The following notes are meant as guidelines only. There is no substitute for looking at each data volume and treating it individually. Each data volume usually brings it's own particular blend of problems to bear. The main key to data processing is in understanding the velocity regime of the data. Other problems are the presence of multiple reflections and imaging in the presence of severe lateral velocity variations such as salt. Most of the comments are more relevant to the North Sea than other areas of the world.
Regional surveys can be 2D or 3D. In modern times these would usually be acquired and processed as speculative surveys by the contractor. For a regional 2D survey line spacing would be around 2km. It is probable that the geological variations over the survey area would change sufficiently to merit changing processing parameters although it is equally likely that some compromise parameters will be chosen in order to process the data quickly. Lateral changes in 2D processing parameters may also exacerbate misties between lines. Providing the acquisition is good, vintage data can almost always be improved by modern processing techniques.
Once suitable targets have been identified a higher resolution 2D survey (500m grid spacing) or 3D survey may be shot over a smaller area. Processing can benefit from lessons learned from the regional surveys or reprocessing older vintages of data. Processing will generally be targeted towards a specific objective. Often some compromises will be made to enhance the defined objective at the expense of other "secondary" targets. Reprocessing may be required if the relative priority of these targets changes at a later date. If onboard processing is initially carried out for rapid turnaround then re-processing should also be considered. The processing on most 3D surveys has a shelf life of 3 years.
A near surface sedimentary section typically consists of sub-horizontal sand and shale sequences with an increasing velocity gradient (1700-2000m/s) controlled by compaction. Frequency content might be as high as 80Hz and processing at 2ms should be considered. The degree of multiple contamination will depend largely on the water depth and character of the seabed although predictive deconvolution will probably work well. DMO and prestack time migration may help to define faults, particularly growth faults. Near surface channels may effect shallow imaging and may cause lateral velocity variations and imaging difficulties. Care should be taken during survey design to ensure adequate near trace offsets and sufficient trace sampling, bearing in mind far offsets may well be muted. Consideration might be made to acquisition with a site survey vessel for optimum near surface resolution. Refraction analysis may be applied to determine near surface velocity variation. For 3D surveys any navigation errors would be more noticeable in the very near surface and seabed section.
Deeper sediments are likely to exhibit higher velocities and moderate structure. Frequencies are likely to be considerably reduced - often 30Hz is the dominant signal in the North Sea Jurassic section, largely due to absorption in the high velocity Cretaceous chalk section. At this depth, deeper than 9000 feet, the residual moveout on multiples may be minimal. Multiple suppression therefore becomes more difficult - Radon demultiple and near trace mute are the standard technologies in use today. Predictive deconvolution in the tau-p domain is an emerging technology in this area. Interbed multiples can become a problem interfering with often weakly reflective target sands. Peg-leg multiples from the base Cretaceous reflector are also a problem and can sometimes be suppressed with a target oriented technique such as SPLATä . At these depths near surface velocity variations are likely to be annealed, but the lateral and vertical velocity variation (particularly beneath chalk) may be difficult to define. 3D prestack depth migration is an emerging (but as yet largely unproven) technology in this area.
These fast (3000-4500m/s) layers can cause considerable problems in seismic processing. As well as causing a series of multiple reflections the fast layers limit usable offsets since critical offset is rapidly approached, typically around 2000m. Lateral velocity variations (often related directly to porosity) are common. If these rocks outcrop at the seabed then multiple reflections and scattered noise may be considerable. Acoustic impedance inversion is commonly applied to determine porosity within chalk reservoirs.
Salt usually creates the biggest imaging problems in seismic processing since it usually has high velocity and steep dips which create rapid lateral velocity variations when the salt is interbedded within low velocity sediments. 3D prestack depth migration is almost always required in these instances whether the target is salt flank or subsalt. Velocity gradients might be required in the velocity model for prestack depth migration. Sometimes the structures are too severe to build an adequate model for prestack depth migration. In these cases a time migration with a low velocity e.g. 90% may be useful. In the Southern Gas Basin the salt is often sufficiently deep (12000 feet) that the lateral contrast with the sediments is minimised. This reduces the difficulty of the imaging problem but the salt flanks may be difficult to pick as there is little velocity contrast between salt and sediment. For sub-salt imaging prestack depth migration is the correct solution for the velocity pull-up due to salt diapirs.