--> Value in Visualization, by Tracy J. Stark, #40133 (2004).

Datapages, Inc.Print this page

Click to view article as PDF

 

GCValue in Visualization*

By

Tracy J. Stark1 

Search and Discovery Article #40133 (2004)

 

*Adapted from the Geophysical Corner column in AAPG Explorer, June, 2004, entitled “Why Do We Need to Have Visualization?” and prepared by the author. Appreciation is expressed to the author, to Alistar R. Brown, editor of Geophysical Corner, and to Larry Nation, AAPG Communications Director, for their support of this online version.   

1STARK Research, Plano, Texas ([email protected])

 

Introduction 

How do you convince “a non-believer,” in a short article with only a few static figures, the need for visualization?  

Within some companies the value of integrating visualization techniques into the exploration workflow is well documented (pun intended). In these companies, the answer is along the lines of: “In order to get a well drilled, it is required by those whose money we are using to drill the well, since it has shown repeatedly to be an excellent risk reducer and a very good return on investment.”  

Visualization encompasses the software, hardware, and workflow combination that allows trained and experienced interpreters to investigate rapidly – and communicate to others – the internal heterogeneities of their 3-D data volume. All three components are important, but the workflow element is probably the most important. It is the workflow that allows you to answer the questions you need to ask of the data. If you choose the wrong workflow, some questions will remain unanswered, or poorly answered. The ability to use a particular workflow effectively depends upon the software package employed. For example, package A is better than package B for quickly comparing multiple attributes on a particular section of data. Yet, package B’s ability to opacity filter large volumes is significantly better than package A’s for identifying regional amplitude anomalies.  

 

uIntroduction

uFigure captions

uCase for visualization

uExamples

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uIntroduction

uFigure captions

uCase for visualization

uExamples

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uIntroduction

uFigure captions

uCase for visualization

uExamples

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uIntroduction

uFigure captions

uCase for visualization

uExamples

 

 

 

 

 

 

Figure Captions

Figure 1. Proportional slices between two closely spaced horizons show rapid changes in depositional patterns not evident in either the volume sculpted view (1), top horizon (2) or base horizon (12). (Seismic data courtesy of Seitel.)

 

Click to view sequence of slices between top and base.

Figure 2. End on volume rendering of a mid-angle stack volume, just showing the largest amplitudes. These data are from a ~300-square-kilometer survey over an undeveloped West Africa field. The identified hydrocarbon sands are between 2400 to 2600 ms. The flat event at 2100 ms was missed using conventional interpretation techniques.

Figure 3. Top down view of a volume-rendered, opacity-filtered, 20 ms thick slab around 2,100 ms illustrates the lateral continuity and extent of the flat spot shown in Figure 2. The five wells (yellow dots) drilled for the deeper target missed the flat spot, even though it covers about 20 percent of the survey, or 60 square kilometers. It appears as if the flat spot should continue off the southern and possible northern ends of the survey.

Figure 4. A single line near the southern end of the survey shows some of the AVO characteristics of the anomaly. The section on the left (A) is from the mid angle stack, while the section on the right (B) is from the full stack. Data between the two interpreted horizons was used to generate the crossplot in the lower right corner. When the anomalous crossplot amplitudes are posted back onto the sections they highlight the flat spot. This figure implies about a 120-meter hydrocarbon column height if the flat spot is a fluid contact.  

 

Click to view in sequence seismic line from midangle stack and full stack.

Return to top.

Case for Visualization 

The growing amount of available 3-D data is one reason an interpreter needs to employ visualization techniques. On a worldwide basis (excluding North America), numbers published by the IHS Energy Group in “First Break” indicate that from 1991 to 2002 the cumulative surface area covered by 3-D seismic data doubled every 2.5 years. By the end of 2001, the equivalent surface area of this cumulative seismic data was larger than the state of Alaska.  

If you assume that a full stack and three additional attribute volumes (such as a near-, mid- and far-stack volume) need to be interpreted, then by the end of 2002 these combined volumes could cover the entire United States. The speed, efficiency, completeness, and multiple workflows available from visualization tools are required to keep up with the growing data volumes. Moore’s Law is yet another reason you need to use visualization techniques, and continually upgrade your computers and graphics system. Moore’s Law implies that if you upgrade your visualization hardware every three years you will catch up on the growth of the 3-D data volume. Your graphics and computing power will increase by ~4x, while the data volume has only increased by ~2x.  

If your competitors are using visualization tools and continually upgrading their hardware and you are not, how much farther behind are you falling? However, the most compelling reason that you need to use visualization tools is that if you don’t, you probably will miss important features of your data volume, such as detailed depositional patterns and large regional flat spots.

 

Examples 

Figure 1 contains 12 sub-images showing changing depositional patterns. The first sub-image is of the volume-sculpted package. The other 11 images are proportional (stratal) slices through this package. Slices (3) through (11) were taken proportional distances from the top and bottom of the two bounding surfaces (2) and (12).  

The depositional patterns in the proportional slices are not apparent on either of the bounding surfaces, nor are they readily apparent in the volume-rendered sub-volume. Such details are important as they indicate possible flow boundaries or conduits as well as give clues where other sands might have been deposited.  

A volume rendering of the largest amplitudes found in the mid-angle stack over an undeveloped West Africa field is shown in Figure 2. This is an end-on view of a ~300-square-kilometer survey. The sands of the field, which are expected to contain in the range of 500 bcfg to 1 tcf gas within the limits of the 3-D survey, are between 2400 to 2600 ms.  

The flat spot at 2100 ms is hard to miss – however, at least seven different evaluation teams did not identify it as a drilling target. Clearly these teams did not generate a similar display. Most of these teams concentrated their efforts on the slightly deeper (250+ ms) objective known to contain hydrocarbons and believed to be part of a giant regional stratigraphic trap.  

How many of us don’t have or take the time to explore the volume above or below our current objective? Do you know what you are missing?   

A thick, volume-rendered, opacity-filtered time-slice around 2100 ms, again just showing the largest amplitudes, is provided in Figure 3. The five wells, drilled for the deeper target, missed hitting the 60-square-kilometer flat spot, even though it covers about 20 percent of the survey. The two wells that clipped the edge of the flat spot should be investigated for oil shows.  

The AVO nature of this event is illustrated with Figure 4. It is not a “textbook” example of a fluid contact. The “reservoir sands” are hard to discern on the vertical sections; they do not have the textbook behavior on either side of the “fluid contact,” and the contact appears locally to “change phase.” The “contact” also appears to have some localized “velocity pull down.” However, until the gathers are evaluated for proper processing and rock property modeling has been done, a hydrocarbon effect should not be ruled out. 

If the flat spot is a fluid contact, then optimistic approximations to the reservoir geometry and properties imply over five billion barrels of oil in place within the limits of the survey. Figure 3 indicates that the flat spot should extend beyond the survey limits. If the right visualization tools and workflow were utilized earlier in the project, slight modifications to two of the drilled well paths could have allowed testing of this potential reservoir. So is this a missed billion-barrel field? Only a well will tell.  

For those of you who still don’t think you need visualization, you might be right, for in the words of Edward Deming: “It is not necessary to change. Survival is not mandatory.”

Return to top.