A mode-converted (P-SV) synthetic seismogram

Jeffrey Blair Thurston, Catherine E. (Tina) Howell, Donald C. Lawton, Robert R. Stewart

A method has been developed for calculating a "zero-offset" converted-wave (P-SV) synthetic seismogram. The algorithm is based on the simplifying assumption of layer thicknesses with constant interval times. These interval times represent the zero-offset travel time of a downgoing compressional wave and an upgoing shear wave. A reflected amplitude is computed for each offset used in the data acquisition, using the exact expression for a P-SV reflection. These reflection amplitudes are stacked to create the zero offset trace. Offsets are mapped to incident angles by raytracing the model. Computation of the synthetic requires either a full waveform sonic log, or a compressional-wave sonic log and a time variant estimate of the VP/VS ratio.

The algorithm has been tested on a full waveform sonic log from southern Alberta, and compared to an offset mode-converted VSP. In general the correlation between the synthetic and the real data is good, as the major reflections are successfully predicted in both time and amplitude.