Seeking the Sweet Spot: Reservoir and
Completion Quality in Organic Shales
Karen Sullivan Glaser
Camron K. Miller
Houston, Texas, USA
Greg M. Johnson
Brian Toelle
Denver, Colorado, USA
Robert L. Kleinberg
Cambridge, Massachusetts, USA
Paul Miller
Kuala Lumpur, Malaysia
Wayne D. Pennington
Michigan Technological University
Houghton, 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.
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.
16
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.
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
Oilfield Review
Winter 2013/2014 17
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-
N
0 km
0 mi
> 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.)
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
18
mining where a resource play may exist and
where RQ and CQ are highest.
Mudstone Characteristics
Geologists 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
600
520
440
Gas production, MMcf
360
280
200
120
2
2
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
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
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.
Oilfield Review
 sCore Classification Clay Clay

10 Clay- 10
dominated
lithotype
20 80 Shale plays 20 80

arg
ne
Barnett

C ar ceous
illa
o
dst
us ch

bon mu
mu
ceo -ri
Marcellus

ate dst
i l l a S ilic a
Mixed

-ric one
argillaceous

h
mudstone Haynesville
arg Fayetteville
50 50 50 50
Argillaceous- Argillaceous- Eagle Ford
siliceous carbonate
mudstone mudstone
(marl) Wolfcamp

car
Mixed
one

bon
mudstone

Cla te mu
us ich
dst

a
y-ri dst
ceo C lay-r
mu

ch one
Mixed Mixed
80 siliceous carbonate 20 80 20
sili

mudstone Carbonate- mudstone

siliceous
mudstone
Silica- Carbonate-
10 10
dominated Carbonate-rich Silica-rich dominated
lithotype siliceous mudstone carbonate mudstone lithotype

QFM 10 20 50 80 Carbonate QFM 10 20 50 80 Carbonate

Wolfcamp Shale Clay Clay

Effective porosity, fraction Minimum stress gradient, psi/ft
10 0.150 10 0.62

20 80 0.120 20 80 0.65
Increasing RQ

Increasing CQ
0.090 0.68

0.061 0.71
50 50 50 50
High RQ
0.031 High CQ 0.74

0.001 0.77
80 20 80 20

10 10

QFM 10 20 50 80 Carbonate QFM 10 20 50 80 Carbonate

Eagle Ford Shale Clay Clay

Effective porosity, fraction Minimum stress gradient, psi/ft
10 0.080 10 0.83

20 80 0.064 20 80 0.85
Increasing RQ

Increasing CQ

0.048 0.88

0.033 0.90
50 50 50 50

0.017 0.93

0.001 0.95
80 20 High RQ 80 20
High CQ
10 10

QFM 10 20 50 80 Carbonate QFM 10 20 50 80 Carbonate

> 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.

Winter 2013/2014 19
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

Plain Light Ultraviolet Light

0
226

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8
Core length, ft

0.9

1.0 227

1.1

0 m 1 1.2
0 mm 5
0 ft 3 1.3

1.4

1.5

1.6

1.7
0 mm 0.5
1.8

1.9

2.0
> 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.)

