GNGTS 2017 - 36° Convegno Nazionale

690 GNGTS 2017 S essione 3.3 Free triangular mesh was built for the whole model, a mesh refinement window of 200x50 m was applied around and below the synthetic array in order to respect a maximum element size suitable for the propagating wavelengths in both materials. Fig. 1 - a) 2D finite-element model used for the generation of the synthetic seismograms. b) Zoom on the upper central part of the model (black rectangle, in a) with the synthetic array configuration, source locations and discontinuity geometry. Tab. 1 - Physical and mechanical parameters adopted in the simulations for materials outside (1) and inside (2) the subsurface heterogeneity. ρ [kg/m 3 ] E [MPa] ν [-] V P [m/s] V S [m/s] V R [m/s] Q factor [-] Material 1 2200 180 0.3 330 175 162 20 Material 2 1900 90 0.3 200 110 102 15 A Ricker wavelet centered at 50 Hz was chosen as seismic input for the model, in seven source positions, located at the ends and within the array, with a step of 12 geophones (S1-S7, in Fig. 1b). Synthetic seismograms were generated for each source location. After recovering geometrical spreading in the seismograms, the total energy of each trace was calculated as the sum of the squared spectral amplitudes in the 5-75 Hz band and then normalized to the maximum of each seismogram. The energy decay exponent was calculated as the local slope of the energy- distance plot in a bilogarithmic scale, shifting a moving window along the seismic line. For each window position, calculated γ were averaged from all the available shots, distinguishing between positive and negative offsets. The standard deviation on each window was retrieved as well to consider estimation uncertainties. Using the same moving windows along the energy- distance plot, the attenuation coefficient of each frequency component was obtained as the local slope of the energy f -distance plot in a double natural logarithmic scale. For each window