--> Abstract: Sub-Salt and Sub-Basalt Imaging Using Long Offset and Multicomponent Seismic Data, by J. Gaiser, N. Jones, and P. Fontana; #90923 (1999)
[First Hit]

Datapages, Inc.Print this page

GAISER, JAMES, NEIL JONES, and PHILIP FONTANA,Western Geophysical Co.

Abstract: Sub-Salt and Sub-Basalt Imaging Using Long Offset and Multicomponent Seismic Data

Introduction

Exploration in potential hydrocarbon producing areas beneath high-velocity layers (HVL) is typically hindered by a variety of seismic propagation effects. Dramatic elastic-impedance contrasts between the HVL and regional sedimentary framework cause P-Previous HitwaveNext Hit energy to be converted to shear-Previous HitwaveNext Hit (S-Previous HitwaveNext Hit) energy. The associated loss of frequency content, unusual amplitude versus offset effects, and numerous multiple reverberations act to limit the signal-to-noise (S/N) ratio of reflections and the ability to image structures below the HVL.

The objective of this paper is to present an overview of a number of different approaches that have proved useful in overcoming some of the above problems to obtain accurate images. The approaches involve techniques in both acquisition and processing of seismic data. For example, whether we should collect data at the sea surface (conventional towed streamer or long-offset towed streamer) or on the sea floor, such as ocean-bottom cable (OBC) data, which is itself dependent on the type of seismic body Previous HitwaveNext Hit we intend to investigate. Conventional P-waves are always important but S-waves can be used to image below HVLs in conventional and OBC data. Processing of these data always requires more accurate information regarding multiples and velocity/depth structure in order to transform and enhance our reflection data to properly image below the HVL.

Conventional Steamer Data

Prestack depth migration has been successful in imaging conventional P-Previous HitwaveNext Hit towed-streamer data.The biggest problem is that the HVL is not only a good P-Previous HitwaveNext Hit multiple generator but it also generates significant S-Previous HitwaveNext Hit energy. The implication is that velocity filters designed for removing P-Previous HitwaveNext Hit multiple Previous HitwaveNext Hit trains may have little if any effect on the S-Previous HitwaveNext Hit mode-converted noise.

The Tertiary basalts of the Faeroes-Shetland region vary in thickness from tens to thousands of meters and overlay potential hydrocarbon-bearing sediments. The velocities vary both vertically and laterally within the region (Fontana and Jones, 1998) and can result in strong impedance contrasts with the adjacent sediments. Ray-based modeling is used to identify P-Previous HitwaveNext Hit multiples and S-Previous HitwaveNext Hit converted modes that interfere with the underlying sediment reflections. The models aid in designing Radon filters for multiple attenuation. Figure 1a shows parabolic Radon transforms of the field data from the Faeroes area and the corresponding synthetic data. Primary reflections are the seafloor (P1), top Oligocene (P2), and top basalt (P3), that can easily be identified and separated from multiples. Diagonal lines indicate the approximate mute boundary. Along with the improved S/N ratio of primaries, the ray-based models help in building improved P-Previous HitwaveNext Hit velocity/depth models for prestack imaging of the base of basalt and below. Figure 1b is a common image-point gather, offset increasing to the right, illustrating the base basalt-reflector at 2700 m and base Cretaceous at 4000 m. Note the large amount of under-corrected mode-converted S-Previous HitwaveNext Hit energy that curves downward below top basalt and interferes with primary reflections.

Alternatively, base of basalt (as well as deeper horizons) can be effectively imaged using.the mode-converted S-waves by replacing the P-Previous HitwaveNext Hit velocity in the HVL with the appropriate S-Previous HitwaveNext Hit velocity (Purnell, 1992). Figure 1c illustrates such an approach where we have derived an S-Previous HitwaveNext Hit velocity structure within the high-velocity basalt layer and used that velocity/depth model to drive a PSSP depth migration. PSSP waves are P-waves that convert to S-waves at the top of basalt, reflect off horizons below, and are subsequently converted back to P-waves at the top of basalt. The result is that slow events in the P-Previous HitwaveNext Hit common image-point gather tend to map to comparable depth structures, e.g., base basalt and base Cretaceous.

Figure 2a is a portion of the velocity/depth model used to drive multiple suppression and depth migration, yielding the associated seismic depth-stack (Figure 2b). Over this limited portion of the model, the basalt thickness varies between 200 m and 280 m. The estimates of thickness are in agreement with values derived from PSSP imaging. Basalt velocities vary laterally between 4100 m/s and 4300 m/s. Notable structures (arrows) are base basalt at 3000 m, the strong unconformity (base Cretaceous) at about 4200 m depth, and the dipping event below 5000 m depth, that is thought to be a pre-rift feature.

Long-Offset Streamer Data

Another method of deriving velocity/depth information in HVL areas makes use of long-offset arrivals rather than dealing with conventional spreads. In this method we make use of diving-Previous HitwaveNext Hit arrivals associated with energy that has continuously turned along ray paths within the subsurface. By deriving traveltime attributes associated with these first breaks we tomographically invert for P-Previous HitwaveNext Hit velocity/depth structure in areas that suffer poor S/N due to multiples and coherent mode-converted noise.Acquiring such towed-streamer data necessitates multi-boat and multi-pass 2-D profiling.We present an example from an experiment conducted in the Faeroes-Shetland region where dual-vessel acquisition gives rise to the formation of super shot gathers ranging in offsets from 0-36 kilometers.Tomographic inversion of the traveltime data gave rise to sub-basalt velocity/depth models, allowing an independent means of corroborating models derived from prestack depth migration.

