P-SV and P-P synthetic stacks

Catherine E. (Tina) Howell, Donald C. Lawton, Edward S. Krebes, Jeffrey Blair Thurston

A method was developed for calculating converted-wave (P-SV) or compressional-wave (P-P) 'zero-offset' synthetic seismograms. The method is based on raytracing various offsets. The algorithm is based on the simplifying assumptions of horizontal homogeneous layers with constant time interval thicknesses, no amplitude attenuation and no multiples. Interval times are calculated, in the converted-wave case, by ray tracing a downgoing compressional wave and an upgoing shear wave. Exact values for phase and amplitude for each raytraced angle are calculated. By convolving with a Ricker wavelet, synthetic traces are generated for each offset. NMO and mute can be applied prior to stacking the traces into the P-SV or the P-P synthetic stack.

The algorithm was tested for an offset, mode-converted, VSP in South Alberta and a multicomponent seismic survey in West Central Alberta. Each set of seismic data correlated well with the appropriate synthetic stack. The P-SV synthetic stack, using either the full-waveform log or a sonic log and a Vp/Vs ratio, provided a reliable match between depth information from well logs and reflections observed in converted-wave seismic data. For the P-SV case, the Vp/Vs ratio strongly affected the traveltimes of events, while selection of NMO and mute had a more significant influence on the character of the synthetic stack. The P-P synthetic stack more accurately preserved the amplitude information than the P-P zero-offset synthetic when the events exhibited a significant amplitude variation with offset.