3D pre-processing techniques for marine VHR seismic data

In VHR3D data, the positions of the source and receivers are required to decimetre accuracy in all three directions to ensure the correct processing of the data. However, current acquisition technology does not permit this level of accuracy in a cost-effective way. Any available instrumentation for the real-time measurement of source and receiver position is both very costly and designed for conventional seismics. During acquisition, movement of the boat and the streamers away from the nominal geometry can affect the recorded data and the quality of the processed results. In conventional seismic acquisition, these variations are small compared to the dimensions of the system and are generally not considered a major problem. In very high-resolution data however, where frequencies above 800 Hz and bin sizes of 1–2 m are common (Table 1), they can severely affect the results. Wave motion and tidal variations can produce degradation of the signal by destructive interference in the stack. Variations in x and y, if uncorrected, can cause traces to have an erroneous source-to-receiver offset distance or to be included in the wrong 3D bins. A schematic diagram of the possible movements that the acquisition system can be subjected to is shown in Fig. 1. The three main movements that are identified produce variations in the theoretical x, y and z positions and require different corrections in processing. Vertical variations due to wave motion and swell require static corrections (or time-shifts). Lateral variations, due to currents and changes in the ship’s heading, require dynamic corrections since they change the source-to-receiver offset. Tidal variations, on the other hand, require time shifts but at zero-offset, i.e. after application of NMO. The vertical variations are analogous to the near-surface problems in land data which require the application of residual static corrections. Algorithms exist in most processing packages to derive these residual static corrections using various correlation techniques. These are generally surface consistent with each shot and receiver having a consistent static. In the marine case, consistency is maintained for the shots but not for the receivers, since they move with the streamer. CDP consistent algorithms can be used to correct these problems and to improve the quality of the final stack section. However, they cannot take into account variations in the sea floor or provide reliable information for the relative positions of shot and receivers if pre-stack imaging or reflection tomography is envisaged.