α RN were stacked in a single plot, in which maximum values are expected to highlight the strongest contrasts. The autospectral density of each seismogram was computed as the sum of the squared real and imaginary parts of each-trace Fourier Spectrum, after recovering geometrical spreading on the seismograms. The resulting autospectrum plots of all the shots were then stacked together to reduce the influence of the source location and highlight the effects of the heterogeneity. T/I spectral ratios were computed considering all the possible pairs between 14 geophones placed after and before the sharp lateral variations (for a total of 196 combinations) in three different T/I locations: inside/outside left the box [(G30…G43)/(G16…G29)], outside right /inside the box [(G44…G57)/(G30…G43)], outside right /outside left the box [(G44…G57)/(G16…G29)]. Results. The results of the five tested SW based methods are summarized in Fig. 2. Given the model geometry and mechanical properties, abrupt variations are expected at the box edges, at distances of approximately 14.25 m and 21.25 m. Dealing with depth estimation, considering an average V R of 150 m/s in the first 10-m of the model and a known depth of 2 m for the box, a cut-off frequency around 23 Hz [f=V R /(3z)] is expected. Theoretical location and cut- off frequency are highlighted in Fig. 2 with back and red dashed lines respectively. All the normalized energy curves (Fig. 2a) show high energy concentration within the box. Particularly, the shot located on the anomaly (S4) highlights sharp narrow energy concentrations at the box edges. Around these locations, the energy decay exponent (Fig. 2b) has marked variations from zero, both for positive and negative offsets. In particular, the real box edges are located between the minimum and maximum of the opposite offset curves. A clear discontinuity location is also obtained from the stacked plot of attenuation coefficient variations (Fig. 2c) and autospectra (Fig. 2d). A clear frequency cut-off is also found in Fig. 2c and Fig. 2d, approximately around the theoretical one. The average T/I spectral ratios (Fig. 2e) confirm the same frequency cut-off, with a local minimum for the T/I ratios related to traces located either across the right box edge or after and before the box. Conversely, while computing the T/I spectral ratios of traces located across the left box edge, given the energy concentration and the high-frequency trapping within the heterogeneity (thus more marked in the transmitted waveforms) a maximum is found at the same cut-off frequency. Conclusions. Several SW based techniques for location and depth estimation of subsurface sharp lateral heterogeneities are available in literature. Since all these techniques were originally developed for single-fold configurations, the final data interpretation often resulted unclear and strongly dependent on the a-priori knowledge of the discontinuity location. Adapting these procedure to multifold data and improving the multi-fold already-existing techniques, towards a single imaging of the subsurface investigated by the array, has the advantage to strengthen the effects due to the discontinuity presence and to consequently improve reliability and interpretation of the results. In the present work, this was demonstrated on a simple synthetic study, but ongoing tests suggest the validity of the presented approach on real case studies, for both rock (i.e. detection and depth estimation of the open joints in a fractured rock mass) and soil mechanics issues (i.e. detection of loose formations in the overburden). Further synthetic tests will help to better constrain the sensitivity and uncertainties of each method to varying heterogeneity depths and to buried sharp lateral variations, not intersecting the free ground surface. References Bergamo P. and Socco L.V.; 2014: Detection of sharp later discontinuities through the analysis of surface-wave propagation . Geophysics, 79 (4), EN77-EN90. Bièvre G., Jongmans D., Winiarski T. and Zumbo V.; 2012: Application of geophysical measurements for assessing the role of fissures in water infiltration within a clay landslide (Trieves area, French Alps) . Hydrol. Process, 26 , 2128-2142.