GNGTS 2014 - Atti del 33° Convegno Nazionale

and central part, overlapping a low Q s anomaly, and a larger low Q S anomaly in the eastern part (see Fig. 2a). We added to the simulated k values a random Gaussian noise with sigma equal to 2.5% of the data value, like for the checkerboard test. The test shows that the simulated anomalies are generally well resolved (see Fig. 2b), although, like in Schurr et al. (2003), amplitudes recovery in general is reduced as a consequence of the damping; in particular the best resolved region is the one in the eastern part, due to the large illumination the area. It is important to remark that these tests are not able to take into account the reliability of the data on V S , that is taken as input and therefore can bias the results. For this reason, we took into account the results on the inversion only for depth lower than 12 km, where we had better reliability of V S data (Bressan et al. , 2012) Results. Fig. 3 shows the sections 4 and 5 of the tomographic 3D reconstruction of Q S for depth <12 km. The earthquakes are indicated as circles whose size corresponds to the magnitude. For the generic n th section characterized by y coordinate equal to y n earthquakes corresponding to y in the range [( y n-1 + y n )/2 , ( y n + y n 1 )/2] are projected. Fig. 3 – Vertical cross sections 4 and 5 (see Fig. 4) of the 3D tomographic Q S image. Circles represent earthquakes. Circle size is proportional to earthquakes magnitude. Q S values are characterized by marked lateral and depth variations. Another study in the same area (Bressan et al., 2012) on sequential integrated inversion of tomographic images and gravity data, which provided the 3-D V P , V P /V S and elastic moduli pattern, found � ���������� a structural heterogeneity that was associated to the variations of the mechanical properties of rocks� �� . At large scale, we recognized three main zones ���� ��������� ����������� ������� ������� �� ��� with different attenuation pattern related to the geologic-structural heterogeneity of the upper crust. The low heat flow and the variable pore pressure can also play a role in low Q S values, but no exhaustive data on crustal fluids are avail- able for the investigated area. GNGTS 2014 S essione 1.2 149