The Road to Prestack Inversion

By

Mark Egan¹, Subhashis Mallick¹, Nader Dutta¹

(1) WesternGeco, Gatwick, United Kingdom

 A common problem in seismic surveys is the lack of acoustic impedance contrast between the reservoir and surrounding country rock. This means that little or no p-wave energy is reflected - thereby rendering the reservoir invisible. By definition, post-stack inversion is pointless in this situation.

A solution that is often proposed is to exploit shear waves recorded via multicomponent phones. In some areas this certainly can be a viable approach. However, this can be costly - and in many other areas, totally impractical for a variety of reasons.

An alternative is to create pseudo shear wave data by using prestack inversion. Indeed the subsequently derived Poisson’s ratio data volume can be an excellent lithology discriminator. This fact is exploited extensively today in geohazard prediction in deepwater environments. Prestack inversion can also have benefits in both thin layer detection and fracture characterization. The challenge is providing the inversion program with prestack records of sufficient signal-to-noise quality.

For this reason, all efforts must be made to ensure that seismic surveys are designed suitably well. An effective tool in this regard is modeling. For each of the candidate survey designs, modeled records containing signal and noise can be input to prestack inversion for direct evaluation. Or, by modeling signal and noise components separately, signal-to-noise ratios can be computed and then empirical rules can be used to predict the success of prestack inversion for each design.

Experience with this methodology generally shows that in the onshore case, high channel count, single sensor systems and Digital Group Forming algorithms are often needed. Benefits are witnessed in marine cases too.