PITFALLS AND SOLUTIONS IN SEISMIC INTERPRETATION:
The World of Stratigraphic Interpretation – SEG Workshop Paper
by Dr Easton Wren

The word pitfall implies consequences and in the realm of stratigraphic interpretation where signal fidelity is cardinal, the stacking process has fundamental consequences. Stacking is a statistical averaging process which does nothing for the signal but improves the signal/noise ratio. The early assumption after NMO correction was that the signal was flat and constant across the gather, thus averaging out perfectly, while the noise was not aligned, or was random, and was therefore at least partially cancelled. The first assumption is incorrect.

Stacking has obvious benefits. It is necessary in the structural realm, particularly with 3D data volumes. In a medical analogy, however, it is similar to chemotherapy where the cancer virus is attacked but the consequences include side effects that destroy otherwise healthy aspects of the patient. Thus the patient (Signal) is affected by the cancer (Noise) treatment.

AVO addresses the desirability of the pre-stack realm. However in the day-to-day world of seismic interpretation on workstations there may be no time for pre-stack analysis and interpretation of the gathers; there may be no interest; there may be no gathers; there may be no workstation facility for analysis of the gathers; there may well be lots of excuses!!

The pre-stack data base is potentially vast with high fold data but contains much useful and honest information, especially in the case of signal attributes. The pre-stack domain also facilitates processing QC, a focus for the quality and INTEGRITY of the stack. Note the potential effect of non-azimuth- varying NMO corrections in 3D data sets, and the consequences on the frequency content of the stacked data.

It is ideally necessary to access the pre-stack data gathers on the fly while interpreting the stacked /migrated data. We would require instant access to any CDP gather or a common offset/Ostrander display by clicking on a stacked trace and then examining the nature/quality of the pre-stack vis-à-vis the stacked output. Other collections such as range limited stacks should be instantly available; spectral plots of frequency as a function of offset would be invaluable. We should be able to ask the machine to hunt for 2D or 3D AVO anomalies as we control the processing.

In his Leading Edge article Brian Russell when SEG President talked about the geophysicist of the future...the enlightened geophysicist would look at every piece of the chain, including the acquisition footprint, distribution of offsets and azimuth for each subsurface bin, the data before and after deconvolution, the velocity cube used for migration etc. Thus the geophysicist of the future will be more like the geophysicist of the past...someone who has insight into all areas of acquisition, processing and interpretation as well as having imagination.

The power of the workstation can thus be used to extend our ability and our imagination in this direction.

This paper was presented at the SEG Conference during the Pitfalls in Seismic Interpretation Workshop. Dr Wren directed a live workstation paper using the EarthWorks Exploration System to show how real-time prestack analysis can prevent common seismic interpretation pitfalls.