16 Oilfield Review

Seeking the Sweet Spot: Reservoir and Completion Quality in Organic Shales

Placement of horizontal wells in shale reservoirs can be a costly and risky business

proposition. To minimize risk, operators acquire and analyze surface seismic data

before deciding where to drill.

Karen Sullivan GlaserCamron K. MillerHouston, Texas, USA

Greg M. JohnsonBrian ToelleDenver, Colorado, USA

Robert L. KleinbergCambridge, Massachusetts, USA

Paul MillerKuala Lumpur, Malaysia

Wayne D. PenningtonMichigan Technological UniversityHoughton, Michigan, USA

Oilfield Review Winter 2013/2014: 25, no. 4. Copyright © 2014 Schlumberger.For help in preparation of this article, thanks to Alan Lee Brown, Raj Malpani, William Matthews, David Paddock and Charles Wagner, Houston; Helena Gamero Diaz, Frisco, Texas; and Ernest Gomez, Denver.sCore is a mark of Schlumberger.

In the late 20th century, E&P geoscientists began to consider shales in a new light. Although production from shales had been established in the early 1800s, operators considered shale for-mations mainly as source rocks and low-perme-ability seals for conventional reservoirs. However, during the 1980s and 1990s, operators showed that the proper application of horizon-tal drilling combined with multistage hydraulic fracturing could make organic shales produc-tive, spurring the exploitation of organic shales as self-sourced reservoirs.1 Despite the success-ful development of the Barnett and Haynesville shales in the US, the industry quickly realized that not all shales are viable targets for eco-nomic hydrocarbon production, and operators sought technologies that could identify appro-priate targets for development.

Shale formations that offer the best potential require a unique combination of reservoir and geomechanical rock properties; such formations are relatively rare. Organic shales have extremely small pore size and ultralow matrix permeability, which makes these unconventional resource plays fundamentally different from most conven-tional reservoirs.2 Furthermore, because hydro-carbon migration paths tend to be short, productive zones of shale reservoirs may be con-fined to a certain area within a basin or restricted to a stratigraphic interval.

The two factors that determine the economic viability of a shale play are reservoir quality and completion quality. Good reservoir quality (RQ) is defined for organic shale reservoirs as the ability to produce hydrocarbons economically after hydraulic fracture stimulation. Reservoir quality is

a collective prediction characteristic that is largely governed by mineralogy, porosity, hydrocarbon saturation, formation volume, organic content and thermal maturity.

Completion quality (CQ), another collective prediction attribute, helps predict successful res-ervoir stimulation through hydraulic fracturing. Similar to RQ, CQ largely depends on mineralogy but is also influenced by elastic properties such as Young’s modulus, Poisson’s ratio, bulk modulus and rock hardness. Completion quality also incor-porates factors such as natural fracture density and orientation, intrinsic and fractured material anisotropy and the prevailing magnitudes, orien-tations and anisotropy of in situ stresses.

To be successful in today’s shale plays, opera-tors drill horizontally within reservoir strata that possess superior RQ and CQ. Stimulation treat-ments are most effective when the induced frac-tures remain propped open, thereby exposing the reservoir to a large fracture surface area and allowing fluids to flow from the reservoir to the wellbore, effectively raising the reservoir’s sys-tem permeability.3

Operators judge the quality of a hydraulic fracture completion design based on a postjob evaluation of data from sources such as micro-seismic monitoring of hydraulic fracturing, flowback tests and initial production to deter-mine how effectively and efficiently the reser-voir was stimulated.

Ideally, an operator places horizontal wells within shale intervals with favorable geologic characteristics and high RQ and CQ and without geohazards.4 Retrospective studies have demon-strated that this strategy would have resulted in

1. Boyer C, Kieschnick J, Suarez-Rivera R, Lewis RE and Waters G: “Producing Gas from Its Source,” Oilfield Review 18, no. 3 (Autumn 2006): 36–49.

Boyer C, Clark B, Jochen V, Lewis R and Miller CK: “Shale Gas: A Global Resource,” Oilfield Review 23, no. 3 (Autumn 2011): 28–39.

2. Nelson PH: “Pore-Throat Sizes in Sandstones, Tight Sandstones, and Shales,” AAPG Bulletin 93, no. 3 (March 2009): 329–340.

3. System permeability refers to the overall permeability of the effective reservoir volume and is the sum of contributions from matrix permeability and natural fracture permeability. Matrix permeability in shale reservoirs ranges from 0.1 to 1,000 nD. Natural and induced fractures are necessary for wells in these formations to be economically productive.

4. Miller C, Waters G and Rylander E: “Evaluation of Production Log Data from Horizontal Wells Drilled in Organic Shales,” paper SPE 144326, presented at the SPE North American Unconventional Gas Conference and Exhibition, The Woodlands, Texas, USA, June 14–16, 2011.

Cipolla C, Lewis R, Maxwell S and Mack M: “Appraising Unconventional Resource Plays: Separating Reservoir Quality from Completion Effectiveness,” paper IPTC 14677, presented at the International Petroleum Technology Conference, Bangkok, Thailand, February 7–9, 2012.

Winter 2013/2014 17

18 Oilfield Review

as much as a tenfold increase in production (below).5 Determining where the best RQ and CQ coincide is therefore an exploration effort, and the best technique to enhance the exploration effort, before drilling the initial well, is interpre-tation of surface seismic data. Recent studies have indicated that seismic interpretation is use-ful for defining production sweet spots within organic shale plays.

In this article, we describe a systematic and strategic approach for using surface seismic data to identify reservoir sweet spots in shale resource plays, starting with basinal and regional RQ and progressing toward local RQ and CQ. Case studies from the Arkoma, Delaware and Williston basins in the US demonstrate how reflection seismic data provide the key to deter-

mining where a resource play may exist and where RQ and CQ are highest.

Mudstone CharacteristicsGeologists define shales as mudstones that exhibit fissility—the ability to split easily, like a deck of playing cards, into individual laminae. The oil and gas industry typically considers resource plays as gas- or liquid-producing “shales.” However, it would be more accurate to call them mudstones or mudrocks, because these “shales” often are not fissile.

Mudstones dominate the sedimentary record and make up roughly 60% to 70% of Earth’s sedimentary rocks.6 They are fine-grained sedi-mentary rocks composed of clay- and silt-size par-ticles with diameters of less than or equal to

62.5 micron [0.00246 in.].7 These small particle sizes result in low permeability; poor sorting—the mixing of various grain sizes—can further reduce both permeability and porosity.

Mudstones have a complex mix of organic mat-ter and clay minerals—illite, smectite, kaolinite and chlorite—along with quartz, calcite, dolomite, feldspar, apatite and pyrite. Geologists with Schlumberger recently introduced the sCore ter-nary diagram mudstone classification scheme, which is built on relationships established between core and log, using clay, QFM (quartz, feldspar and mica) and carbonate as the corner points. The sCore diagram defines 16 classes of mudstones and can classify a sample as an argilla-ceous (clay-rich), siliceous or carbonate mud-stone. This classification scheme allows geologists and engineers to examine empirical relationships between mineralogy and factors that influence the RQ and CQ of mudstones by overlaying points that include indications of RQ, CQ or both (next page).8 Productive mudstones most sought by oil compa-nies tend to be dominated by nonclay minerals, principally silicates and carbonates, and therefore lie in the lower part of the diagram, away from the clay point; higher RQ and CQ rocks are near the edges of the triangle.9

Several factors control the physical proper-ties of mudstones: the mineralogy and propor-tions of grains, the fabric of the originally deposited muds and the postdepositional pro-cesses—such as resuspension, redeposition, dia-genesis, bioturbation and compaction—that convert mud into rock.10 Mudstones tend to be highly heterogeneous, and this heterogeneity can vary horizontally and vertically, originating from the sequence of depositional environments and tectonic regimes that prevailed as the mud strata stacked up through geologic time.

