Abstract

3D seismic surveys are replacing conventional 2D seismic surveys as the accepted technique for seismically-driven exploration and development. The Encyclopedic Dictionary of Exploration Geophysics defines a three dimensional survey as: “A survey involving collection of data over an area with the objective of determining spatial relations in three dimensions, as opposed to determining components along separated survey lines. Various field arrangements are used. The data from such a survey constitute a volume which can be displayed in different ways.”

In other words, a three dimensional survey provides a finely-spaced volume of information which, after a 3D migration technique, leads to a more nearly correct subsurface image. However, this technique is easily compromised if due care and attention is not given to the survey design.

There are many aspects to be considered in designing a successful 3D survey, some of which are:

Economics

Fold

Distance Distribution

Azimuth Distribution

Economics is probably the most important factor in a 3D design. The importance of getting a satisfactory return on investment in a seismic and drilling program can not be understated, but there is a temptation to increase the return by decreasing the seismic effort in the field and in data processing. Common ways in the field to reduce acquisition costs are to reduce the number of source points or to reduce the sweep effort on vibroseis surveys. In processing, it is commonplace not to perform refraction statics estimation or dip moveout. In many instances, this reduction may be acceptable, but there have been cases where these economies have compromised the results of the survey and have had a detrimental effect on the overall results of the project.

Fold refers to the number of source-to-receiver midpoints that lie within a subsurface cell or bin. In processing, the individual contributors in a bin are summed together to form a common-midpoint stack trace. This summation has a number of benefits, a key one being the improvement of the reflected signal-to-noise ratio. In general, the higher the fold, the better the resulting data quality. But increasing the fold on a survey will result in an increased cost. There is usually a point of diminishing return when the improvement in data quality arrived at by increasing the fold is not worth the expense incurred by the extra field effort. In general, the fold on a 3D survey can be less than the fold on a 2D survey in the same area. Fold is also an important factor because as the number of traces in a bin increases, so does the accuracy of the velocity analysis and some surface-consistent processes such as the estimation of residual statics.

Distance distribution is also referred to as offset distribution. Both terms refer to the distribution of traces within a common mid-point as a function of source-to-receiver distance. Usually, surveys are designed with the maximum source-to-receiver distance or offset being approximately equal to the greatest depth of interest. This is to maximize the accuracy of the stacking velocity analysis which, in turn, is used to construct the velocity field used for the 3D migration. For example, in a 20-fold survey where source-to-receiver offsets lie between 0-10,000 ft, ideally the design should yield one trace with an offset between 0-500 ft, one with an offset between 501-1000 ft, one with an offset between 1001-1500 ft, and so on. In practice, this is rarely achieved but it is still an important consideration when evaluating a 3D design.

Azimuth distribution is used in reference to the geographical orientation (N-S, NNE-SSW, etc.) of individual source-to-receiver traces. In some instances, this may be an important factor such as in areas of complex geology or moderate-to-extreme velocity anisotropy where the distortion of the seismic wavefield varies according to the direction of propagation. Surveys can be designed to accommodate a wide range of azimuths or a narrow range. Typically, 3D surveys in the Permian basin are designed with little regard to azimuthal concerns, but there have been instances when this has caused problems in the quality of the final migrated volume.

Probably the most commonly used 3D technique in the Permian basin today is the orthogonal swath. In this method, a number of lines of geophone groups are laid out parallel to each other at intervals of perhaps a quarter of a mile. Then a line of source positions is shot in the center of the geophone lines at right angles to them. The source line typically extends for one geophone line interval, but on some surveys it may extend further. This layout is then “rolled” in either the geophone line direction or in the source line direction and another source line is collected. The number of lines of geophone groups, the number of live geophone groups per line, and the source-line interval are all determined from the fold and offset requirements of the survey.

A very common design utilizes 6 lines of geophone groups spaced at quarter-mile intervals with 80 live geophone groups per line. The source-line interval may be a quarter or a third of a mile and this layout results in a very even fold distribution. Offset and azimuth distributions are fairly even also and this design has proved to be very economical to shoot. A recording system that has 480 live channels is required for this and it is not uncommon for designs to require 600-800 live channels. As a consequence, there has been an influx into the region of modern recording systems capable of recording 1000 channels or more.

