GNGTS 2014 - Atti del 33° Convegno Nazionale

GNGTS 2014 S essione 3.1 87 shifted upwards until the end of the line, at station 115. ����� ���� ������ ��� ������� Using this scheme the maximum nominal coverage was 24, further doubled in the processing phase applying an interleaving procedure. To exploit the polarization of SH waves, each shot point was acquired using both directions of energisation and in the processing lab these two records were appropriately summed to enhance the SH-waves while attenuating other unwanted phases in case they were recorded. To increase the signal to noise ratio 6 shots were done for each polarity (i.e. directions of energisation). In the processing lab, these 6 shots with NE-SW polarization were added together and the same operation has been done for the 6 shots belonging to the same station but with SW- NE polarization, after reversing the polarity of these traces. The two resulted shots were then summed together to enhance the SH-waves while attenuating other unwanted recorded phases. Figs. 2a and 2b show an example of a raw shot gather before the vertical stacking for each direction of energisation (NE-SW and SW-NE respectively) and Fig. 2c shows the resulting shot after vertical stacking and appropriate polarization summation. The SH events identified on the seismogram in Fig. 2c are: the direct waves in orange, the refracted waves in red and green and strong event with a characteristic dispersion trend corresponding to the Love waves. No reflections can be observed at this stage but they are evident after the de-noising procedure aimed at attenuating the various types of noise present. This is clearly shown in Fig. 2d which displays the same gather as in Fig. 2c at the end of the processing steps applied to reduce the noise. Taking advantage of the symmetry of the path trajectory in case of SH-SH reflection, for the processing of the SH wave data we can use the same software and algorithms developed for the P-P reflection seismic data, tailoring the parameters accordingly to the required needs. The main additional and different step required in the sequence consists in a calibrated difference between the records of the sources at the same location with opposite polarization, as described before. Besides the conventional steps of database building and cmp (common mid point) gathering (with interleaving), kill of anomalous traces, static corrections, band-pass filter, velocity analysis, normal move out, residual static application and stack, some dedicated operations are applied to remove the Love waves. They constitute a particular type of coherent noise that is characterized by high energy, dispersion behaviour, recording time and frequency band coincident with those of the desired signal. With an accurate design of directional (frequency- wavenumber) and eigenvalues filters applied on the shot domain, it was possible to attenuate these wave modes early in the processing, thus increasing the signal to noise ratio at the pre- stack level as shown in Fig. 2d. The resulting stack section was depth migrated by means of the Kirchhoff algorithm using as the velocity model a smoothed and Dix converted version of the stacking velocity field. The migrated seismic section exhibits a good signal to noise ratio from the very shallow layers up to 40 m in depth, with many continuous events displaying a trend in accordance to the expected geomorphological and geological setting. This image and the depth converted P-wave seismic section in Stucchi et al. (2014) are the ones used for a combined interpretation. Comparison between SH- and P-waves seismic sections. The main feature that induces to use the SH-wave reflection seismic compared to P-wave reflection seismic is the higher resolution that can be achieved with SH-waves at very shallow depths (Guy et al. , 2003; Guy, 2006). Accepting as threshold for the vertical resolution a quarter of the dominant wavelength (λ/4) (Sheriff and Geldart, 1982; Yilmaz, 2001; Sheriff, 2002), it is easy to deduce that SH seismic method benefits of a higher resolution than P seismic method. In fact, if we consider for the Patigno case a mean velocity of ~280 m/s and a dominant frequency of 25 Hz for SH waves, a resolution around 3 m results at the slip surface depth, compared to approximately 10 m for P-waves (using 2000 m/s and 50 Hz for velocity and dominant frequency respectively). Then the shorter wavelength provides a higher resolution for the SH-waves, but at the same time it causes an earlier signal attenuation because of increased absorption.