Ocean-Bottom Cable Data

The recent advent of OBC technology has brought with it the ability of recording high-fold, regularly sampled S-Previous HitwaveNext Hit data using multicomponent instruments. This involves detecting sea-bottom particle motion with three-component geophone receivers (two horizontals and a vertical) and a hydrophone. Combining the vertical geophone and hydrophone data appropriately attenuates water-born multiples to provide high S/N P-Previous HitwaveNext Hit reflections for prestack imaging the base of an HVL as well as structures below (Kendall et al., 1998). The above allows us to record the same Previous HitwaveNext Hit types as observed in our conventional towed-streamer data (e.g., P-waves and PSSP-waves). On the other hand, combination of the two horizontal geophones (based on survey geometry) allows us to analyze S-waves that have converted from P-waves at the target reflectors (PS-waves).The PS-Previous HitwaveNext Hit reflections travel back to the sea floor where they are detected by horizontal receivers. This is in contrast to S-waves used for imaging in towed-streamer data that must convert back to P-waves on their journey to the hydrophones.

Imaging with PS-Previous HitwaveNext Hit data from OBC surveys requires additional processing effort compared to that used in P-Previous HitwaveNext Hit depth migration. Using the P-Previous HitwaveNext Hit velocities and Vp/Vs information from the converted-Previous HitwaveNext Hit processing, prestack depth migration of the PS-Previous HitwaveNext Hit data can also be accomplished to give independent subsurface images of base salt and the structure beneath to complement the P-Previous HitwaveNext Hit images. Figure 3 illustrates such an example of P-Previous HitwaveNext Hit and PS-Previous HitwaveNext Hit depth migrations of an OBC experiment from the Gulf of Mexico. Although no parabolic Radon multiple attenuation was performed, possible sub-salt reflections are present (arrows) in both sections.

Conclusions

Ray-theoretical modeling has been effective in aiding multiple attenuation of data from the Faeroes region where thick basalt layers overlay sediments. The result is improved P-Previous HitwaveNext Hit S/N and velocity/depth models for prestack imaging of the base of basalt.Also, base of basalt (as well as base of salt intrusions in the Gulf of Mexico) can be effectively imaged using mode-converted shear waves by replacing the P-Previous HitwaveNext Hit velocity in the HVL with the appropriate S-Previous HitwaveNext Hit velocity. In addition, depth/velocity models of sediments below HVLs can be obtained from long-offset streamer data by tomographically inverting refraction first-breaks. Finally, OBC technology has brought the ability to image the base of an HVL with PS-waves to complement P-Previous HitwaveNext Hit images.

Acknowledgments

The authors thank Gillian Brown at Western Geophysical and Doug Toomey of the University of Oregon for their help in creating the velocity models for the Faeroes data. A special thanks goes to Greg Wimpey of Western Geophysical for his assistance in depth imaging the Gulf of Mexico data. Robert Bloor and Rich Van Dok, also of Western Geophysical, provided much support in processing the Gulf of Mexico data.

References

Fontana, P. and Jones, N., 1998, Long offset and multicomponent recordings: application to sub-basalt imaging, Presented at the Advances in Seismic Technology Conference, Stavanger Norway.

Kendall, R.R., Gray, S.H., and Murphy, G.E., 1998, Subsalt imaging using prestack depth migration of converted waves: Mahogany Field, Gulf of Mexico, 68th Ann. Internat. Mtg., Soc. Expl. Geophys., New Orleans, Expanded Abstracts, 2052-2055.

Purnell, Guy W., 1992, Imaging beneath a high-velocity layer using converted waves, Geophysics, 57, p. 1444-1452.

Figure 1: Parabolic Radon transforms (a) of field data from the Faeroes (upper) and synthetic data (lower). P1, P2, and, P3 are primary reflections and all other events are multiples. Diagonal lines indicate the approximate mute boundary. P-Previous HitwaveNext Hit common image gather (b), and PSSP common image gather (c) of Faeroes data.The PSSP mode conversions are P-waves that convert to S-waves within the basalt and after reflection convert back to P-waves at the top of basalt.

Figure 2: Portion of the P-Previous HitwaveNext Hit velocity/depth model (a) derived from prestack depth migration (b) in the Faeroes area. Good similarity exists between the model and the depth section in the character and structure of the top and base-basalt reflector, the base Cretaceous unconformity, and the steeply dipping event.

Figure 3: P-Previous HitwaveNext Hit (a) and PS-Previous HitwaveNext Hit (b) depth sections from an OBC experiment in the Gulf of Mexico. Using the P-Previous HitwaveNext Hit velocities and Vp/Vs information from the PS-Previous HitwaveNext Hit processing, prestack depth migration of the PS-Previous HitwaveNext Hit data can also be accomplished to give independent subsurface images of base salt and structure beneath to complement the P-Previous HitwaveTop image. Although no parabolic Radon multiple attenuation was performed, possible sub-salt reflections are present in both sections (arrows).

AAPG Search and Discovery Article #90923@1999 International Conference and Exhibition, Birmingham, England