Your account already exists. Please login first to continue managing your settings.
New methodologies promise to solve the mysteries of carbonate reservoirs.
There is an industry-wide awareness of the importance of carbonates in the commercial development and production of hydrocarbons. Yet there is a mystique about these that arises partially from the nature of the genesis and also from the higher velocity/density nature (even when geologically young), which has always made the seismic imaging a bit more challenging.
Simple traditional hydrocarbon indicators such as bright spots in carbonates, while not unheard of, are rare to say the least, and not necessarily as diagnostic as might be hoped.
Developments in seismic imaging over the last 40 years have been significant and along several avenues, responding principally to particular problem objectives having economic consequence.
For carbonates, a geologically and geometrically conformable nature, cyclicity in development, and sometimes seismically “thin” character can present difficult obstacles for imaging. Many advances address these issues, in particular better- and higher-resolution processing and improved seismic data presentations.
Zero-phase wavelet processing with better and more frequent velocity analyses were important first steps in improving resolution. Next, following the early lead of Roy Lindseth, the use of seismic inversion or velocity-scaling of seismic amplitudes offered great advantages for detecting velocity changes, especially in higher-velocity formations.
Extended, visual-dynamic range (EVDR) color displays then presented 20 to 25 times more information to the interpreter. Most recently, replacing standard signal processing methods by holography has produced highest-possible resolution (HPR) seismic imaging. Now the image frequency content can move beyond the source bandwidth, both on the low- and high-frequency end.
Typical imaging bandwidths developed from conventional seismic data acquisition using HPR imaging are three to five times broader than the result of usual signal processing methods. Such increased resolution often provides remarkable information about deposition, pore fluids, porosity, lithology and stratigraphy. This information could not be accessed using standard methods.
In addressing the approaches described here, there is no intent to diminish or understate the importance of auxiliary analyses and computations. Amplitude vs. offset (AVO), attribute calculations and novel displays all have contributed greatly to additional information extraction, particularly for carbonates starting from conventional survey results. These include identifying zones of higher fracture density and small fault trends.
Better images, better interpretations and other available analyses bode well for the future of finding, defining, and exploiting carbonate reservoirs. Reviewing examples spanning 40 years emphasizes this point.
The importance of data presentations cannot be overstated. Two issues should be raised: why inversions, and why such strange color combinations? Inversion displays are always preferable to reflectivity presentations, which today still remain the dominant format. A single key advantage favors inversion, especially for carbonate reservoir regimes.
In many instances, hydrocarbon-filled porosity is signaled by an interval velocity drop. This can usually be identified with ease. The inversion described here is a composite display. It is a blending of a simple trace integration and a coarse interval velocity model developed from very detailed moveout velocity information. No information from well logs is used in producing these displays.
Hence, the inversion displays treat hydrocarbon indications in high and low acoustic impedance circumstances without bias. With an effective use of highly contrasting colors, visual dynamic range is extended significantly, 20-fold or more as compared to conventional black and white presentations and fivefold or more as compared to typical color data displays. Such composite inversion displays look more like geology.
In conventional seismic imaging, four “ground rules” exist. There is a propagating waveform, all boundaries are sharp and well-defined, reflections are “scaled” copies of that waveform that broaden with reflection time as higher frequencies are lost, and resolution decreases with reflection time.
In holographic imaging, these are quite different. There is no propagating waveform; there is only propagating wave energy, the frequency of which really does not matter. Holographic images, after all, can be formed using only a single frequency.
Boundaries now are viewed as having realistic geologic properties, and while they could be “specular” in nature as always assumed in the original ground rules, this is not a common circumstance except for starved basins where flat, uniform, and vertically abrupt boundaries predominate.
Each boundary will then be represented by that frequency content necessary to represent its image. Low-energy boundaries such as shale/shale contrasts can have frequencies in the hundreds of Hz, while higher energy shale/sand boundaries top out at 60-160 Hz. Even higher energy reefal environments typically reach only 40-90 Hz. Modest as these limits seem, they are three to five times the values attained by standard signal processing approaches.
An important corollary of HPR is that since any frequency or range of frequencies can form the image, the loss of resolution with increased reflection time is far less pronounced. It diminishes somewhat as the imaging curves that need to be applied become less appropriate approximations to the real physics, and the requisite velocity information also is less precise. The bandwidths typically used for seismic acquisition offer quite robust holographic imaging.
Some of the earliest work on reshaping seismic wavelets to zero-phase rather than minimum-phase character was pioneered by a group at Geoquest International Ltd. spearheaded by J.P. Lindsey. While the basic precepts of classical seismic imaging remained unchanged, impressing a symmetric wavelet signature improved resolution, simplified the interpretation, and made reconciliation with well information via synthetic seismograms more definitive.
