GNGTS 2015 - Atti del 34° Convegno Nazionale

GNGTS 2015 S essione 3.2 83 cost, non-invasive, complementary strategy would be the use of methods based on surface waves (SW), such as SASW (Nazarian and Stokoe, 1984) or MASW (Park et al. , 1999; Socco and Strobbia, 2004). In the classic SASW and MASW, Rayleigh (and alternatively Love) SW are excited by an active source and recorded by a linear array of receivers deployed on the ground. The dispersive nature of the recorded surface waves is used to derive the vertical (1D) subsurface profile of the shear velocity V S by an inversion procedure. In detail, the propagation of SW along the array allows for the construction of a dispersion pattern, which is retrieved by transforming the experimental seismograms from the time-space to a more suitable domain by using a specific numerical transform. For example, in the MASW framework, the frequency- wavenumber (f-k) and the frequency-Rayleigh (f-V R ) domains are frequently used and also occasionally the τ-p transform (McMechan and Yedlin, 1981). The obtained dispersion pattern is then an entire portion of a two parameters domain and in order to capture the dispersion pattern the spectral maxima are picked to yield the so-called dispersion curve. Since SW are multimodal, this approach is capable of separating multiple phase velocity values at the same frequency. In the SASW framework the dispersion pattern is retrieved by calculating the cross spectrum of experimental seismograms recorded at two, opportunely spaced, receivers. The dispersion pattern, in this case, is represented as a cloud of points in the frequency-velocity domain (f,V R ). Unfortunately, by this approach an unique value of the velocity is obtained for each frequency, so that when multiple modes contribute to the real propagation these are not identified as separated and misleadingly used as a whole “apparent” propagation mode during the inversion. Despite these differences however, the dispersion pattern is inverted to estimate shear wave velocity distribution. Unfortunately, available inversion algorithms assume the subsurface model as a stack of homogeneous parallel layers, hence capturing only vertical variations of the subsurface elastic properties (e.g. Aki, 2002; Kausel and Roesset, 1981). Consequently, these algorithms are of limited use when lateral heterogeneities are known to exist. Indeed, there is a growing interest toward applications of the MASW technique for 2-D and 3-D subsurface imaging (Boiero and Socco, 2010; Vignoli et al. , 2011, 2015; Bignardi et al. 2012, 2014; Masoni, 2014; Socco et al. 2014, 2015). Such interest points out how the laterally heterogeneity identification is of primary interest in the near-surface investigations. Dealing with this issue, Bignardi et al. (2014) showed that in a MASW survey, the presence of a moderate lateral heterogeneity can be detected in the f-Offset domain while its effects are difficult to recognize when data are transformed in the f-V R domain; i.e. in this domain the information of the “locality” is lost. This leads to the consideration that lateral heterogeneity could be retrieved by separately elaborating the signals recorded at pairs of receivers in a similar way as it is done in the SASW technique. In what follows we shall use part of the SASW workflow to establish a strategy that can be used both as a feasible inversion strategy or alternatively for the direct interpretation of active- source datasets. Following the second course, we shall show that the Direct Interpretation of Phase Lags (DIPL) (Bignardi et al. 2015a, 2015b), which uses the frequency-dependent phase lags among pairs of seismic signals, is capable of retrieving a satisfactory 2-D and 3-D V S subsurface image also in complex subsurface environments without the need of inversion. We shall discuss the workflow, the pros and cons of 2-D and 3-D applications. Finally, we shall present and discuss field examples. Method. Let’s consider a source S and two receivers, R 1 and R 2 placed along a line; a SASW-like data processing concerns calculating the phase of the cross spectrum between the signals recorded at the receiver pair. This information can be used to get the local dispersion pattern as a cloud of points (COP) in the f-V R domain (see e.g. Fig. 1) through Eq. 1, [please refer to Nazarian’s and Stokoe’s (1984) paper for the details]. (1) and