An individual layer of mud, called a lamina-tion, is typically about a millimeter thick. Laminations stack up to form laminae sets, called beds. Beds in turn stack up to form bed sets that group together into members and then into geo-logic formations. The mineral and organic com-position of each layer depends on the sequence

> Best 12-month production results. This 50-mi2 [130-km2] area of the Barnett Shale play in northwest Tarrant County, Texas, USA, shows the first year’s gas production for more than 650 horizontal wells. Black dots represent surface locations of well pads, which may service multiple wells. Areas of warm colors (top of scale) are production sweet spots and areas of cool colors (bottom of scale) are not. (Adapted from Baihly et al, reference 5.)

120

280

200

360

440

520

600

Gas

prod

uctio

n, M

Mcf

0 km

0 mi 2

2

N

5. Baihly JD, Malpani R, Edwards C, Han SY, Kok JCL, Tollefsen EM and Wheeler CW: “Unlocking the Shale Mystery: How Lateral Measurements and Well Placement Impact Completions and Resultant Production,” paper SPE 138427, presented at the SPE Tight Gas Completions Conference, San Antonio, Texas, November 2–3, 2010.

6. Aplin AC, Fleet AJ and Macquaker JHS: “Muds and Mudstones: Physical and Fluid-Flow Properties,” in Aplin AC, Fleet AJ and Macquaker JHS (eds): Muds and Mudstones: Physical and Fluid-Flow Properties. London: The Geological Society, Special Publication 158 (1999): 1–8.

7. A micron, or micrometer, is equal to one millionth of a meter or one thousandth of a millimeter. Its abbreviation

is μ, μm or mc. In Imperial units, a micron equals 3.937 × 10–5 in.

8. For more on the sCore classification scheme: Gamero-Diaz H, Miller C and Lewis R: “sCore: A Mineralogy Based Classification Scheme for Organic Mudstones,” paper SPE 166284, presented at the SPE Annual Technical Conference and Exhibition, New Orleans, September 30–October 2, 2013.

9. Loucks RG and Ruppel SC: “Mississippian Barnett Shale: Lithofacies and Depositional Setting of a Deep-Water Shale-Gas Succession in the Fort Worth Basin, Texas,” AAPG Bulletin 91, no. 4 (April 2007): 579–601.

Passey QR, Bohacs KM, Esch WL, Klimentidis R and Sinha S: “From Oil-Prone Source Rock to Gas-Producing Shale Reservoir—Geologic and Petrophysical

Characterization of Unconventional Shale-Gas Reservoirs,” paper SPE 131350, presented at the CPS/SPE International Oil and Gas Conference and Exhibition in China, Beijing, June 8–10, 2010.

Lash GG and Engelder T: “Thickness Trends and Sequence Stratigraphy of the Middle Devonian Marcellus Formation, Appalachian Basin: Implications for Acadian Foreland Basin Evolution,” AAPG Bulletin 95, no. 1 (January 2011): 61–103.

10. Aplin AC and Macquaker JHS: “Mudstone Diversity: Origin and Implications for Source, Seal, and Reservoir Properties in Petroleum Systems,” AAPG Bulletin 95, no. 12 (December 2011): 2031–2059.

Winter 2013/2014 19

> sCore classification tool. In the clockwise direction, the corners of the sCore ternary diagram (top left) are clay, carbonate and quartz plus feldspars plus micas (QFM). The diagram defines 16 classes of mudstones based on mineralogy. The mudstones (top right) sought by oil companies tend to have less than 50% clay. In the Wolfcamp Shale (middle), siliceous mudstones exhibit high RQ and CQ. In the Eagle Ford Shale (bottom), carbonate mudstones have high RQ and CQ. In these examples, RQ is directly proportional to effective porosity and CQ is inversely proportional to the stress gradient of the minimum in situ principal compressive stress.

Carbonate-richsiliceous mudstone

Silica-richcarbonate mudstone

Silica-dominatedlithotype

QFM 20

20

50

50

80

80

80

50

20

10

80

50

20

10

10

10

20

50

80

10

20 50 80 10

Clay Clay

Carbonate

Shale plays

Barnett

Marcellus

Haynesville

Fayetteville

Eagle Ford

Wolfcamp

QFM Carbonate

Mixedsiliceousmudstone

Argillaceous-siliceousmudstone

Mixedmudstone

Argillaceous-carbonatemudstone

(marl)

Mixed argillaceousmudstone

Clay-dominatedlithotype

MixedcarbonatemudstoneCarbonate-

siliceousmudstone

Clay

-rich

silice

ous m

udsto

ne

Silic

a-ric

h

argil

laceo

us m

udsto

ne Carbonate-rich

argillaceous mudstone

Clay-rich

carbonate mudstone

Carbonate-dominatedlithotype

sCore Classification

QFM QFM

80

50

20

10

80

50

20

10

20

50

80

10

20

50

80

10

10 20 50 80 10 20 50 80

Clay Clay

Carbonate

High RQHigh CQ

Carbonate

Effective porosity, fraction

Incr

easi

ng R

Q

0.150

0.120

0.090

0.061

0.031

0.001

Minimum stress gradient, psi/ft

Incr

easi

ng C

Q

0.62

0.65

0.68

0.71

0.74

0.77

Wolfcamp Shale

QFM QFM

80

50

20

10

80

50

20

10

20

50

80

10

20

50

80

10

10 20 50 80 10 20 50 80

Clay Clay

High RQHigh CQ

CarbonateCarbonate

Minimum stress gradient, psi/ft

Incr

easi

ng C

Q0.83

0.85

0.88

0.90

0.93

0.95

Effective porosity, fraction

Incr

easi

ng R

Q

0.080

0.064

0.048

0.033

0.017

0.001

Eagle Ford Shale

20 Oilfield Review

or history of geologic conditions of the area through time. Geologists use the principles of stratigraphy to decipher this geologic history.11

Layering has a particular effect on some rock properties: It is a fabric that causes anisotropy.12 A rock is anisotropic if its properties vary with direction.13 A consequence of layering is that the composition, size, shape, orientation, packing and sorting of the particles in the layer tend to vary more quickly perpendicular to, rather than parallel to, layers. As a result, rock properties tend to vary with direction. They are different if measured layer parallel than if measured layer perpendicular. Another aspect of rocks that can lead to anisotropy is the presence of roughly par-allel open fracture networks, which can control the efficiency of reservoir stimulation. Because anisotropy is observable in seismic data, geophys-

icists are able to characterize it for geologists and engineers to use in their various geologic, geomechanical and fluid-flow models of the pro-spective reservoir (below).

Mudstones play an important role in a petroleum system. Their small grain sizes and sorting characteristics contribute to their char-acterization as low-porosity rocks with low to ultralow permeability and high fluid-displace-ment entry pressures. Consequently, when mud-stones are in the correct stratigraphic and structural location and configuration, they form the seals that cap and delimit conventional hydrocarbon reservoir geometries.