Hence, the achievements related to imaging carbonates from the 1970s through the 1980s became substantially more effective than earlier efforts.
In the Eastham State Farm gas field on the prison-farm grounds north of Houston, the main reservoir is the Glenrose C formation. Visibility on the conventional display (Slide 2 in attached PowerPoint file) is very limited. The C Member Reef (Slide 3) is about 60 m (200 ft) thick, 750 m (2,500 ft) across, and with a carbonate velocity some 1,830 m/sec (6,000 ft/sec) higher than the surrounding shales.
The gas-filled porosity in the reef is 750 m/sec lower in velocity than the tight carbonate rock. Gas effects in the carbonates are usually seen as velocity reductions in the range of 300-750 m/sec.
Condensates and oil typically reduce velocities in the range of 60-180 m/sec (200-600 ft/sec), which at the lower end introduces greater ambiguity.
Using the data skip zone of the prison building and the prominent Navarro reflection (top of the Cretaceous sediments) for reference, reef visibility can be compared. The composite inversion in this case is a simple trace integration blended with an interval velocity model developed from closely spaced velocity analyses (high-resolution computations by 1970 standards).
An inversion trace nearest the wellbore is aligned with acoustic impedance as computed from the well logs. The comparison is excellent. Also, the C member reef velocities may be mapped from the available 2-D seismic profiles (Slide 4). Enhanced information content as elicited via the wavelet processing often overloaded the visibility capacity of standard seismic displays. Also, the inversion display appeared to provide spatially consistent results, which correlated well with data from wells.
In 2007, this play was revisited seeking thinner Glenrose targets. When compared to standard imaging, holographic imaging accesses frequencies to 500 Hz and has twice as many data traces (Slide 7). For this case, the data acquired was of excellent quality, and the standard processing was very well done.
Even on the standard processing there are strange, nearly vertical features that can be seen in several places on the section. These are understood to be conduits by which fluids move vertically in the subsurface, and these are extraordinarily common features.
Often, when the fluids encounter faults, they “hitch” rides along these conduits for varying distances. These are seen more clearly on the holographic view with the broadened frequency range. Many of the newer attribute presentations also clearly show and even emphasize these conduits, which have been long overlooked.
It is the bimodal character of the frequency spectrum of the holographic result that may at first seem surprising. Each boundary responds with the frequency content necessary to represent its image.
Low-energy shale sections have many low-amplitude boundaries that are quite well defined and give rise to these high frequencies. In fact, these are at the heart of the more pronounced anisotropy (vertical versus horizontal) that is characteristic of such lithologies.
The higher energy sand boundaries are not nearly so well defined since they have greater and more rapid horizontal variations and vertical grading. In fact, the highest frequencies reached for such boundaries are actually some quantitative measure of the depositional energy.
There are important lessons here. First, the full data content of the holographic image requires a more effective display. Second, the importance of the fluid conduits must not be underestimated.
Third, access to higher frequencies will image vastly greater detail, but expectations must be in line with geological reality.
Improved matches to synthetic seismogram calculations will not be found in high-energy environments, and in fact these will be seen to limit the interpreter’s ability to develop effective ground-truth correlations.
In 2008, a project was undertaken in McMullin County, TX. A small 3D dataset was imaged using HPR. When shown as a standard data presentation, the HPR image appears unremarkable (Slide 13), as the display format masks most of the desired information content.
However, the composite inversion image of Slide 14 is quite overwhelming owing to its small size and the density of the available information. An anomalous feature is seen in the Edwards formation above this image, even though the area is beyond the traditionally mapped Edwards reef trend. The logical step to investigate this feature would be to “zoom” in on it while exploring some more unorthodox display formats.
Using only those inversion velocities approximately characterizing Cretaceous carbonates, the reefal feature is viewed in close-up in Slide 15. The cyclic development of the feature becomes eminently clear, as do the front and back reef environments.
The survey profile can be related to a time slice mapped through the survey cube. Here features map readily and show classic carbonate geometries including barriers, surge channels, and even atolls (Slide 17).
Such interpretations would likely not be feasible using standard seismic imaging approaches. Further analyses using attribute calculations and AVO methods would be further enhanced having first applied holographic data treatments.
In the 1980s, the Austin Chalk was a quite common exploration play over a fairly large swath of Texas. Production from this objective was typically quite prolific but short-lived, often depleting in just a few months but producing good economic returns nevertheless. About this time, horizontal drilling technology was evolving and growing in its use and sophistication.
With such exploitation, Chalk wells would still produce at extraordinary rates and again for rather short periods, but the additional revenues would more than offset the added drilling costs. Slide 26 shows a conventional display of a seismic profile encompassing the Austin Chalk, with a small area of the Chalk highlighted.
