Source depopulation potential and surface-wave tomography using a crosscorrelation method in a scattering medium

Abstract

We use seismic prospecting data on a 40  40 regular grid of sources and receivers deployed on a 1 km  1 km area to assess the feasibility and advantages of velocity analysis of the shallow subsurface by means of surface-wave tomography with Green's functions estimated from crosscorrelation. In a first application we measure Rayleigh-wave dispersion curves in a 1D equivalent medium. The assumption that the medium is laterally homogeneous allows using a simple projection scheme and averaging of crosscorrelation functions over the whole network. Because averaging suppresses noise, this method yields better signal-to-noise ratio than traditional active-source approaches, and the improvement can be estimated a priori from acquisition parameters. We find that high-quality dispersion curves can be obtained even when we reduce the number of active sources used as input for the correlations. Such source depopulation can achieve significant reduction in the cost of active source acquisition. In a second application we compare Rayleigh-wave group velocity tomography from raw and reconstructed data. We can demonstrate that the crosscorrelation approach yields group velocity maps that are similar to active source maps. Scattering has an importance here as it may enhance the crosscorrelation performance. We quantify the scattering properties of the medium using mean free path measurements from coherent and incoherent parts of the signal. We conclude that for first-order velocity analysis of the shallow subsurface, the use of crosscorrelation offers a cost-effective alternative to methods that rely exclusively on active sources.

Surface wave eikonal tomography in heterogeneous media using exploration data

“…This heterogeneity would complicate any tomography based on ray tracing, such as traditional or double‐difference tomography, and it is likely that in such applications much of the structural details inferred here would have been lost due to regularization. We note that the averaged slowness values at each frequency match the values found by Gouédard et al (2011, fig. 2) for a laterally homogeneous equivalent medium for this area.…”

“… Rayleigh wave phase‐velocity maps obtained from eikonal tomography at (a) 10 Hz; and (b) 20 Hz. (c) Rayleigh wave fundamental‐mode group velocity in the 10–25 Hz frequency band (with a dominant frequency around 12 Hz) obtained using traveltime tomography (from Gouédard et al 2011). …”

“…In Fig. 4(c), we show the group‐velocity model obtained by Gouédard et al (2011) from the same data. Gouédard et al (2011) used a traveltime tomography approach to produce a fundamental‐mode Rayleigh wave group‐velocity map (as opposed to phase velocity here) in a broader frequency band (10–25 Hz, but with a dominant frequency around 12 Hz) than considered here.…”

“…4(c), we show the group‐velocity model obtained by Gouédard et al (2011) from the same data. Gouédard et al (2011) used a traveltime tomography approach to produce a fundamental‐mode Rayleigh wave group‐velocity map (as opposed to phase velocity here) in a broader frequency band (10–25 Hz, but with a dominant frequency around 12 Hz) than considered here. Despite the differences in inversion method, frequency content and wave type, the results are consistent: while the tomography result is smoother due to regularization and the use of lower frequency data, the main structures of Figs 3(a), 4(a) and (b) are also present in Fig.…”

Multiple scattering from icequakes at Erebus volcano, Antarctica: Implications for imaging at glaciated volcanoes

“…As such, we directly recover both frequency-dependent scattering and absorption mean free paths, which further allow for constraints on the average size of primary scatterers within the volcanic edifice. This has, to our knowledge, only been previously demonstrated for numerical models, the human body, and for a densely gridded anthropogenic source experiment [De Rosny and Roux, 2001;Gouedard et al, 2011;Conti et al, 2004]. Broader application of this method in seismology would present new opportunities for constraining otherwise elusive scattering parameters.…”