Some mudstones are characterized as organic rich, and these have been viewed historically as the source rocks that, through secondary migra-tion, supply hydrocarbons to adjacent and remote

conventional and unconventional continuous res-ervoirs. These same organic-rich mudstones have also proved to be self-sourcing reservoirs, yielding hydrocarbons that have been expelled and under-gone primary migration to be stored within the source rocks themselves.14 For example, the Eagle Ford Shale in south Texas, USA, is a mudstone that sources the prolific Austin Chalk fractured reservoir, which has been explored and produced for more than 80 years. Now, operators recognize the Eagle Ford Shale itself as a reservoir capable of producing oil, condensate, wet gas and dry gas that simply never left the source rock.15

Not all mudstones contain sufficient hydro-carbons to be considered potential reservoir rocks. Mudstones are defined as organic rich if their total organic carbon (TOC) concentration is greater than 2 weight %.16 The preservation and

>Mudstone layering at many scales. Layering may be observed in photographs of outcrop, core and thin section. The Eagle Ford Shale outcrop (left) is in Lozier Canyon in Terrell County, Texas. The images of core (plain and ultraviolet light, center) and thin section (original and close-up, right) are of lower Eagle Ford Shale from the BP-Schlumberger Lozier Canyon number 1 well. The 2-ft [0.6-m] core length was cut from depths 226 to 228 ft [68.9 to 69.5 m]. The thin section is of fossiliferous siliceous-calcareous mudstone and has a mineralized fracture running along its right side, which has been stained with potassium ferricyanide and alizarin red S dye to distinguish carbonate minerals. In the close-up view from the thin section, there is evidence that the fracture propagated, stopped and restarted along a different path. (Outcrop photograph courtesy of Karen Sullivan Glaser. Core and thin section images courtesy of Schlumberger and BP America Incorporated.)

0 mm 5

0 mm 0.5

0 m

0 ft 3

1

226

227

Core

leng

th, f

t

Plain Light Ultraviolet Light0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

Winter 2013/2014 21

richness of organic matter depend on the relative rates of its production, dilution and destruction (right).17 Inorganic matter deposited at the same time as organic matter will dilute organic matter concentration. Destruction of organic matter occurs through bacterial consumption, by oxida-tion reactions at shallow depths and by deeper thermally activated reactions, which transform part of the organic matter into oil and gas before it ultimately changes to graphite, or dead carbon. The primary portion of organic matter in source rock is kerogen, which is insoluble in common organic solvents; the other portion is bitumen, which is soluble.

Kerogen has petrophysical characteristics that differ significantly from those of the mineral con-stituents in shale, and these characteristics affect the overall bulk properties of the reservoir rock. For example, depending on kerogen type and maturity, the density of kerogen can vary from 1.1 to 1.4 g/cm3, considerably less than the bulk den-sity of its shale host rock.18 Consequently, the bulk density of organic-rich shales appears lower (as if the shale has a higher porosity) than that of shales containing lower concentrations of kerogen.

The distribution of kerogen varies from iso-lated particles dispersed through the mudstone matrix to lenses and sheets aligned with mud-stone laminae. Investigators have found that kerogen particles contain secondary porosity that likely formed during thermal maturation.19 This organic porosity occurs as nanopores, defined as less than 1 micron in diameter.

Kerogen fabric affects the physical properties of organic-rich mudstones. When the organic con-tent is high and the kerogen forms interconnected layer-parallel networks through the mudstone

frame, the organic porosity may be sufficient for hydrocarbon storage and for providing permeabil-ity to gas and liquid hydrocarbon in an otherwise extremely low-permeability matrix.20

> Organic matter. The thin section (left), which has been stained with potassium ferricyanide and alizarin red S dye on its left side, is of calcareous pelletal mudstone. In the close-up view (right), the layer is composed of planktonic foraminifera (white and pink), flattened fecal pellets (reddish brown) and organic matter (black). (Core and thin section images courtesy of Schlumberger and BP America Incorporated.)

0 mm 5

0 μm 200

11. Neal J, Risch D and Vail P: “Sequence Stratigraphy— A Global Theory for Local Success,” Oilfield Review 5, no. 1 (January 1993): 51–62.

12. Fabric refers to the spacing, arrangement, distribution, size, shape and orientation of the constituents of rocks, such as minerals, grains, organic matter, porosity, layering, bed boundaries, lithology contacts and fractures. Fabric elements contribute to the anisotropy of materials when the elements have preferential orientation along crystallographic axes, fractures and elongated and flat particles.

13. For more on permeability anisotropy: Ayan C, Colley N, Cowan G, Ezekwe E, Wannell M, Goode P, Halford F, Joseph J, Mongini A, Obondoko G and Pop J: “Measuring Permeability Anisotropy: The Latest Approach,” Oilfield Review 6, no. 4 (October 1994): 24–35.

For more on elastic anisotropy: Armstrong P, Ireson D, Chmela B, Dodds K, Esmersoy C, Miller D, Hornby B, Sayers C, Schoenberg M, Leaney S and Lynn H: “The Promise of Elastic Anisotropy,” Oilfield Review 6, no. 4 (October 1994): 36–47.

For more on the anisotropy of electrical properties: Anderson B, Bryant I, Lüling M, Spies B and Helbig K: “Oilfield Anisotropy: Its Origins and Electrical Characteristics,” Oilfield Review 6, no. 4 (October 1994): 48–56.

14. Primary migration refers to the flow of newly generated hydrocarbon fluids within source rocks. Secondary migration refers to the flow of free hydrocarbon fluids away from source rocks toward adjoining or distant reservoir rocks.

15. Martin R, Baihly J, Malpani R, Lindsay G and Atwood WK: “Understanding Production from Eagle Ford–Austin Chalk System,” paper SPE 145117, presented at the SPE Annual Technical Conference and Exhibition, Denver, October 30–November 2, 2011.

16. Boyer et al (2006), reference 1. Loucks and Ruppel, and Lash and Engelder, reference 9. The bulk volume fraction or percent of TOC in rock is

roughly twice its weight fraction or percent. A concentration of 2 weight % [0.02 kg/kg] TOC in rock is approximately equivalent to 4 bulk volume % [0.04 m3/m3] TOC. The exact calculation depends on the density and maturity of the organic matter and the bulk density of the host rock.

17. Bohacs KM, Grabowski GJ Jr, Carroll AR, Mankiewski PJ, Miskell-Gerhardt KJ, Schwalbach JR, Wegner MB and Simo JA: “Production, Destruction, and Dilution—The Many Paths to Source-Rock Development,” in Harris NB (ed): The Deposition of Organic-Carbon-Rich Sediments: Models, Mechanisms, and Consequences. Tulsa: Society of Sedimentary Geology, SEPM Special Publication 82 (2005): 61–101.

For more on source rock geochemistry: McCarthy K, Rojas K, Niemann M, Palmowski D, Peters K and Stankiewicz A: “Basic Petroleum Geochemistry for Source Rock Evaluation,” Oilfield Review 23, no. 2 (Summer 2011): 32–43.

18. The density of kerogen increases as organic carbon matures from immature, generative organic carbon to overmature, nongenerative organic carbon. For more on kerogen: Jarvie DM, Jarvie BM, Weldon WD and

Maende A: “Components and Processes Impacting Production Success from Unconventional Shale Resource Systems,” Search and Discovery Article 40908, adapted from an oral presentation at the 10th Middle East Geosciences Conference and Exhibition, Manama, Bahrain, March 4–7, 2012.

