Oil companies around the globe have made substantial investments in both software and computer resources to support advanced interpretation techniques for 3-D seismic. It has been observed, however, that the adoption and regular use of these systems have not spread to the entire population of geoscientists and that section-based interpretation software is still the most commonly used application, even in companies where virtually all the data consists of 3-D seismic.

A number of reasons have been advanced to explain this phenomenon, but one item has been frequently overlooked: Can the interpreter leverage all the available information? In many cases, 2-D seismic is still an important component of an interpretation project. It ties together different seismic campaigns, both 2-D and 3-D, that span an area of interest for exploration. It ties a small 3-D survey to surrounding wells. Regional 2-D grids make it possible to put structural and depositional systems into a broader context. And yes, there are still many places where 2-D data is the only data available. As exploration activities move into new frontier regions where 3-D data is still scarce, the efficient and accurate interpretation of 2-D seismic is vital to a rapid economical appraisal of vast permit areas.

Displaying 2-D data in a three-dimensional viewer is nothing new. But the critical mass of interesting capabilities was apparently not reached, so the many users of section-based interpretation systems felt vindicated in their continued use of section-based systems. A "tipping point" has been reached, however, with the recent introduction of a full-fledged 2-D and 3-D interpretation system designed from the ground up to satisfy the needs of conventional interpretation workflows. In addition, this system offers enticing new capabilities that will attract a whole new population of users to embrace three-dimensional tools for their daily work.

Building 3-D pictures from 2-D data

As trivial as this may seem, the ability to view data not only on the specific seismic sections but also projected in front of them makes it much easier to understand the spatial relationship between fault blocks, particularly on strike lines.

Lining up sections for viewing along a specific trend line is another intrinsic quality of three-dimensional representation. The difference from conventional section-based work lies in the fact that all displays and combinations of interpretation objects are occurring in a single window. The user can zoom, rotate or tilt the display up and down, while turning on or off any section to get a clear view of the data in the most advantageous manner. Well data, particularly in areas with dense drilling and deviated boreholes, can be more readily associated to seismic, with the ability to view the true geometry of both the well and all the surrounding seismic sections. Marker correlation becomes significantly easier, and the verification of fault geometries against the occurrence of faults in a well is much simpler and more straightforward.
Beyond these immediate features, the value of the complete three-dimensional solution lies in its comprehensive capabilities. The fault interpretation process builds tessellated surfaces as the interpreter progresses through the data. This makes it much easier to assess the consistency of the interpretation while it is ongoing. Horizons can be gridded on the fly, and their surfaces or contours can be verified directly on the data.

3-D seismic

While remaining in the same three-dimensional window, 3-D seismic data can be brought into the display for a combined interpretation exercise. There are no limits to the amount of 3-D seismic on view or to the number of different surveys used concurrently. An intelligent cache process maximizes use of computer memory resources while making it possible to view data volumes many times larger than the computer's live memory. The interpretation process can skip back and forth between 2-D and 3-D datasets, all the time checking for inconsistencies. Grids, maps and solid models can be built from the combined data and further used for depth conversion, structural analysis or property mapping.

Is this enough?

The foregoing arguments and features represent significant new capabilities for the use of three-dimensional visualization and interpretation tools to address all seismic data, 2-D and 3-D. Yet the early reactions to this new offering indicate that it is mainly understood and adopted by the avant-garde user community, which is already well versed in the use of such systems for the interpretation of 3-D seismic. What could be missing?

Looking at the few illustrations in this article so far, we have only shown 2-D seismic sections as opaque displays of variable density sections. Even if we were to use a color table, it would still be a section of seismic data, with the addition of perspective views of the data and the interpretation objects.
If we want to assess progress in our interpretation and obtain an overall view of the information, we need to switch off the seismic and view in 3-D the horizons and faults we have created. This is not an ideal workflow as it implies interpreting data before we can get a view of the information in some form that would expose and illuminate particularities in the data, ahead of interpretation.

Transparency

If we look at what has defined 3-D volume interpretation since the early days of Vital Images and VoxelGeo, it has been the ability to rapidly scan through the data and identify anomalies by applying opacity curves to slabs of data and lining up contiguous events in areas of challenging signal quality; this technique is referred to as optical stacking. In 3-D datasets, optical stacking is routinely used for enhancing subtle structural or stratigraphic details or delineating the extent of fluid contacts. It consists of thresholding a dataset to show only specific values and then seeking the optimal viewing angle that will align with the structural grain both in terms of direction and of the vertical angle of viewing such that the significant events stack up and unrelated information appears as scattered and diffuse. The ability to superimpose optically a set of seismic sections with full control of the angle of viewing greatly facilitates the qualification of the features that are observed individually. Figures 5(left) and 5 (right) clearly demonstrate how the horizons and faults are delineated once the correct angle of viewing is achieved.
Two-D data offers both challenges and advantages in pursuing optical stacking. Contrary to 3-D seismic, we will not have the density of sampling that will result in the deterioration of noise while coherent events are reinforced. On the other hand, the cloud effects generated by dense 3-D will not be as prominent for 2-D data.

It must be accepted in all cases that 2-D data suffers from limitations inherent to the nature of the data. A 2-D section includes information that is not within the plane it occupies in space. Unless it is a perfect dip section from the surface down to the objective, artifacts will be present in the section. The definition of a fault plane is ambiguous if the angle between the profile and the fault is not close to perpendicular. So there are some caveats, but on a sample of datasets from various regions and geological settings tested to date, results have added benefit by expanding the understanding of 3-D intricacies in the 2-D dataset.

Productivity/Efficiency

The argument is not whether the final interpretation could or could not have been made using conventional section-based tools. The value of 2-D seismic interpretation in a three-dimensional environment centers on the acceleration it gives to the initial phase of interpretation (understanding the general structural and depositional context), an acceleration that continues all the way through to the quality control of a final interpretation and the subsequent mapping and solid modeling that may ensue. Is this the watershed that has been so long awaited, finally convincing the industry-at-large that insofar as our planet is a three-dimensional body, we should be using three-dimensional tools in every aspect of our appraisal of the subsurface?

Comments