Multichannel analysis of surface waves

Editor's Notes

• #23: The Spectral analysis of surface waves method is the successor to the steady state Rayleigh wave method developed in the 1950’s. Much of the development of the modern SASW method was carried out at UT Austin in the early 1980’s. SASW testing is used to obtain a shear wave velocity profile It is non-invasive and non-destructive - testing is performed on the ground surface and strains are in the elastic range Instead of measuring shear wave velocity directly, Rayleigh wave velocities are measured and Vs is inferred.
• #24: The general testing setup is shown here. A seismic source generates surface waves, which are monitored by two in-line receivers. Both transient and continuous dynamic sources are used to generate surface waves, with the data usually cleaner from continuous swept-sine sources. A vibroseis truck (slide) was used for the long wavelengths and various hand-held hammers (slide) were used for the short wavelengths. Several factors must be considered in receiver geometry. To avoid near field effects associated with Rayleigh waves and body waves, the distance from the source to the receiver, d1, is at least half of the maximum recorded wavelength. Attenuation reduces signal quality if d1 is greater than 4-10 wavelengths, depending on the source. Therefore, an expanding receiver spread is used, with overlap between the wavelengths recorded in each setup. To minimize lateral variability, forward and reverse profiles are taken, usually with a common centerline. The time records from the two geophones are transformed to the frequency domain to generate the dispersion curve. The most important data are the phase of the cross-power spectrum and the coherence. It is important that the frequency domain calculation be done in the field so that the experiment can be modified as needed.
• #25: From the unwrapped phase of the cross power spectrum, the Rayleigh wave velocity is calculated, given the frequency and interreceiver distance. The dispersion curves from each receiver spacing are combined to generate the composite dispersion curve, which is representative of the site. Several theoretical solutions are used to model the dispersion curve: Fundamental mode Rayleigh waves only, and full stress wave solutions that incorporate higher modes of Rayleigh wave propagation, body wave energy, and receiver location. The full stress wave solution generally gives better results so the results from that model will be shown. The parameters in the layered earth model used to calculate the dispersion curve consist of layer thickness, shear wave velocity, Poisson’s ratio, and mass density. Usually only shear wave velocity and layer thickness are adjusted to match the dispersion curve, since they have the largest influence. The resolvable depth varies from about one half the longest wavelength to one fifth or less, depending on the site.
• #26: Obviously, the results from SASW testing were closer to those from downhole testing at some sites. There are several possible reasons for this; SASW testing samples a much larger volume of material than downhole testing. The SASW results are averaged across several hundred meters. If the subsurface varies laterally or is non-homogeneous the material sampled in SASW and downhole seismic testing may be different. At sites where the shear wave velocity increases gradually, the SASW data are easiest to interpret and most accurate. Large shear-wave velocity gradients limit the resolvable depth, because the dispersion curve never levels out to a velocity representative of individual layers or a half space. SASW testing does not resolve layer boundaries as well as average properties. Interface resolution also decreases with depth. The value of Poisson’s ratio assumed in the model for Rayleigh wave dispersion has a greater effect than is often thought. A sensitivity study showed that in saturated sediments.