GNGTS 2015 - Atti del 34° Convegno Nazionale

the near-field hydrophone. These data were transmitted by a radio-link (RTS-100) to a dedicated onshore seismograph. We used three Seismic Source Daq Link III seismographs (two for the 48 hydrophones, the third one to record the near-field signature and GPS PPS signal) with 0.125 ms sampling rate and 5 s data length. Underwater seismic refraction data have a good signal to noise ratio and the first arrivals are clearly detectable also a larger offsets (Fig. 2a). ������ ���� ���������� ��� ������� �� ���� Travel time tomography was applied to each seismic line to obtain the corresponding 2D depth velocity section and to estimate the depth of the sediment/Flysch interface. We invert the first arrivals picked in each seismic line. First, we separated two different arrivals in the picked travel times for each common shot gather: the direct arrivals and the refracted arrivals from an horizon below the sea bottom, interpreted as the top of Flysch. For this purpose, we applied a procedure which automatically identifies the knee point (crossover distance) separating the two different arrivals. Then, we inverted the refracted arrivals by using an inversion approach based on minimum dispersion of the refracted points (Carrion et al. , 1993): depth and geometry of the refracted horizon are defined by an iterative procedure which minimized the difference between the refracted points and the estimated surface. The same procedure reconstructed also the lateral velocity gradient of the layer below the refracted interface. In the inversion we constrained the sediment velocity (1610 m/s) associated to the first layer below sea bottom. Fig. 2 – Underwater refraction seismic data: a) three common-shot-gathers recorded along the line 2 (perpendicular to the TMT front) and b) P wave velocity model obtained by the first arrivals tomographic inversion (line 2). GNGTS 2015 S essione 3.2 67