A few years ago when the use of 3D in the Permian basin began to increase, there were few of the modern seismic recording systems available in the region. Also, the economics and benefits of 3D surveys were not widely appreciated. As a result, a number of the surveys recorded at that time were designed in a way that would be viewed somewhat critically today. One such example is a survey in Scurry County, just southwest of Snyder. The technique used a 240-channel “patch” that involves laying out 240 groups of geophones and then shooting through it. The patch is then picked up and laid over a different part of the survey and shot through again. This was repeated until the survey area was covered. The source and receiver lines were laid at half-mile intervals, but this varied around cultural obstacles. The resulting fold has an uneven distribution over the survey area and reaches a maximum of 7 in places. The offset distribution is far from even as well. However, the survey did lead to the identification of a number of potential drilling locations, most of which have been successfully completed.

As the availability of the newer recording systems improved, and as 3D surveys became more widespread, the design of a typical 3D survey evolved toward the method described earlier. However, there is still a requirement to keep the cost as low as practical. One method that has been used incorporates a rolling swath. However, the source interval is doubled (from 220 ft to 440 ft for example). To retain the desired subsurface bin size, alternate source lines are moved 220 ft so that the source positions occupied are staggered from line to line. The fold plot for such a design shows that the fold is evenly distributed except for some stripes at the edges that are a result of the staggering. The fold is lower also but this is a very attractive design because the number of source points is halved for the survey and there is a significant cost saving. The inline sections from such a survey will give a good indication of data quality with little evidence of any anomalies caused by the design. The inline direction is used to describe the direction of the geophone lines and the crossline direction is orthogonal to that, in this case in the direction of the source lines. However, the crossline sections may show some anomalies, often a lack of continuity in the data that is not evident on the inline sections. A review of time slices through the volume may be especially revealing. There may be linear anomalies in the slice that tie in with the direction of the recording geometry. As a side note, the time slice is an excellent way of reviewing a 3D volume and whenever anything shows on the slice that aligns with the geometry of the survey, it should be investigated as to its cause.

It is instructive to generate two-fold plots for this survey to help understand the cause of the anomalies. The first plot uses only the odd-numbered source lines and the second uses only the even-numbered source lines. Careful scrutiny of the plots shows that the reflection points covered by the records from the odd-numbered source lines do not occupy the same subsurface positions as the reflection points covered by records from the even-numbered source lines. In effect, two different surveys have been recorded that are interleaved. While they occupy many of the same surface geophone positions, they do not occupy the same subsurface positions. This is referred to as decoupling. Depending on the cause and severity of the decoupling, it may be compensated for in the processing center.

A similar situation can arise in surveys where the source lines are run parallel to the receiver lines. While the fold plot appears acceptable, a review of the inlines, crosslines, and time slices may show some lack of continuity in the crossline direction and some linearities on the time slices. It turns out that while there is overlap on the surface from swath to swath, there may not be any overlap in the subsurface giving rise to insufficient coupling.

Designs using source lines parallel to the receiver lines are not very common, in part because of the problem described above. However, in some circumstances this design may be the best way to collect usable data. One such example is in an area where there is a shelf edge present in the shallower part of the section. This shelf edge gives rise to a rapidly changing geology and velocity field. Two surveys were shot in this area, one with an orthogonal swath technique and the other with source lines parallel to the receiver lines. The orthogonal swath survey showed a lack of data continuity at the objective while the other survey showed the objective very clearly. The orthogonal swath had a very wide range of source-to-receiver azimuths whereas the other survey had a very narrow range of source-to-receiver azimuths. Due to the presence of the shelf edge, the survey with the wide range of azimuths is affected by rapid velocity changes and ray-path distortion in the subsurface for which processing algorithms were not able to compensate. The survey with the restricted range of azimuths remains relatively unaffected and provides a very interpretable product.

Even a well-designed, acquired, and processed 3D survey can present a challenge for the interpreter. Typically, 3D migrated volumes are in time and it is left to the interpreter to convert to depth. Rapid velocity changes, such as those giving rise to velocity “pull ups”, may not be readily extracted from the seismic data without additional analysis which can add both time and expense to the survey. Even a depth migration or a conversion to depth of a time-migrated data set requires a very detailed knowledge of the velocity field in the subsurface. This knowledge can only be obtained by careful analysis of the seismic data volume and integration of that analysis with other information such as checkshot surveys and sonic logs.

The 3D seismic survey has become an accepted technique in reducing the risk inherent in exploring for hydrocarbon reserves. 3D has become economically feasible for many oil companies and the results have shown the ability of 3D to produce far more detailed pictures of the subsurface than were previously available. As with any new technique, the effectiveness of the method can be compromised, but as familiarity with the technique increases, the information obtained from 3D surveys becomes more reliable. The effectiveness of 3D seismic surveys has also impacted the financial aspect of the oil and gas industry contributing to an improved return on the investment in exploration and development projects.