Okiongbo KS, Aplin AC and Larter SR: “Changes in Type II Kerogen Density as a Function of Maturity: Evidence from the Kimmeridge Clay Formation,” Energy & Fuels 19, no. 6 (November 2005): 2495–2499.

19. Loucks RG, Reed RM, Ruppel SC and Jarvie DM: “Morphology, Genesis, and Distribution of Nanometer-Scale Pores in Siliceous Mudstones of the Mississippian Barnett Shale,” Journal of Sedimentary Research 79, no. 12 (December 2009): 848–861.

Curtis ME, Cardott BJ, Sondergeld CH and Rai CS: “Development of Organic Porosity in the Woodford Shale with Increasing Thermal Maturity,” International Journal of Coal Geology 103 (December 1, 2012): 26–31.

20. Wang FP and Reed RM: “Pore Networks and Fluid Flow in Gas Shales,” paper SPE 124253, presented at the SPE Annual Technical Conference and Exhibition, New Orleans, October 4–7, 2009.

Ambrose RJ, Hartman RC, Diaz-Campos M, Akkutlu IY and Sondergeld CH: “Shale Gas-in-Place Calculations Part I: New Pore-Scale Considerations,” SPE Journal 17, no. 1 (March 2012): 219–229.

Curtis ME, Sondergeld CH, Ambrose RJ and Rai CS: “Microstructural Investigation of Gas Shales in Two and Three Dimensions Using Nanometer-Scale Resolution Imaging,” AAPG Bulletin 96, no. 4 (April 2012): 665–677.

22 Oilfield Review

In addition, kerogen fabric affects the elastic and mechanical properties of reservoir rocks.21 Generally, mudstones that contain intercon-nected kerogen within their frame are character-ized by lower elastic moduli and higher ductility compared with those mudstones that have iso-lated kerogen particles dispersed through their matrix. Kerogen content distributed parallel to the laminae may profoundly affect the anisotro-pic elastic and mechanical properties of mud-stones.22 These effects will be enhanced if, in addition to creating secondary porosity within kerogen, hydrocarbon generation and charging of the kerogen-rich laminae cause overpressuring, a condition that is conducive to creating layer-par-allel microcracks, which strike parallel to layers and open perpendicular to them.23 Because matrix permeability in shale reservoirs is excep-tionally low, ranging from 10–7 to 10–3 mD, natu-ral fractures play a significant role in reservoir completions and hydrocarbon production.

Natural fractures contribute to the perfor-mance of hydraulic fracture stimulations by pro-viding planes of weakness and conduits for fluid flow.24 As planes of weakness, natural fractures may dictate the propagation and development of induced fracture networks, especially if the in situ stress anisotropy is low.25 As conduits for fluid flow, these fractures may enhance the extent of effective reservoir volume drained by the wellbore, and they may admit high-pressure fluids, which could cause permanent shear slip-page along their fracture planes and increase fracture aperture and conductivity.

Local RQ sweet spots in prospective mud-stone reservoirs often contain natural fractures that provide flow pathways connecting matrix porosity and storage to hydraulic fractures and the well. Natural fractures may also affect CQ through the geometry of stimulation-induced hydraulic fracture networks, which tend to become more widespread and complex when pre-existing natural fracture networks are oriented at an angle to the present-day principal horizon-tal stress.26 When mudstone reservoirs lack natu-ral fractures, operating companies must rely on hydraulic fracture stimulations to create induced fracture networks that connect production from the reservoir matrix to the wellbore. Therefore, natural fractures, which can increase both RQ and CQ, are a seismic exploration target in the hunt for sweet spots in shale reservoirs.

Through the analysis of seismic attributes, geo-physicists detect and characterize fracture networks. This process takes advantage of the volume-averaged response from the entire reservoir interval contain-ing an open natural fracture system.27

Numerous fracture detection methods use seis-mic attributes. When natural fractures align in a consistent strike orientation, they cause elastic properties and seismic attributes to vary with azi-muth, including velocity and reflection ampli-tude.28 Geophysicists observe these variations from analysis of 3D surface seismic surveys that have been acquired along multiple azimuths.29 Azimuthal analysis of shear waves (S-waves) has proved to be a good fracture detection method.30 Analysis of seismic waveform scattering, which was often treated as noise in the past, may also reveal information about fracture orientation and spacing through frequency analysis.31 In addition, combinations of attributes such as reflection strength and seismic variance—the variation between seismic samples—may be blended, or superimposed, to expose subtle structural features that have fracture systems associated with them.32

Regional- or Basin-Scale Sweet SpotsDuring the initial years of the current surge in activity in shale plays, some operators were able to develop shale plays on the basis of hydrocar-

bon shows on mud logs that were recorded in shales encountered while drilling conventional reservoirs within a basin. The regions within these basins where organic shales are thermally mature were already known to the industry; therefore, for many of the shale plays in North America, it was unnecessary for operators to investigate the plays’ thermal maturity.

Because of the successful development of the Barnett Shale in the Fort Worth basin in north-central Texas, operators widened their search for shale gas beyond North America to basins that are less explored. In certain basins around the world, few wells have been drilled, and operators lack the level of understanding of the strati-graphic and structural framework to anticipate where potential shale resource plays exist. In these basins, initial exploration for potential shale reservoirs relies on evaluating preexisting 2D seismic surveys and additional structural data from remote sensing analyses and outcrop stud-ies of surface geology.

Geoscientists evaluate these data to establish the structural framework of the major basinal

21. Suárez-Rivera R, Deenadayalu C and Yang Y-K: “Unlocking the Unconventional Oil and Gas Reservoirs: The Effect of Laminated Heterogeneity in Wellbore Stability and Completion of Tight Gas Shale Reservoirs,” paper OTC 20269, presented at the Offshore Technology Conference, Houston, May 4–7, 2009.

22. Vernik L and Landis C: “Elastic Anisotropy of Source Rocks: Implications for Hydrocarbon Generation and Primary Migration,” AAPG Bulletin 80, no. 4 (April 1996): 531–544.

Vernik L and Milovac J: “Rock Physics of Organic Shales,” The Leading Edge 30, no. 3 (March 2011): 318–323.

Sayers CM: “The Effect of Kerogen on the Elastic Anisotropy of Organic-Rich Shales,” Geophysics 78, no. 2 (March–April 2013): D65–D74.

23. For more on layer-parallel microcracks: Lash GG and Engelder T: “An Analysis of Horizontal Microcracking During Catagenesis: Example from the Catskill Delta Complex,” AAPG Bulletin 89, no. 11 (November 2005): 1433–1449.

24. Miller C, Hamilton D, Sturm S, Waters G, Taylor T, Le Calvez J and Singh M: “Evaluating the Impact of Mineralogy, Natural Fractures and In Situ Stresses on Hydraulically Induced Fracture System Geometry in Horizontal Shale Wells,” paper SPE 163878, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, February 4–6, 2013.

25. Weng X, Kresse O, Cohen C, Wu R and Gu H: “Modeling of Hydraulic Fracture-Network Propagation in a Naturally Fractured Formation,” SPE Production & Operations 26, no. 4 (November 2011): 368–380.

