Converted‐wave seismic exploration: Methods

Abstract

Multicomponent seismic recording (measurement with vertical-and horizontal-component geophones and possibly a hydrophone or microphone) captures the seismic wavefield more completely than conventional single-element techniques. In the last several years, multicomponent surveying has developed rapidly, allowing creation of converted-wave or P-S images. These make use of downgoing P-waves that convert on reflection at their deepest point of penetration to upcoming S-waves. Survey design for acquiring P-S data is similar to that for P-waves, but must take into account subsurface V P /V S values and the asymmetric P-S ray path. P-S surveys use conventional sources, but require several times more recording channels per receiving location. Some special processes for P-S analysis include anisotropic rotations, S-wave receiver statics, asymmetric and anisotropic binning, nonhyperbolic velocity analysis and NMO correction, P-S to P-P time transformation, P-S dip moveout, prestack migration with two velocities and wavefields, and stacking velocity and reflectivity inversion for S-wave velocities.Current P-S sections are approaching (and in some cases exceeding) the quality of conventional P-P seismic data. Interpretation of P-S sections uses full elastic ray tracing, synthetic seismograms, correlation with P-wave sections, and depth migration. Development of the P-S method has taken about 20 years, but has now become commercially viable.

Elastic least-squares reverse time migration

“…For highly dipping events, the PP-reflections may not be recorded due to the limited recording aperture, whereas the PS-reflections, which have smaller reflection angles, are easier to record and use for migration (Stewart et al, 2002(Stewart et al, , 2003, as shown in Figure 14. In elastic LSRTM, the PS-reflections mainly contribute to construct the S-image, which helps to identify complex structures.…”

Transmission Amplitude Variation with Offset (TAVO)

“…A possible way to reduce these errors in the estimation is to assume a constant value for β/α as it is usually done in reflection AVO studies. The ratio could be obtained from the general empirical relationship between P-wave and S-wave (Castagna et al, 1985) or using full-waveform sonic measurement (Stewart et al, 2002). This would make estimation of the other two parameters from T P S more accurate as it would serve as a constraint to the inversion from the regression coefficients.…”

Wavelets and warping PS seismic images

“…Events in a PS image appear at a later time than in a PP image because the events in a PP image are a result of a downward propagating P-wave that is reflected upward, while the events in a PS image are a result of a downward propagating P-wave that is converted to an upward propagating S-wave upon reflection. Because S-waves are slower than P-waves, it takes more time for the information seen in a PS image to reach the receivers (Stewart et al, 2002). In both cases, warping (stretching or squeezing) a trace will produce wavelet distortion, which can cause problems for later processing steps.…”