Wavelet processing with dense, high-resolution velocity work had been applied. An early (1980) composite inversion (developed from wavelet processed data) is shown in Slide 27 with the then-contemporary well control. An EVDR color scale has been used.
The Chalk in Slide 27 in seismic terms is seen as having three members. The lower member has the highest velocities; and the middle member has the lowest values. It is fortuitous that in the middle member, velocity drops of over 330 m/sec (1,000 ft/sec), indicating gas-filled porosity, could be seen as prominent orange zones.
In fact, the velocity drops in each member could be readily recognized individually as decreases of 330 m/sec or greater. Velocity drops in the high velocity and very consistent Buda could also be noted.
Using information like this, dozens of Austin Chalk wells were successfully drilled with vastly improved economics as the vertical position within the section and lateral extent of the porosity developments were visible and known in advance. Slide 28 compares the seismic visibility of the original processed data and the counterpart composite inversion. It would not at all be feasible to perform the same interpretations using the original display. Slide 29 shows a profile from a 3D survey over an area of South Texas where Austin Chalk is again present. This holographic imaging and inversion was performed in 2011.
The figure shows an interpreted profile from the composite inversion cube. The horizontal red line indicates a time-slice to be viewed below. Nominal interpretative boundaries are shown indicating top of Chalk and base of Chalk, which is the Eagle Ford top. As before, velocity variations within the Chalk are visible.
The Buda, Del Rio, and Georgetown below are readily recognized. A well tie is included for reference. Note that synthetic seismograms cannot be used to make effective ties because of the restricted frequency content and inability to track the imaged detail.
The Austin Chalk is now quite different and shows what appear to be several cycles of reefing.
Two separate reefs are positioned on a discernible slope change on the Eagle Ford. Chalk velocities are at about 6,700 m/sec (22,000 ft/sec) in the lower members, but only 4,700 m/sec (15,500 ft/sec) in the upper cycles, which possibly may be detrital material. These are still significantly above the 3,650 to 3,950 m/sec (12,000 to 13,000 ft/sec) velocities characteristic of the shales at this depth.
It is interesting to note the level of detail presented by comparing Slides 27 and 29. Rapid insights can be gained by viewing the time-slice noted previously at the horizontal red time line. This is shown in Slide 30. Profile A’A is shown on this display.
The two reefs are related by an atoll feature that has a very distinct geometry. Such an environment can explain the cycles of possible detrital material that we see above the harder Chalk. Within the body of the reef itself, velocities as low as 5,150 m/sec (17,000 ft/sec) may be seen.
These reefs would likely not be seen in conventionally treated seismic data even if present as Slide 28 indicates. The Chalk counterpart in South America is known to show reefing, but these reefs are visible using standard approaches. Slide 31 shows a photo of a modern-day atoll (Woleai) taken from the Internet. The similarity is striking.
Slide 32 is a standard seismic data presentation of the parent data from which the composite inversion of Slide 29 is developed. In Slide 33, the standard data presentation is compared directly to the composite inversion. The two reefs would not be readily identified in the conventional presentation even of the HPR data.
Slide 34 from the same composite inversion cube focuses on the Eagle Ford formation. The harder parts of the Eagle Ford show velocities of about 5,150 to 5,760 m/sec (17,000 to 19,000 ft/sec) and must be very “limey.” Velocities at the lower end range from 4,240 to 4,550 m/sec (14,000 to 15,000) ft/sec.
All of these trends can be mapped and are conformable with the regional dips, for the most part. Fracture imaging has not been attempted, but if present. these would be more effectively detected with the much superior HPR image data. Of course, most typical attribute calculations may be applied to this improved data.
Progress providing greater seismic visibility of subsurface formations, particularly carbonates, begins to erase much of the mystery surrounding the genesis and development. Improving the images has resulted in bringing them closer to the geological models that are typically formulated.
Advancing seismic imaging and applying geologic reasoning are worthy goals in the quest for finding and producing the hydrocarbons so vital to our society. The improvements cited here certainly do not represent the only progress, but holographic imaging and improved seismic displays can open new doors of opportunity and understanding.
Similar progress could be shown for clastic environments, subtle stratigraphy, small faults and deeper objectives.
It has already been demonstrated that standard seismic data presentations cannot show the detail that can now be drawn from seismic data. Similarly, the synthetic seismogram is unable to address the correlation with log information for want of frequency content. It also embodies constraints not present in holographic imaging.
Work using the methods described has been applied in most of the major oil and gas provinces globally, both onshore and offshore.
Editor's Note: This is the online, expanded version of the article "Revealing The Secrets of Carbonate Formations" that appeared in the January 2012 issue of Hart's E&P magazine.