Statistical multioffset phase analysis for surface-wave processing in laterally varying media

Standard procedures for dispersion analysis of surface waves use multichannel wavefield transforms. By using several receivers, such procedures integrate the information along the entire acquisition array. That approach improves data quality and robustness significantly, but its side effects are spatial averaging and loss of lateral resolution. Recently, a new approach was developed to address that issue and maximize lateral resolution. The new method uses multioffset phase analysis to detect and locate sharp lateral variations in velocity. By using the phase analysis approach, the number of usable channels can be maximized, thereby gaining data quality without compromising lateral resolution. In fact, such preliminary data analysis also allows selection of the appropriate traces on which to perform multichannel processing. Such multioffset phase analysis can be enhanced by f-k filtering, which assures the selection of only one wave-propagation mode, and by a statistical analysis that takes advantage of data redundancy of multishot data, usually collected, for example, in land refraction surveys. Moreover, this novel statistical method with f-k filtering can be used to retrieve a dispersion curve, in principle, for each receiver location. The quasi-continuous pseudoimage of shear-wave velocity as a function of offset and frequency allows a characterization of lateral variations in velocity, whether they are sharp or smooth.