Kresse O, Cohen C, Weng X, Wu R and Gu H: “Numerical Modeling of Hydraulic Fracturing in Naturally Fractured Formations,” paper ARMA 11–363, presented at the 45th US Rock Mechanics/Geomechanics Symposium, San Francisco, June 26–29, 2011.

26. Miller et al, reference 24.27. For more on fracture detection using reflection

seismology: Aarre V, Astratti D, Al Dayyni TNA, Mahmoud SL, Clark ABS, Stellas MJ, Stringer JW, Toelle B, Vejbæk OV and White G: “Seismic Detection of Subtle Faults and Fractures,” Oilfield Review 24, no. 2 (Summer 2012): 28–43.

28. For more on elastic anisotropy: Armstrong et al, reference 13.

29. For more on azimuthal seismic anisotropy analysis: Barkved O, Bartman B, Compani B, Gaiser J, Van Dok R, Johns T, Kristiansen P, Probert T and Thompson M: “The Many Facets of Multicomponent Seismic Data,” Oilfield Review 16, no. 2 (Summer 2004): 42–56.

30. Hardage B: “Fracture Identification and Evaluation Using S Waves,” Search and Discovery Article 40792, adapted from five Geophysical Corner columns by B Hardage in AAPG Explorer 32, no. 4–8 (April–August 2011).

31. Burns DR, Willis ME, Toksöz MN and Vetri L: “Fracture Properties from Seismic Scattering,” The Leading Edge 26, no. 9 (September 2007): 1186–1196.

32. High seismic variance occurs where seismic data vary rapidly, such as when crossing faults or stratigraphic boundaries.

33. For more on petroleum system modeling: Al-Hajeri MM, Al Saeed M, Derks J, Fuchs T, Hantschel T, Kauerauf A, Neumaier M, Schenk O, Swientek O, Tessen N, Welte D, Wygrala B, Kornpihl D and Peters K: “Basin and Petroleum System Modeling,” Oilfield Review 21, no. 2 (Summer 2009): 14–29.

Peters KE, Magoon LB, Bird KJ, Valin ZC and Keller MA: “North Slope, Alaska: Source Rock Distribution, Richness, Thermal Maturity, and Petroleum Charge,” AAPG Bulletin 90, no. 2 (February 2006): 261–292.

Peters K, Schenk O and Bird K: “Timing of Petroleum System Events Controls Accumulations on the North Slope, Alaska,” Search and Discovery Article 30145, adapted from an oral presentation at the AAPG International Conference and Exhibition, Calgary, September 12–15, 2010.

Higley DK: “Undiscovered Petroleum Resources for the Woodford Shale and Thirteen Finger Limestone–Atoka Shale Assessment Units, Anadarko Basin,” Denver: US Geological Survey Open-File Report 2011–1242, 2011.

34. Boyer et al (2006), reference 1.35. US Department of Energy Office of Fossil Energy and

National Energy Technology Laboratory: “Modern Shale Gas Development in the United States: A Primer,” Washington, DC: US Department of Energy, April 2009.

36. The shorter wavelength portions of the seismic signals are scattered enough to become incoherent and cancel themselves out.

Winter 2013/2014 23

stratigraphic units, including the locations of major fault zones and other tectonic features. Once they complete this analysis, basin analysts can use the structural framework as input to petroleum system modeling to determine if organic shale formations might be thermally mature, and if so, where in the basin these may occur.33 When this information is combined with regional mapping of available TOC data, regional- or basin-scale sweet spots may be identified, enabling operators to select optimal locations for drilling initial vertical pilot wells in the next phase of exploration.

Local or Operating Area Sweet SpotsPetroleum system modeling predicts the location and characteristics of basin-scale sweet spots, including the distribution of kerogen content, its thermal maturity and the pore pressure within the prospective interval. However, these predic-tions can be confirmed only by drilling a pilot well. Core and logging measurements from the vertical pilot well provide data to update the modeling and determine whether the pilot well has intersected a sweet spot. Engineers can cat-egorize local sweet spots by analyzing RQ and CQ using the newly acquired core and log data.

Local sweet spots of high RQ have one or more of three properties. Local sweet spots may have high matrix porosity that contains signifi-cant amounts of free gas, which may be produced at high rates during initial production, allowing for rapid payout of a horizontal appraisal well.

In addition, sweet spots may have significant concentrations of kerogen. Kerogen-rich sweet spots also contain large volumes of adsorbed gas, which is stored mainly on kerogen surfaces.34 This adsorbed gas contributes to sustained production as pressure drops during reservoir depletion, long after the free gas has been consumed.

Local RQ sweet spots may also have dense networks of open microfractures. Similar to high-porosity sweet spots, densely fractured sweet spots contain free gas that is produced during the early production of a well. In addition, microfrac-tures also increase the system permeability within a shale reservoir.

The best RQ sweet spots include all three properties—increased porosity, kerogen and microfracturing—which in turn affect various attributes of seismic data through their effect on rock properties. Increased porosity and the pres-ence of fractures typically cause decreases in seis-mic velocity and increased attenuation of high frequencies. Concentrations of kerogen can also

lower the elastic moduli and density of mud-stones, but to a lesser degree. Changes in certain seismic attributes associated with these rock properties may be used to identify RQ sweet spots.

Correlating Frequency Anomalies to

Production BehaviorIn the Arkoma basin of southeast Oklahoma, USA, gas production has been established from the Woodford Shale, a Late Devonian–Early Mississippian age organic-rich mudstone. Its mineralogy is primarily quartz and illite, with small quantities of pyrite and dolomite. Porosity ranges from 3% to 9% and TOC ranges from 1 to 14 weight % [0.01 to 0.14 kg/kg].35

An operator targeting Woodford Shale gas production had drilled six vertical wells within a 4-mi2 [10-km2] area. The wells’ production rates varied widely. In a 2.5-year period, cumulative gas production per well ranged from 18 to 372 MMcf [0.51 to 10.5 million m3] with average cumulative production from the five lowest pro-ducing wells of 40 MMcf [1 million m3]. The operator had acquired a 3D seismic survey over the field and requested that Schlumberger ana-lysts interpret the data to determine why the production was so variable and to locate areas of potentially higher production.

The 3D seismic data provide far greater cover-age of the reservoir interval than could be achieved by vertical or horizontal well data. The 3D seismic data were initially interpreted to locate faults and any other geohazards within the area, but the observed faulting and fracturing associated with fault damage zones could not explain the production history or the well-to- well variability.

Geophysicists analyzed the data for seismic attributes that would reveal RQ sweet spots. They identified a seismic frequency attribute that at certain frequencies corresponded with areas of higher production. These seismic sweet spot anomalies were areas in which the dominant seismic frequency proved to be relatively low, apparently the result of scattering of waves from networks of natural fractures or microfractures.36 The anomalies appeared as isolated patches within the field, and the team interpreted them to represent areas of increased porosity and microfracturing within the shale reservoir. Productive wells were located within these anomalous areas, while the underperforming wells were not. The well with the highest produc-tion is situated within a large anomaly (above). At the time of the study, this well had produced nine times the average production of the other

> Reservoir quality sweet spots in a shale reservoir. Six vertical wells (red dots) were drilled into the Woodford Shale in the Arkoma basin in southeast Oklahoma. Their cumulative gas production after approximately 2.5 years through June 2009 varied widely. Interpretation of a 3D seismic dataset revealed faulting (black). However, the wells’ proximity to faults, which often is associated with fracture density in the faults’ damage zone, did not explain the production variation. An analysis of seismic frequencies in the dataset revealed a frequency attribute that interpreters identified with RQ sweet spots (dashed red outlines) when this attribute was strong. Gas production correlated with the size and strength of the seismically identified sweet spots.

