One-way wave equation migration. (Images courtesy of ION)

Traditional methods of depth imaging require a mix of different migration algorithms and pose questions later in the workflow. This situation has been caused by the expense of running the algorithm and by the fitness for purpose of that algorithm. Recent advances allow reverse time migration (RTM) to be used throughout the model building sequence and in final imaging. Since the RTM algorithm gives superior image fidelity, this is the ideal tool for both model building and final imaging. This article shows that by using a single algorithm in both the model building and final imaging phases of a project, the cycle time can be collapsed. The two-way wave equation algorithm is proposed as a tool to help significantly reduce total project turnaround.

The Kirchhoff algorithm has been used as a significant part of imaging projects, first for small projects in 2-D and then for full volumes with additional iterations as the norm. Later, one-way wave equation migrations were introduced solely for the final migration; however, the same benefits could be useful for the interpretation of boundaries during the model building process such as the base salt and top salt. The complementary capabilities of one-way wave equation and Kirchhoff migrations led to workflows that would often involve both algorithms at multiple stages during the project. In some cases, the use of the combined volumes would be straightforward; however, the problem of combining interpretations would always be present, and the ambiguity was responsible for many delays.

For example, the imaging problem might contain problematic propagation angles or a problem in the identification of the correct arrival for a Kirchhoff algorithm, so one was never sure which image to believe.

Over the past few years we have seen enormous growth in the use of two-way wave equation migration or RTM for imaging. Initially, RTM tended to be confined to the final migration to highlight some of the model deficiencies, and it was apparent that using RTM during the model building phase would give additional and significant benefits. One reason this happened was the absence of any dip limitations or restrictions on the propagation path, leading to a more consistent image wherever events meet, such as against a salt body or fault. This consistency can be extended to any case where a reflector could be imaged in different ways, like a salt wall imaged by diving waves through sediments and via energy propagating through the salt itself. Since RTM is a single algorithm solution, it has become the workhorse for the majority of the steps in the imaging workflow. The RTM algorithm is robust enough to resolve the underlying geological issues, so there is no need to introduce a second or third image to the solution. Thus ambiguity is removed and cycle time is reduced.

A more accurate and valid model can clearly help interpreters add clarity to their assessment of the subsurface. By using a single optimal migration algorithm, the interpreter has the data necessary for the shortest time needed to reach an interpretation with confidence. A nice illustration of one of these effects can be seen in photo, which show the subsalt reflectors from the model used for the 2004 EAGE workshop on velocity model building. In this example, the one-way wave equation, even with the exact model, produces an image which appears to have a number of faults, which are seen below the salt body.

Even in a synthetic dataset imaged with the exact model, there can be imaging issues which could badly mislead the interpreter.

We have identified two benefits in the model building and interpretation workflow: the use of a single algorithm to address all imaging needs and the use of an optimal algorithm to best image complex structures. This is particularly true of those that require a wide dip range in a single location such as the imaging of variations in amplitude along an event or the termination of events. In both cases these will reduce the difficulties in making decisions on the data, thus enhancing our convergence to an accurate model and a confident interpretation with the associated reduction in cycle time.

Business drivers

In the last few years while we have been witnessing the growth and benefits of RTM, we have seen a number of changes in the business environment. These changes have been worldwide but are illustrated well in the Gulf of Mexico, where the high lease turnover rate has caused significant work for the oil companies bidding in that area. The same effect is present elsewhere where competition for assets has been significant. As a result, a large number of projects have required a rapid convergence to an interpretable volume. The interpretation staff does not have the luxury of conducting interpretations multiple times nor the willingness to work with anything other than an optimal image. In this highly competitive environment, anything other than very rapid project turnaround is unacceptable. The quality requirements have pushed strongly towards the extensive use of RTM, and the need to satisfy that quality requirement has ensured a focus on making the technology available in a timely manner.

Since the first introduction of RTM as a commercial product, there has been a massive focus on the optimization of the algorithm from the computer science and the geophysical perspective while maintaining and enhancing the integrity of the algorithm. The appropriate use of significant computational hardware and the optimal pairing of the computer science and hardware has brought us to the point where even the most aggressive turnaround times are feasible, including cases where the projects are as large as 1,900 sq miles (5,000 sq km). In fact, in most cases where large RTM imaging projects are performed, the migration is the most scalable part of the project and is often a smaller part of the elapsed time than most other parts of the total project turnaround such as the awarding of the project, transfer and preparation of the data, and the delivery and loading of the data to the interpretation environment.

Summarizing the above points, we see an environment where the drivers for imaging projects are optimal quality and minimal turnaround time. RTM provides the quality that no other algorithm can deliver in a single volume and, for many features, will surpass any combination of alternative migration algorithms. In addition, when turnaround time is critical regardless of project size with suitable hardware, an RTM solution is a viable option. If quality and turnaround time are the most important parts of the imaging needs, then RTM is the imaging algorithm that can deliver the best combination of the two.

Acknowledgment
Special thanks to BP for providing access to the included synthetic data examples.