The accuracy of dipole sonic logs with implications for synthetic seismograms and wavelet estimation

Laurence R. Lines, P. F. Daley

Sonic logs contain errors due to mud invasion and cycle skipping, and repeat logs may be recorded to validate measurements. For repeated dipole sonic logs, it is interesting to note differences in the (compressional) P-wave and (shear) S-wave velocities, as well as the resulting differences in reflectivity sequences and synthetic seismograms. For synthetic seismograms with low- frequency wavelets, the differences are often barely perceptible, especially for P-wave synthetic traces. When correlating these different synthetic traces with reflected events on real seismic data, our interpretations would often not be affected. However, for the purposes of deconvolution, seismic wavelets are often estimated by using both sonic logs and real seismic data. In some cases, where there are noticeable differences in estimated log-based wavelets, it is advisable to check log-based wavelet estimates using statistical methods, such as minimum phase wavelet estimation. Also in these comparisons of dipole sonic logs, synthetic seismograms and wavelet estimates, we have generally found the repeatability of P-wave logs to be superior to that of the shear-wave logs. This is not surprising due to the difficulty of picking shearwave arrivals compared to P-wave first break picks. In summary, repeat measures of dipole sonic logs will be worthwhile for insuring that the P-wave synthetic seismograms, shear-wave synthetic seismograms and wavelet estimates are accurate.