18 MMcf

23 MMcf

Untestedanomaly

22 MMcf

80 MMcf

61 MMcf

372 MMcf

N

0 km

0 mi 2

2

Freq

uenc

y an

omal

y

Strong

Weak

24 Oilfield Review

five wells combined. This observation is in line with the tenfold increase observed in the Barnett Shale for wells located within sweet spots.37

In another shale play, an operator was devel-oping a combination fractured carbonate and gas shale unconventional reservoir in the Delaware basin of southern New Mexico and western Texas, USA. The company had drilled a number of hori-zontal boreholes at the interface between the carbonate and the underlying shale. Production from these wells varied significantly.

Schlumberger geophysicists analyzed a 3D seis-mic volume to help determine the location and extent of potential RQ sweet spots and define their geologic nature. The geophysicists performed prestack azimuthal inversion and several fre-quency-related studies. The results of these sepa-rate investigations converged on the same locations within the shale reservoir as potential RQ sweet spots. These sweet spots manifested themselves through specific frequency-related seismic attri-bute anomalies that were also coincident with zones of S-wave anisotropy. The team interpreted these areas to be volumes of enhanced microfrac-turing in the upper portion of the gas shale (left).

The operator drilled three horizontal wells along the carbonate/shale interface in the hope of encountering fractures within the carbonate formation and zones of high gas content in the shale. Production rates from these wells appeared to be directly related to the magnitude and size of the frequency anomalies and S-wave anisotropy. Well A was drilled across the top of a gentle anti-clinal feature, where high seismic variance indi-cated the presence of faulting along the crest of the fold. At the time of the study, Well A was the best producer, with an average production rate of 64 MMcf [1.8 million m3] of gas per month. Well B was drilled near a smaller seismic frequency anomaly and its monthly production rate was 28 MMcf [0.79 million m3], less than half that from Well A. Well C did not penetrate a frequency anomaly and its monthly production rate was a poor 7 MMcf [0.2 million m3].

The team believed that the frequency anoma-lies highlighted zones within the shale that con-tained more microfractures than other locations. The concentration of microfractures at the crest of the anticline is consistent with the tectonic extension the layers experienced during anti-cline formation. Other evidence suggests that this fracturing did not extend through the full shale thickness. Zones of enhanced microfractur-ing within the shale were encountered by Well A and, to much lesser degree, by Well B, and resulted in the improved production observed in both wells compared with that in Well C.

> Fracture detection with seismic frequency attributes. A seismic fence diagram composed of seismic cross sections and a horizon slice shows a frequency-related seismic attribute. The horizon slice is also blended with the seismic variance attribute (grayscale); only high variance values are shown. The fence diagram (inset) is formed from seismic sections along the trajectories of Wells A, B and C (dark blue). The horizon slice, which is taken along the top of the formation immediately below the shale reservoir, is curved by a gentle anticline. Along the anticlinal crest, the seismic variance and frequency attributes are high. Average monthly gas production rates, shown above each well’s lateral, illustrate how each well’s production rate corresponds to its proximity to strong frequency anomalies.

Freq

uenc

y an

omal

y

Weak

Strong

Variance

HighLow

64 MMcf 28 MMcf 7 MMcf

High variance fault indicator

Anticlinalcrest

Well A Well B Well C

Frequencyanomaly

No faults

Horizon along topof next deeper

formation

Seismic fence diagram

> Gas shows encountered during the drilling of Well A (black line). A seismic section (background) is shown in a perspective view looking down and into it. The section is parallel to the trajectory of Well A and cuts through the 3D volume of the frequency attribute. High values of the frequency attribute (red and pink) appear as clouds coming out of the section. Gas chromatograph readings (blue curve), obtained from the mud log, are shown along the horizontal portion of Well A. Perforation cluster locations (cyan diamonds) align with the mud log depth points (small red triangles below the log curve). Gas shows from the mud log were strong when the wellbore was near high values of the seismically derived frequency attribute.

Freq

uenc

y an

omal

y

Weak

Strong

Frequencyanomaly highs

Farther fromanomaly

Close to anomaly

Mud loggas shows

Strong shows Weak shows

Winter 2013/2014 25

37. Baihly et al, reference 5.38. Sturm SD and Gomez E: “Role of Natural Fracturing in

Production from the Bakken Formation, Williston Basin, North Dakota,” Search and Discovery Article 50199, adapted from a poster presentation at the AAPG Annual Convention and Exhibition, Denver, June 7–10, 2009.

39. Pitman JK, Price LC and LeFever JA: “Diagenesis and Fracture Development in the Bakken Formation, Williston Basin: Implications for Reservoir Quality in the Middle Member,” Denver: US Geological Survey Professional Paper 1653, 2001.

Pollastro RM, Roberts LNR and Cook TA: “Geologic Assessment of Technically Recoverable Oil in the Devonian and Mississippian Bakken Formation,” in US Geological Survey Williston Basin Province Assessment Team (ed): Assessment of Undiscovered Oil and Gas Resources of the Williston Basin Province of North Dakota, Montana, and South Dakota, 2010, Denver: US Geological Survey Digital Data Series DDS–69–W (2011): 5-1–5-34.

40. A continuous petroleum system is one that displays little if any buoyancy, or gravity, segregation of the reservoir fluids. Generated oil or gas migrated directly into reservoir storage within the source rock or adjacent formations. This contrasts with conventional petroleum systems in which generated oil or gas migrated from source rocks into traps that lie beneath a reservoir seal. Conventional reservoirs display distinct fluid contacts, which are products of gravity segregation.

Examination of the gas shows encountered during the drilling of Well A also supported this interpretation (previous page, bottom). The stron-gest gas shows coincided with strong seismic fre-quency anomalies. Where the frequency anomalies were weaker, gas shows were not as strong.

In another location within the same Delaware basin study area, the operator drilled two hori-zontal wells from a vertical pilot well. The wells were drilled from east to west, hydraulically stim-ulated in multiple stages and monitored for induced microseismicity. The team was able to correlate the microseismic event locations to areas where seismic frequency anomalies were strongest (above). It was evident that high levels of frequency anomaly corresponded to RQ sweet spots or, more specifically, to zones of high poros-ity and increased microfracture density. In addi-tion, these zones appeared to have favorable CQ.

Associating Anisotropy to Production PatternsThe Bakken Formation is an oil-producing petro-leum system. Its stratigraphy represents deposi-tion in a restricted, shallow water environment that existed throughout most of the Williston

basin, which covers portions of Alberta, Saskatchewan and Manitoba in Canada and Montana, North Dakota and South Dakota in the US.38 The Bakken Formation is of Late Devonian–Early Mississippian age and lies unconformably above the Late Devonian Three Forks Formation and conformably below the Early Mississippian Lodgepole Limestone.39 The Bakken Formation has been subdivided into lower, middle and upper members. The middle member is the reservoir and is a mixed clastic-carbonate interval consist-ing of dolomitic sandstones, dolomites and lime-stone. The upper and lower members are organic-rich shales that serve as the seal and source for hydrocarbons.

