Joint P- and S-wave prospection and a new method of S-wave extraction

The shallow seismic prospection is most often carried out using the P-waves as "illuminating" rays. It is sufficient in many cases, however, sometimes also other information is necessary. In hydrogeology, for instance, porosity and fracturing of rocks play the key role. In geotechnics, the stress state of the rock is of vital interest. The Lamé parameters μ and λ (or the E and σ parameters) are used when evaluating bedrock state for constructions. Moreover, the joint P- and S-wave interpretation gives much more detailed image of the subsurface than the interpretation of only one wave type. However, problems with the identification of S-wave onsets and S-wave wave forms make it a very demanding task. Usually, a special S-wave source is needed for the S-wave survey. Unfortunately, the S-waves sources suffers from a very low generated energy or are heavy, clumsy and expensive. Even though conventional P-wave sources also produce sufficient amount of S-waves, their usage in S-wave prospection is limited due to the problems with onset identification.
Over the years, many different techniques for S-wave identification were developed. They are often based on polarisation analysis of multicomponent records. Unfortunately the polarisation analysis is not always successful with the near surface data. Therefore a new method of S-waves extraction was developed. This method is based on comparison of neighbouring records.
The three neighbouring records are shifted in time and summed in such a way that the P-waves are summed destructively and cancelled out and the S-waves are left. Now the S-wave onsets can be conveniently picked. However, simple linear summation (stacking) is not sufficient for this task and some kind of non-linear summation must be used. The generalized average of signals (the arithmetic average multiplied with a power of the ratio of the first and second moment) works well for this task.
This S-wave extraction method was used on a three-component field data (from the common P-wave source -- a sledgehammer) to resolve the boundary between the weathered carboniferous arkoses and heavily weathered Proterozoic phyllites. This boundary could not be delineated from velocity cross-sections of solely P- or S-waves. However, when the P- and S-waves (SV and SH) velocity information is combined together then also other features (water-table level, fault zone) than solely the boundary could be inferred.
The field results showed that the described method of S-wave extraction is capable of extraction of the S-waves from the simple P-wave source records and can be recommended for intensive testing. The resulting joint P- and S-waves velocity cross-sections can produce impressively detailed image of the subsurface.