20 Oilfield Review
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.
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.
Winter 2013/2014
0 μm 200
0 mm 5
> 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.)
Kerogen fabric affects the physical properties frame, the organic porosity may be sufficient for
of organic-rich mudstones. When the organic con- hydrocarbon storage and for providing permeabil-
tent is high and the kerogen forms interconnected ity to gas and liquid hydrocarbon in an otherwise
layer-parallel networks through the mudstone extremely low-permeability matrix.20
15. Martin R, Baihly J, Malpani R, Lindsay G and Maende A: “Components and Processes Impacting
Atwood WK: “Understanding Production from Production Success from Unconventional Shale
Eagle Ford–Austin Chalk System,” paper SPE 145117, Resource Systems,” Search and Discovery Article
presented at the SPE Annual Technical Conference and 40908, adapted from an oral presentation at the 10th
Exhibition, Denver, October 30–November 2, 2011. Middle East Geosciences Conference and Exhibition,
16. Boyer et al (2006), reference 1. Manama, Bahrain, March 4–7, 2012.
Loucks and Ruppel, and Lash and Engelder, reference 9. Okiongbo KS, Aplin AC and Larter SR: “Changes in
Type II Kerogen Density as a Function of Maturity:
The bulk volume fraction or percent of TOC in rock is
Evidence from the Kimmeridge Clay Formation,”
roughly twice its weight fraction or percent. A
Energy & Fuels 19, no. 6 (November 2005): 2495–2499.
concentration of 2 weight % [0.02 kg/kg] TOC in rock is
19. Loucks RG, Reed RM, Ruppel SC and Jarvie DM:
approximately equivalent to 4 bulk volume % [0.04 m3/m3]
TOC. The exact calculation depends on the density and “Morphology, Genesis, and Distribution of Nanometer-
maturity of the organic matter and the bulk density of the Scale Pores in Siliceous Mudstones of the Mississippian
host rock. Barnett Shale,” Journal of Sedimentary Research 79,
no. 12 (December 2009): 848–861.
17. Bohacs KM, Grabowski GJ Jr, Carroll AR,
Curtis ME, Cardott BJ, Sondergeld CH and Rai CS:
Mankiewski PJ, Miskell-Gerhardt KJ, Schwalbach JR,
“Development of Organic Porosity in the Woodford
Wegner MB and Simo JA: “Production, Destruction,
and Dilution—The Many Paths to Source-Rock Shale with Increasing Thermal Maturity,” International
Development,” in Harris NB (ed): The Deposition of Journal of Coal Geology 103 (December 1, 2012): 26–31.
Organic-Carbon-Rich Sediments: Models, Mechanisms, 20. Wang FP and Reed RM: “Pore Networks and Fluid Flow
and Consequences. Tulsa: Society of Sedimentary in Gas Shales,” paper SPE 124253, presented at the
Geology, SEPM Special Publication 82 (2005): 61–101. SPE Annual Technical Conference and Exhibition,
For more on source rock geochemistry: McCarthy K, New Orleans, October 4–7, 2009.
Rojas K, Niemann M, Palmowski D, Peters K and Ambrose RJ, Hartman RC, Diaz-Campos M, Akkutlu IY
Stankiewicz A: “Basic Petroleum Geochemistry for and Sondergeld CH: “Shale Gas-in-Place Calculations
Source Rock Evaluation,” Oilfield Review 23, no. 2 Part I: New Pore-Scale Considerations,” SPE Journal 17,
(Summer 2011): 32–43. no. 1 (March 2012): 219–229.
18. The density of kerogen increases as organic carbon Curtis ME, Sondergeld CH, Ambrose RJ and Rai CS:
matures from immature, generative organic carbon to “Microstructural Investigation of Gas Shales in Two and
overmature, nongenerative organic carbon. For more on Three Dimensions Using Nanometer-Scale Resolution
kerogen: Jarvie DM, Jarvie BM, Weldon WD and Imaging,” AAPG Bulletin 96, no. 4 (April 2012): 665–677.
21
 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
22
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 Spots
During 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-
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.
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
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.
Oilfield Review
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 Spots
Petroleum 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
Winter 2013/2014
18 MMcf Strong
61 MMcf
23 MMcf
Untested
Frequency anomaly
anomaly
22 MMcf
N
80 MMcf
372 MMcf
Weak
0 km 2
0 mi 2
> 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.
lower the elastic moduli and density of mud- The 3D seismic data provide far greater cover-
stones, but to a lesser degree. Changes in certain age of the reservoir interval than could be
seismic attributes associated with these rock achieved by vertical or horizontal well data. The
properties may be used to identify RQ sweet spots. 3D seismic data were initially interpreted to
locate faults and any other geohazards within the
Correlating Frequency Anomalies to area, but the observed faulting and fracturing
Production Behavior associated with fault damage zones could not
In the Arkoma basin of southeast Oklahoma, explain the production history or the well-to-
USA, gas production has been established from well variability.
the Woodford Shale, a Late Devonian–Early Geophysicists analyzed the data for seismic
Mississippian age organic-rich mudstone. Its attributes that would reveal RQ sweet spots. They
mineralogy is primarily quartz and illite, with identified a seismic frequency attribute that at
small quantities of pyrite and dolomite. Porosity certain frequencies corresponded with areas of
ranges from 3% to 9% and TOC ranges from higher production. These seismic sweet spot
1 to 14 weight % [0.01 to 0.14 kg/kg].35 anomalies were areas in which the dominant
An operator targeting Woodford Shale gas seismic frequency proved to be relatively low,
production had drilled six vertical wells within a apparently the result of scattering of waves from
4-mi2 [10-km2] area. The wells’ production rates networks of natural fractures or microfractures.36
varied widely. In a 2.5-year period, cumulative The anomalies appeared as isolated patches
gas production per well ranged from 18 to within the field, and the team interpreted them
372 MMcf [0.51 to 10.5 million m3] with average to represent areas of increased porosity and
cumulative production from the five lowest pro- microfracturing within the shale reservoir.
ducing wells of 40 MMcf [1 million m3]. The Productive wells were located within these
operator had acquired a 3D seismic survey over anomalous areas, while the underperforming
the field and requested that Schlumberger ana- wells were not. The well with the highest produc-
lysts interpret the data to determine why the tion is situated within a large anomaly (above).
production was so variable and to locate areas At the time of the study, this well had produced
of potentially higher production. nine times the average production of the other
23
 Seismic fence diagram 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,
Well C
USA. The company had drilled a number of hori-
Well A Well B
zontal boreholes at the interface between the
Strong
carbonate and the underlying shale. Production
from these wells varied significantly.
Schlumberger geophysicists analyzed a 3D seis-