The model for the Bakken Formation is one of a continuous petroleum system.40 The organic-rich upper and lower Bakken shale members have 8 to 10 weight % [0.08 to 0.1 kg/kg] TOC and are source rocks that generated oil that had migrated locally into reservoirs hosted by the adjacent middle Bakken member and the under-lying Pronghorn, which includes the Sanish Sand member of the Three Forks Formation. Because of the relatively closed nature of this petroleum

> Comparison of microseismicity and frequency attribute anomalies indicating zones of good CQ. This perspective view looks down into a west-to-east seismic section. The seismic section is fully opaque, showing all values of the frequency attribute. Two horizontal wells (black curves) were kicked off from a vertical pilot well in the east. Strong values of the seismic frequency attribute within the 3D seismic volume and limited to the upper portion of the shale reservoir are shown as clouds (tan to red). Microseismic events (dots), color-coded by stimulation stage, tend to occur where values of the frequency anomaly are high (white ovals). This relationship suggests that strong values of this frequency attribute may also indicate zones of good CQ.

Freq

uenc

y an

omal

y

Weak

Strong

N

S

EW

system, overpressuring occurs in deeper parts of the basin, where the most hydrocarbon genera-tion has taken place. Pore space and fractures within the upper and lower Bakken shale mem-bers also provide reservoir storage.

26 Oilfield Review

Natural fractures can occur locally in the Bakken Formation, and where their intensity is sufficiently high, such as along the Antelope anticline in North Dakota, they can affect pro-duction. In general, the fractures are vertical to subvertical, bed bounded and partially to totally filled by quartz, calcite or, rarely, pyrite cements. Some vertical microfractures appear to be expul-sion, or fluid release, fractures that form when fluid pressures exceed the prevailing minimum principal compressive stress, allowing oil to migrate from the source rocks into adjacent res-ervoir members.

The RQ (porosity and permeability) of the middle member, along with the degree of over-pressuring, plays a large role in determining the productivity of the Bakken Formation. The ability to predict where better reservoir quality occurs dramatically increases the chance of success in this play.

For this reason, an E&P company operating in the Williston basin contracted with Schlumberger, whose geophysicists reprocessed a proprietary 3D multiazimuth seismic survey over an area within the Bakken play of North Dakota. The target res-ervoir horizon was in the middle Bakken member. The company wanted to base drilling locations on

patterns of initial production and seismic attri-butes, which are both affected by characteristics of reservoir geology. The company hoped to move away from drilling wells based on geometric pat-terns—lease boundaries or the Public Land Survey System—which ignores geologic heteroge-neity, and to take a deliberate approach to locate, orient and drill infill horizontal wells into highly productive reservoir locations.41

Geoscientists constructed a calibrated geo-logic model that was constrained by all available data, including well logs, borehole image logs and core samples. Geophysicists processed the 3D

> Rocks under stress. Randomly oriented soft—malleable and yielding—fabrics (top left, blue) within a host matrix (tan) may open in any direction in an isotropic stress field; soft fabrics may include pores, kerogen particles and microcracks. Under an anisotropic stress field (top right), such fabrics will be preferentially squeezed in the maximum compressive stress direction (orange arrows) and have their shapes modified less in the other principal stress directions. The N–S oriented maximum compressive stress (σHmax, bottom left) causes incident SW–NE polarized S-waves (gray arrows) to split into N–S polarized fast S-waves (brown arrows) and W–E polarized slow S-waves (gold arrows). In addition, incident P-waves (green arrows) resolve into P-waves that are fastest (red arrows) parallel to the N–S maximum compressive stress and slowest (blue arrows) perpendicular to it; the sinusoid (bottom right) shows the full azimuthal P-wave velocity variation.

NESW

Fast S-wave, N–S

Slow S-wave, W–E

σHmax

σV

σHmin

Isotropic Stress Field Anisotropic Stress Field

Azimuth

Fast P-wave, N–S

Slow P-wave, W–E

Incident P-wave

Incident P-wave

Incident S-wave

P-w

ave

velo

city

East

North South

West

Winter 2013/2014 27

41. Johnson GM and Miller P: “Advanced Imaging and Inversion for Unconventional Resource Plays,” First Break 31, no. 7 (July 2013): 41–49.

For more on the Public Land Survey System: “US Topo Quadrangles—Maps for America.” http://nationalmap.gov/ustopo/ (accessed January 17, 2014).

42. Johnson GM and Dorsey J: “Modeling Overburden Heterogeneity in Terms of Vp and TI for PSDM, Williston Basin, U.S.A.,” Expanded Abstracts, 80th SEG Annual Meeting, Denver (October 17–22, 2010): 4062–4065.

43. For more on OVT processing: Stein JA, Wojslaw R, Langston T and Boyer S: “Wide-Azimuth Land Processing: Fracture Detection Using Offset Vector Tile Technology,” The Leading Edge 29, no. 11 (November 2010): 1328–1337.

44. Johnson and Miller, reference 41.

seismic data to account for horizontal variability and anisotropy of seismic velocities in the strata above the reservoir.42 Seismic processors sorted the seismic data into offset vector tile (OVT) gathers, in which traces share similar source-to-receiver offset and azimuth.43 Using high-resolu-tion, multiazimuth OVT tomography, the processors modeled seismic velocities and anisotropy and used them for prestack depth migration (PSDM) of the OVT gathers. If there was disagreement between seismic picks of for-mation-top depths from PSDM and those from well data, the velocity and anisotropy model parameters were readjusted, and the tomography and PSDM steps were repeated until there was acceptable agreement between the geologic model and PSDM image.

Once the geologic model and PSDM image matched, subsequent processing could focus on the seismic anisotropic effects at middle Bakken reservoir depths that appeared to result from ori-ented geologic fabrics or stress anisotropy (previ-ous page). The geophysicists used the fitted elliptical anisotropy from traveltimes (FEATT) workflow to find the fast and slow compressional-wave (P-wave) velocities and directions at the reservoir level.

The FEATT workflow starts by converting the PSDM OVT gather from depth to two-way travel-time. The analyst or an automated routine picks residual traveltimes across common offset-azi-muth time horizons, converts the traveltimes to interval velocities and fits an ellipse to the veloci-ties. The ellipse’s major and minor axes and their orientations provide estimates of the fast and slow P-wave velocities and directions (right).

Following application of the FEATT workflow, the geophysicists applied amplitude variation with offset and azimuth (AVOAZ) analyses to esti-mate S-wave velocity anisotropy.44 The AVOAZ analysis of S-waves may provide higher vertical resolution of the anisotropy variation than P-wave

> Azimuthal anisotropy. The seismic data were sorted into offset vector tiles (OVTs) and converted to depth by conventional migration (top left) and by anisotropic prestack depth migration (PSDM) and tomography (top right). The latter process reduced the waviness in the data attributable to overburden effects and produced datasets appropriate for azimuthal anisotropy analysis. In both panels, the yellow zigzag line gives the azimuth distribution in the OVT, and offset increases from left to right. The PSDM OVT data (cyan box) were converted from depth to time (bottom left), and a horizon (red) was selected for fitted elliptical anisotropy from traveltimes (FEATT) analysis (bottom right). In this example, the seismic processors selected the minimum number of three points (red) required to fit an ellipse; in practice, processors use many more than three. Processors converted the residual moveout at each azimuth to P-wave velocity (the radius of the radial plot) and fitted a FEATT ellipse (blue points, black points and radii) to the input points. The ellipse yielded a fast P-wave velocity azimuth of 114.24° with a slow-to-fast P-wave velocity ratio of 0.974, or P-wave velocity anisotropy of 2.6%. (Adapted from Johnson and Miller, reference 41.)

