In simple terms, every subsurface point at the target should be properly illuminated during a 3-D seismic survey and should have reflected seismic energy with a uniform distribution of source-receiver offsets, azimuths and incidence angles. This definition forms a classic criterion for high-resolution, high-quality seismic imaging.

Figure 1. Schematic WATS acquisition configuration for one streamer vessel and two (fore and aft) source vessels, acquired for two “tiles.” In this example, the streamer length is 26,250 ft (8,000 m). (All figures courtesy of PGS)
In the most extreme cases, target events cannot be imaged because of the lack of target illumination. Data processing cannot generate data that have not been recorded, whatever sophisticated processing and imaging technologies are employed. Furthermore, wave propagation effects may be so complex that the resulting seismic images are incoherent and ambiguous — even if the target is illuminated. The most typical example is subsalt challenges in the Gulf of Mexico. So-called wide-azimuth towed streamer (WATS) acquisition is increasingly used in an attempt to overcome incoherent imaging problems.

Wide-azimuth and multi-azimuth seismic in practice

Wide-azimuth seismic is typically considered in either of two modes: using several marine seismic vessels, or using a regular grid of (autonomous) seafloor receivers and a regular shot grid at the surface. The seafloor option is expensive and can only be considered on a prospect scale, whereas WATS becomes increasingly cost-effective as the survey size increases. To date, WATS applications have all been for subsalt exploration, where the highly 3-D nature of salt geometries can make the planning of optimal azimuths for multi-azimuth (MAZ) surveys quite challenging. A large variety of source and streamer vessel configurations for WATS seismic have already been published in the literature, and each will have its own merits. BP acquired a WATS survey over the Mad Dog field in the Gulf of Mexico using two additional source vessels placed at equal lateral distances from the streamer vessel.

Each source was fired in sequence until all sources had fired, and the sequence cycled back to the first source. Each source line in the Mad Dog WATS survey was acquired four times such that the lateral offset of the streamer vessel was increased by 3,280 ft (1,000 m) each time the line was shot. Each such acquisition pass is referred to as a “tile.” The number of tiles and/or the lateral offset chosen will be dictated by both cost and geophysical considerations. Note that the BP approach also has the advantage of being expandable by later surveys — the shot lines can be repeated with new lateral offsets, and the collective data will be reprocessed to output a new dataset. Geophysically, repeat shot lines for each tile are preferential for addressing the spatial sampling requirements of pre-stack depth imaging.

Figure 1 shows the WATS approach with two tiles, each using a common shot line but with the source vessels placed a different cross-line offset from the streamer
vessel. Most receivers in the streamer spread will therefore acquire a reasonably large range of source-receiver azimuths (0 - 180Þ), a large range of inline source-receiver offsets, and a (coarsely sampled) range of cross-line source-receiver offsets, as dictated by the shooting strategy used in the WATS survey (Figure 2).

Figure 2. Offset/azimuth coverage for the WATS configuration in Figure 1, using a cross-line vessel pass separation of 3,937 ft (1,200 m) and a sail line separation of 1,970 ft (600 m).
Figure 2 shows the surface offset and illumination coverage for the WATS configuration in Figure 1 using a cross-line vessel pass separation of 3,937 ft (1,200 m, two tiles), and a sail line separation of 1,970 ft (600 m). While the offset coverage is heavily biased towards the shooting direction, it is evident that the offset/azimuth coverage is smoothly varying in all directions. Furthermore, the stacking of many coherent noise events (notably diffracted multiples) from different acquisition tiles will attenuate such events due to their differential moveout on normal moveout (NMO) corrected gathers.

In contrast to WATS, seafloor acquisition with either (coarsely sampled) autonomous nodes or “cross-shooting” with seafloor cables will result in more azimuthally symmetric offset/azimuth coverage. Operationally, both seafloor wide-azimuth solutions will be significantly more expensive than WATS acquisition.

Figure 3 shows the illumination results modeled for a tabular salt model with 3-D ray tracing for both MAZ and wide-azimuth seafloor seismic. The rose diagrams show the surface source-receiver azimuth (radial lines) and offset (circles) distribution for both streamer acquisition and 3-D seafloor acquisition. Each rose diagram has a maximum offset of 26,250 ft (8,000 m). The seafloor survey was modeled with several seafloor cables and a large areal shot grid where the shot lines were orthogonal to the seafloor receiver cables. When two streamer azimuths are added together there is still quite a narrow range of source-receiver azimuths in comparison to the cross-shooting seafloor survey. In terms of target illumination below the salt, the orthogonal streamer surveys yield complementary non-ideal illumination, whereas the seafloor survey yields quite uniform illumination. When added together, the MAZ streamer illumination is greatly improved and is a reasonable approximation to the much more expensive seafloor result. The MAZ streamer illumination could be further improved by acquiring additional complementary azimuths.

In the context of “survey expandability,” MAZ streamer seismic offers a relatively low-cost geophysical incentive. Interpretation of 3-D data acquired along a single azimuth using conventional vessel geometry commonly reveals deficiencies in the target imaging.

Post-survey modeling using an interpreted

3-D model and 3-D ray tracing will often demonstrate that alternative shooting azimuths will yield complementary (but different) target illumination and that the combination of different
Figure 3. Comparison of the surface source-receiver offset/azimuth distribution and subsurface target illumination for streamer vs. cross- shooting seafloor acquisition. The target illumination area shown is 7.3 miles by 7.3 miles (12 km by 12 km), and the maximum offset shown in the rose diagrams is 4.8 miles (8 km). The top two rows are single-azimuth and multi-azimuth streamer acquisition, while the cross-shooting seafloor results are on the bottom row.
azimuths will inevitably yield superior illumination, image quality and resolution. Provided that the existing 3-D data were acquired with acceptably good spatial sampling and far-offset coverage, it reasonable to expect that a new 3-D survey acquired in the orthogonal direction can be used to process a new MAZ dataset. If the original dataset was poorly acquired, the new survey data will be compromised accordingly during the MAZ processing. Several options exist during MAZ processing, and a variety of prestack time and depth products can be delivered with varying degrees of cost and effort. This flexibility is highly attractive as MAZ datasets can be created using a single streamer vessel without the expensive requirement of additional source vessels and the multi-pass (“tile”) shooting strategy used for WATS surveys.

Survey planning

Each seismic imaging challenge also can be considered to be an illumination challenge — potentially addressed by MAZ or WATS seismic. Wherever a dense 2-D dataset or a 3-D dataset exists, it will be possible to build a representative 3-D model and use 3-D ray tracing to perform illumination analyses to compare and contrast the relative merits of narrow-azimuth, MAZ and WATS seismic. Elastic ray tracing or finite-difference modeling and processing can also be used to generate synthetic seismic data volumes, thus extending the pre-survey planning to also incorporate imaging considerations.

Overall, a range of acquisition and processing strategies are available to address the seismic imaging challenges encountered throughout the world. Of particular interest is the potential to acquire new 3-D surveys along azimuths orthogonal or oblique to existing 3-D surveys, thus inexpensively forming a new MAZ dataset. Global experience demonstrates that two or three complementary azimuths quickly build up the uniform distribution of offsets and azimuths required for optimal target imaging and resolution in most geological settings. Likewise, new WATS tiles with larger cross-line source-receiver offsets can be acquired over existing WATS surveys, further improving the target illumination.

Comments