© 2007 EAGE 37 technical article first break volume 25, November 2007 The world’s demand for energy is accelerating, while its hydrocarbon reserves are diminishing. Producers are com- pelled to explore and produce oil and gas in more chal- lenging environments and to maximize recovery in existing reservoirs. New technology has always been a key to success. One such new technology is seismic acquisition using ocean bottom station (OBS) nodes (Berg et al., 1994; Ronen et al., 2003; Amal et al., 2005; Docherty et al., 2005; Granger et al., 2005). Surface-towed streamers provide excellent seismic data for exploration, development, and production monitoring. How- ever, streamers have limitations and this motivates a quest for alternative technologies. When using streamers, obstacles such as production platforms require undershooting and the data lack near offsets and have anomalous azimuth distributions. OBS nodes are much less sensitive to obstacles and allow more complete coverage. When deployed by a remotely-operated vehicle (ROV), they can be placed very accurately, right up to, or even under, production installations. Obstacles are not the only motivation for OBS tech- nology. Even in the absence of obstacles, OBS nodes facili- tate wide-azimuth geometries, important for imaging struc- tures under complex overburdens such as salt environments. In particular, by moving the receivers to the seabed and recording a dense shooting grid on the sea surface, we can create a dataset that is suitable for wave-equation migration, is well-populated in azimuth and offset, and provides opti- mal subsurface illumination; conventional towed-stream- er acquisition cannot provide this type of dataset because of the constraints imposed by the fixed source-receiver geome- try. Indeed, while the wide azimuth towed streamer (WATS) method is becoming popular, and can also provide data suit- able for wave-equation migration (Threadgold et al., 2006), it requires separate source vessels as well as repeated shot lines with different streamer locations. Such effort comes at a cost and, for small areas (up to about 400 km 2 ), OBS tech- nology can be less expensive than WATS. Similarly, monitor- ing with time-lapse seismic surveys (4D) requires acquisition repeatability; OBS nodes deployed by ROVs provide better repeatability than streamers, which are subject to feathering due to currents. Indeed, although the limitations of stream- ers are reduced with streamer-steering, streamer-steering can only have a small effect of just a few degrees, while feath- ering is sometimes 10 0 or more. Eiken et al. (2003) report a natural streamer feathering within +/- 6 o for 95% of Nor- wegian Sea operations, and a streamer steering capability of +/- 3 o that is not sufficient to achieve a zero-feather sur- vey in a cost effective manner. Streamer steering also intro- duces noise that can be difficult to remove. Finally, nodes deployed on the seabed record not only P-waves but also S-waves; these are a useful complement to P-waves but do not travel in water, and therefore cannot be recorded by sur- face-towed streamers. OBS nodes can provide four-compo- nent (4C) seismic data from a hydrophone and a three-com- ponent geophone, thereby enabling elastic-wave analysis, as well as wavefield separation and multiple removal. Applica- tions of elastic-wave analysis include imaging beneath gas clouds and imaging low P-impedance reservoirs and fracture characterization. One alternative to OBS nodes is use of ocean bottom cables (OBC), where sensors are embedded in cables rath- er than nodes. OBC operations have lower cost than OBS in shallow water, but are more limited by depth and obsta- cles such as pipelines and seabed installations. Also, cables have an inherent in-line and cross-line asymmetry which may compromise vector fidelity. Although shear-waves are a significant motivation for recording four-component data on the seabed, the focus of this paper is P-wave imaging of OBS data with sparse receiv- er sampling. Deploying OBS nodes takes considerable time and ROV boats have a considerable cost. Therefore, a prac- Mirror imaging of OBS data Sergio Grion, 1 Russell Exley, 2 Michel Manin, 4 Xiao-Gui Miao, 3 Antonio Pica, 4 Yi Wang, 1 Pierre-Yves Granger, 4 and Shuki Ronen 1 1 CGGVeritas, 10300 Town Park Drive, Houston, TX-77072, USA. 2 School of Geography, Earth & Environmental Sciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK. 3 CGGVeritas, 715 5th Avenue SW, #2200, Calgary, Alberta T2P 5A2, Canada. 4 CGGVeritas, 1, rue Leon Migaux, 91341 Massy Cedex, France. Figure 1 Poor illumination of sparse OBS. Note the gaps in shallow reflectors coverage. This problem is exasperated if any OBS nodes fail. (Shots are shown as black stars, receivers dead and alive are yellow and black spots respectively.)