As an example, consider the problem of converting three dimensional time surfaces and fault planes to depth. The success of this operation depends on:

1. The correctness of the time interpretation and the accuracy with which the seismic data were processed.
2. The intelligence which went into the building of an accurate earth velocity model.

It is this second aspect which has been neglected for so many years. However, we now possess, courtesy of recently developed software packages, computer tools specifically designed to perform the manipulations that we should have been doing all along. We are able, for instance, to 'insert' a time horizon or fault surface from an interpretation -workstation into a three dimensional cube of velocity, and then directly make the depth conversion (See Figure 2). However, what is needed before this can be done, is the velocity model itself.

Fig. 2

In order to transform a time horizon to depth, we need to use a field comprised of accurate vertical propagation velocities. It is incorrect to use raw processing derived velocities, even if they happen to come from the processing of a pre-stack depth migration. Appropriate velocity fields are formed in the computer from carefully scrutinized and edited vertical checkshots. Any horizontal component in the checkshot will compromise the result. The problem, even with vertical checkshot derived velocities, is that wells are often drilled in non-ideallocations for the purposes of good velocity work. Structural highs and fault zones, which involve localized anomalies in interval velocity, do not provide perfect velocity measurement conditions. However, the checkshots are still the best data we have for depth conversion purposes.

The construction of an accurate velocity model requires that the time interpretation be i an integral part of its formation. Whether the derived velocity field is a simple cube or a more complex layered model, the success of depth work still relies on the interpretation of the geologically well informed geoscientist. In the end, exploration success depends more than anything else on the accuracy of the depth model, and that model is derived from interpretations of both time AND velocity.

Paul Kennedy