Pre-Stack Depth Migration from Regional 2-D Seismic Surveys, a Cost and Time Effective Way of Optimising Structural Detail

Chris Benson and Dane Southward
Business Development, GGS-Spectrum Ltd, Woking, United Kingdom

The initial exploration of almost all frontier areas will almost certainly include the acquisition of regional 2D seismic data. This data will then most probably be processed through a conventional time processing sequence and thereafter interpreted to identify prospects for further more detailed seismic studies.

With the significant decrease in computer run times and the proportional reduction in processing costs now achievable with scalable, modern PC clusters, an increasing amount of seismic data is being processed using more sophisticated techniques. Pre-Stack Depth Migration (PSDM), for example, which, due to excessive computer run times was prohibitively expensive only a few years ago, is now increasingly utilised to provide more accurate imaging of geological structures and hydrocarbon prospects.

PSDM requires an accurate interval velocity model to achieve optimum results. This can be difficult in frontier areas with little or no well control. However, allowing for this limitation, it is possible to build a regional interval velocity model in a time and cost effective manner which can be utilised in the PSDM process to generate a more optimal subsurface image. The process requires utilisation of RMS velocities from the seismic data. The derived velocity field is interpolated and smoothed using proprietary software. This ensures that both the seismic data and velocity fields tie following the migration process. This regional RMS velocity field is subsequently converted into a regional interval velocity field. Any available geological information is incorporated at this stage to add some constraints to the model. The interval velocity model is then sub-divided into discrete velocity layers which match as closely as possible the geological horizons. The velocity model is refined and updated though a number of PSDM iterations using tomographic updates.