Conclusions

I chose here a particular example of relating velocity measurements at different scales, namely a well log and a surface seismic data set. I have applied Dix and Schoenberg&Muir averaging to the well log and compared it with the surface seismic velocity analysis. Both approaches replace the overburden with an homogeneous equivalent. However the sense of averaging is different. Overlaying velocity analysis, Dix rms velocity and §+M derived velocities show that Dix rms velocity seems consistently higher (up to 15%). The §+M average gives velocities that seem closer to the analysis picks (up to 5% deviation); At large depths however it is noticeably off the velocity picks. The §+M average is consistently lower than the Dix rms average, as we would expect since the high frequency average is based on the Fermat's principle and is thus least travel time.