0

180

Trav

eltim

eDe

pth

Offset

P-w

ave

velo

city

, m/s

3,000

3,200

3,400

3,600

3,800

Offset

Input pointsFitted pointsEllipse definition

PSDM OVT Gather

Conventional Migration of OVT Advanced PSDM Migration of OVT

0

30

60

90

120

150

180

210

240

270

300

330

Major axis 114.24°Aspect ratio 0.974

Azim

uth,

deg

ree

28 Oilfield Review

velocity anisotropy methods because of its sensi-tivity to interface contrasts rather than to the average cumulative response of overlying strata.45

The present-day Bakken Formation maximum in situ principal horizontal compressive stress direction determined from the hydraulic fracture stimulations is generally NE–SW.46 Vertical natu-ral fractures observed in wells within the area of investigation were oriented NW–SE, in the pres-ent-day minimum in situ horizontal compressive stress direction. The fractures tended to be min-eralized, had permeabilities in the microdarcy to nanodarcy range and were not believed to be contributing to production.47 In addition, the RQ of the Bakken Formation in the area of investiga-tion was poor to fair, which explains the low pro-duction rates.

The team compared the seismic anisotropy results to the first 90 days of production from wells across the field. Areas of low production correlated to those having weak P- and S-wave anisotropy, and areas of high production were associated with strong anisotropy (left). Anisotropy was weak to the west and strong in the east, which helped explain why the eastern part of the field was more productive than the western part. Along with production improve-ment from west to east in the area of interest, the anisotropy orientation changed from NW–SE in the west to NE–SW in the east. An improvement in matrix properties is one explanation for this change; in addition, geophysicists speculate that this change in anisotropy direction represents a change in natural fracture orientation from one side of the field to the other. In the east, NE–SW oriented fractures would be parallel to the pres-ent-day maximum principal stress direction and would tend to be open (left).

In addition, anisotropy appeared strong in close proximity to source and reservoir rock con-tacts. From bottom to top, strong anisotropy occurred around the upper Three Forks–lower Bakken contact, the lower Bakken–middle Bakken contact, the middle Bakken–upper Bakken contact and through the upper Bakken member into the lower Lodgepole Limestone (next page). This result indicates that anisotropy from 3D multiazimuth surface seismic data may be used to delineate the areal and depth distribu-tion of sweet spots and future drilling targets.

> Production sweet spots. A seismic horizon through the middle Bakken member shows the slow-to-fast S-wave velocity ratio derived from AVOAZ inversion. The black arrows represent the relative magnitude of the estimated S-wave anisotropy; the arrow directions provide the fast S-wave vector azimuth from the inversion. The colored circles mark the average location of long horizontal wells and show the initial 90 days of oil production within the mapped area. To the west, the production was low to moderate, and S-wave velocity anisotropy is weak (blue to purple); the fast S-wave direction tends to be NW–SE. In the east, the production was higher, anisotropy is stronger (yellow to red) and the fast S-wave direction has a SW–NE trend, which is consistent with the present-day regional maximum in situ principal compressive stress direction. Initial production tends to be higher where the anisotropy is stronger. Analysts interpret the anisotropy to be associated with production sweet spots that are potential targets for drilling. (Adapted from Johnson and Miller, reference 41.)

3,000

12,000

0

0

m

ft

10,000

20,000

30,000

40,000

50,000

70,000

80,000

60,000

Firs

t 90

days

of o

il pr

oduc

tion,

bbl

Anisotropic Isotropic

0.8 1.0Slow/fast shear velocity ratio

N

> Volumes of high anisotropy. This view of S-wave velocity anisotropy within the middle Bakken member is looking down and north. The orange and red clouds are volumes of low ratios of slow-to-fast S-wave velocity, equivalent to high anisotropy, extracted from 3D seismic data between the upper and lower Bakken members. The anisotropy is strong in the east and south and weaker in the northwest. The blue surface underneath the clouds is from within the lower Bakken member and shows the ant-tracking seismic attribute (black to white), which accentuates traces of faults and fractures. (Adapted from Johnson and Miller, reference 41.)

Slow

/fas

t she

ar v

eloc

ity ra

tio

Anisotropic

Isotropic1.0

0.8

Top of theupper Bakken

Top of thelowerBakken

Top of themiddleBakken

1,500 m4,900 ft 1,5

00 m

4,900

ft

N

45. Hall SA and Kendall JM: “Constraining the Interpretation of AVOA for Fracture Characterization,” in Ikelle L and Gangi A (eds): Anisotropy 2000: Fractures, Converted Waves, and Case Studies. Tulsa: Society of Exploration Geophysicists (2001): 107–144.

46. Sturm and Gomez, reference 38.47. Sturm and Gomez, reference 38.

Winter 2013/2014 29

The Value of Seismic DataThese examples of using surface seismic data to understand patterns of production have been ret-rospective rather than prospective. Operators continue to test and appraise the identified sweet spots with new wells.

An increasing number of operators are acquir-ing and analyzing 3D surface seismic data during the early stages—exploration, pilot and appraisal phases—in the operating cycle of organic shale plays. Suitably analyzed and interpreted seismic data have proved to be invaluable for guiding the

placement of initial wells within a frontier shale basin, appraisal wells within a prospective shale basin and infill wells as part of a field develop-ment program in a mature basin. —RCNH

> Slow-to-fast S-wave velocity ratio near the middle Bakken member boundaries. The S-wave velocity ratio was calculated from an AVOAZ inversion for a pair of crossing seismic sections. The red rectangle (top) shows the middle Bakken reservoir interval displayed in the main figure (bottom left). The vertical black dashed line marks the intersection of the inline and crossline seismic sections and the approximate location of a vertical well. The relative shear velocity ratio within the middle Bakken Formation at this location is higher (blue to purple) in the center and lower (green to yellow) at the formation boundaries, indicating that anisotropy increases from the formation’s center to its boundaries. The log plot (bottom right) shows two tracks. Track 1 (left) displays well log traces of bulk density (pink), P-wave slowness (red), P-wave impedance (blue) and tops of geologic formations and members. Track 2 (right) shows the slow-to-fast S-wave velocity ratio (blue) from the AVOAZ inversion result along the well trace in the main display; it also shows formation and member tops. There is a difference in resolution between the well logs and the inversion result. Locations of tops are sharp and clear in the well log display and not as clear in the inversion display because of surface seismic resolution limitations. (Adapted from Johnson and Miller, reference 41.)

Isotropic

Slow-to-fast shearvelocity ratio

Anisotropic0.820.84

0.86

0.88

0.90

0.92

0.94

0.96

0.981.0

MiddleBakken

30 m

120 m

Bulk Density

P-Wave Slowness

P-Wave Impedance

1.5 g/cm3

μs/ft

3

147

15,700 (ft/s)×(g/cm3) 60,500

24

Isotropic

Slow-to-Fast ShearVelocity Ratio

Anisotropic

Sanish Sand

Three Forks

Sanish Sand

Three Forks

False Bakken

Scallion

Upper Bakken

Middle Bakken

False Bakken

Scallion

Upper Bakken

Middle Bakken

Lower Bakken Lower Bakken

X,235

Dept

h, ft

X,240

X,245

X,250

X,255

Low

erLo

dgep

ole

Bakk

en