GNGTS 2018 - 37° Convegno Nazionale

392 GNGTS 2018 S essione 2.2 The visualization of the traces in the fig.1b shows a significant dispersion of the pulse with increasing depth. For this reason the RT method seems to be particularly suited for this dataset. The key-point of the SR method is the identification of the first cycle of the signal over which perform Fourier Transform. I’ve firstly carried out a pack of operation technically called conditioning operations. These operations are usefull to improve the quality of the seismograms in relation of our studies, and their related spectra. After stacking, the primary objective is to identify and isolate the part of the seismogram relative to the passage of the P-waves and eliminate the contribute of slower waves and multiples in seismogram. Another difference between a classic VSP data and a near-surface VSP is even the different processing and conditioning operations. The downgoing wavefield was extracted using a wave-by-wave separation approach considering the near-surface wavefield observed (only Downgoing). A 300 Hz lowpass filter is applied in order to decrease interference and noise at higher frequencies. Then we performed a windowing operation using a Blackman Window function. In the end, we applied a polynomial function in order to remove significant trends of each seismic signal. We operated an FFT to estimate amplitude spectra of each trace. The natural logarithms of the spectral ratios between the reference signal and the arrivals recorded at subsequent geophones were then computed and plotted as a function of frequency to estimate the slope of the logarithm spectral ratio, alpha, which, according to Equation (1-1), is a function of Q and the travel time between the two arrivals. We can see on Fig.2 (a-b-c) the reduction in the spectral amplitude due to the attenuation of the signal in depth. The slope was plotted and obtained over a frequency band of 15-150 Hz, when the spectrum shows the peak of energy. Fig. 2 - (a) Fourier spectra of seismic traces at 10-15-20 meters depth, (b) Spectrogram of 15 th trace. Data results. The key parameter used for Q P estimate for SR method is , the slope of the regression (Equations (1-1). The spectral ratios show a very good linear trend between 15 and 150 Hz, especially for deeper ratios. Once this threshold is exceeded, a drastic change in slope is noted (Fig.3a). The slope values obtained from these ranges are equal to the differential attenuation . As we expected, plotting the differential attenuation values vs depth, we see an increasing. Some deviating values, most in the first layers, can essentially be due to scattering phenomena. The same is observed in the RT values (Fig.3c), which increase in depth, and vary from 0.069 to 0.33 ms. The slope values ​are divided by a minimum of 3.165 x 10 -3 up to a maximum equal to 2.622 x 10 -2 recorded in the last meters. At this point it is possible to obtain, with reference to the formulas (1-1) and (1-2), the values ​of the quality factor for the compressional P-waves for both of methods (Fig.3d). For the Risetime method, 3 different intervals were considered in relation to layers observed before. The results of the tests, comparing the values ​obtained from the two methods, show an excellent correspondence and a minimum