Surface-wave analysis for building near-surface velocity models — Established approaches and new perspectives Laura Valentina Socco 1 , Sebastiano Foti 1 , and Daniele Boiero 1 ABSTRACT Today, surface-wave analysis is widely adopted for build- ing near-surface S-wave velocity models. The surface-wave method is under continuous and rapid evolution, also thanks to the lively scientific debate among different disciplines, and interest in the technique has increased significantly during the last decade. A comprehensive review of the literature in the main scientific journals provides historical perspective, methodological issues, applications, and most-promising re- cent approaches. Higher modes in the inversion and retrieval of lateral variations are dealt with in great detail, and the cur- rent scientific debate on these topics is reported. A best-prac- tices guideline is also outlined. INTRODUCTION Since Lord Rayleigh first predicted their existence in 1885 Ray- leigh, 1885, surface waves have attracted the interest of a constantly increasing number of researchers from different disciplines, includ- ing solid-state physics, microwave engineering, geotechnical engi- neering, nondestructive testing, seismology, geophysics, material science, and ultrasonic acoustics. Despite marked differences in scales and methods, these disciplines share the goal of exploiting the surface waves that propagate along the boundary of a domain to ob- tain information on one or more scalar fields inside that domain. Surface waves are interesting because they can be used to develop noninvasive techniques for characterizing a medium at a small scale e.g., engineers use ultrasonic surface waves to identify material de- fects, at a large scale e.g., seismologists use surface waves to inves- tigate the structure of the earth’s crust and upper mantle, and at an intermediate scale e.g., geophysicists and geotechnical engineers use surface waves to characterize near-surface geomaterials. All of these applications share the same principles: they use the geometric dispersion of surface waves to infer the properties of the medium by identifying the model parameters. In near-surface applications, most surface-wave tests estimate the shear-wave velocity profile. This is usually accomplished by adopt- ing a strategy based on estimating the experimental dispersion curve from field data and subsequently solving an inverse problem. This latter step implies the choice of a reference model for the interpreta- tion, which in most cases is a stack of homogeneous linear elastic layers. Surface-wave analysis is usually performed using Rayleigh waves because they are easy to generate and detect on the ground surface; however, Love waves, Scholte waves, and other kinds of guided waves that may be generated in specific stratigraphic condi- tions can also be analyzed. Regardless of the type of surface wave used, the standard proce- dure for surface-wave analysis can be divided into three main steps: 1 acquire the experimental data 2 process the signal to obtain the experimental dispersion curve 3 solve the inverse problem to estimate model parameters Each step can be performed using different approaches, according to the scale of the problem, the target, the complexity of the subsoil property distribution, and the available equipment and budgets. For applications on an engineering scale, the acquisition is conducted with a multichannel layout of vertical low-frequency geophones and an impact source in an off-end configuration. The processing is per- formed with automatic picking of the frequency/wavenumber f -k or frequency/slowness - p spectral maxima, which are then trans- formed to the dispersion curve. The inverse problem is usually solved with linearized algorithms that use a 1D forward model and yield a 1D S-wave velocity profile. This basic scheme is used extensively and is suitable for many near-surface applications. Alternatively, the analysis may be per- formed adopting the full-waveform inversion approach, in which extracting the dispersion curves is unnecessary. Nevertheless, the full-waveform approach requires a realistic simulation of the dy- namics of the propagation that accounts for source, attenuation phe- nomena, and soil-receiver coupling and requires complex computa- tional approaches. For these reasons, the full-waveform approach is seldom applied. Manuscript received by the Editor 13 January 2010; revised manuscript received 14 April 2010; published online 14 September 2010. 1 Politecnico di Torino, Torino, Italy. E-mail: valentina.socco@polito.it; sebastiano.foti@polito.it; daniele.boiero@polito.it. © 2010 Society of Exploration Geophysicists. All rights reserved. GEOPHYSICS, VOL. 75, NO. 5 SEPTEMBER-OCTOBER 2010; P. 75A83–75A102, 18 FIGS., 1 TABLE. 10.1190/1.3479491 75A83 Downloaded 15 Sep 2010 to 130.192.28.138. Redistribution subject to SEG license or copyright; see Terms of Use at http://segdl.org/