Frequency anomaly
64 MMcf 28 MMcf 7 MMcf 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
Anticlinal Weak spots. These sweet spots manifested themselves
crest No faults through specific frequency-related seismic attri-
High variance Frequency Horizon along top bute anomalies that were also coincident with
fault indicator anomaly of next deeper
formation zones of S-wave anisotropy. The team interpreted
Variance
these areas to be volumes of enhanced microfrac-
turing in the upper portion of the gas shale (left).
Low High The operator drilled three horizontal wells
> Fracture detection with seismic frequency attributes. A seismic fence diagram composed of seismic along the carbonate/shale interface in the hope
cross sections and a horizon slice shows a frequency-related seismic attribute. The horizon slice is of encountering fractures within the carbonate
also blended with the seismic variance attribute (grayscale); only high variance values are shown.
formation and zones of high gas content in the
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. Production rates from these wells appeared
shale reservoir, is curved by a gentle anticline. Along the anticlinal crest, the seismic variance and to be directly related to the magnitude and size of
frequency attributes are high. Average monthly gas production rates, shown above each well’s lateral, the frequency anomalies and S-wave anisotropy.
illustrate how each well’s production rate corresponds to its proximity to strong frequency anomalies.
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
Strong 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
Strong shows Weak shows
28 MMcf [0.79 million m3], less than half that
Frequency anomaly

Mud log
from Well A. Well C did not penetrate a frequency
gas shows anomaly and its monthly production rate was a
poor 7 MMcf [0.2 million m3].
Farther from The team believed that the frequency anoma-
anomaly
lies highlighted zones within the shale that con-
tained more microfractures than other locations.
Close to
Frequency anomaly The concentration of microfractures at the crest
anomaly highs Weak
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
> Gas shows encountered during the drilling of Well A (black line). A seismic section (background) is shown shale thickness. Zones of enhanced microfractur-
in a perspective view looking down and into it. The section is parallel to the trajectory of Well A and cuts ing within the shale were encountered by Well A
through the 3D volume of the frequency attribute. High values of the frequency attribute (red and pink)
and, to much lesser degree, by Well B, and
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) resulted in the improved production observed in
align with the mud log depth points (small red triangles below the log curve). Gas shows from the mud log both wells compared with that in Well C.
were strong when the wellbore was near high values of the seismically derived frequency attribute.

24 Oilfield Review
 Strong

Frequency anomaly
Weak

W E

> 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.

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 Patterns
The 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
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.
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.

Winter 2013/2014 25
 Isotropic Stress Field Anisotropic Stress Field

Incident S-wave
Incident P-wave
SW NE

σHmin Fast P-wave, N–S

North South

P-wave velocity
East West
Slow P-wave, W–E

Azimuth
Slow
σHmax S-w
ave,
W–
E
Incident P-wave S
N–
a ve,
-w
tS
Fas

σV
> 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.

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

26 Oilfield Review
seismic data to account for horizontal variability Conventional Migration of OVT Advanced PSDM Migration of OVT

Azimuth, degree
0
and anisotropy of seismic velocities in the strata
above the reservoir.42 Seismic processors sorted 180

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

Depth
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 Offset
reservoir depths that appeared to result from ori- PSDM OVT Gather
ented geologic fabrics or stress anisotropy (previ-

velocity, m/s
ous page). The geophysicists used the fitted

P-wave
elliptical anisotropy from traveltimes (FEATT)
workflow to find the fast and slow compressional-
3,800 0
wave (P-wave) velocities and directions at the
330 30
reservoir level. 3,600

The FEATT workflow starts by converting the 300 3,400 60

PSDM OVT gather from depth to two-way travel-
Traveltime

3,200
time. The analyst or an automated routine picks
residual traveltimes across common offset-azi- 270 3,000 90

muth time horizons, converts the traveltimes to

interval velocities and fits an ellipse to the veloci-
240 120
ties. The ellipse’s major and minor axes and their
orientations provide estimates of the fast and
210 150
slow P-wave velocities and directions (right). 180
Following application of the FEATT workflow,
the geophysicists applied amplitude variation Input points
with offset and azimuth (AVOAZ) analyses to esti- Fitted points
Ellipse definition
mate S-wave velocity anisotropy.44 The AVOAZ
analysis of S-waves may provide higher vertical
Major axis 114.24°
resolution of the anisotropy variation than P-wave Aspect ratio 0.974

41. Johnson GM and Miller P: “Advanced Imaging and Offset

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.
> 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.)

Winter 2013/2014 27
 velocity anisotropy methods because of its sensi-
tivity to interface contrasts rather than to the
average cumulative response of overlying strata.45
80,000
The present-day Bakken Formation maximum
70,000 in situ principal horizontal compressive stress
First 90 days of oil production, bbl

direction determined from the hydraulic fracture

60,000 stimulations is generally NE–SW.46 Vertical natu-
ral fractures observed in wells within the area of
50,000
investigation were oriented NW–SE, in the pres-
40,000
ent-day minimum in situ horizontal compressive
stress direction. The fractures tended to be min-
30,000 eralized, had permeabilities in the microdarcy to
nanodarcy range and were not believed to be
20,000
contributing to production.47 In addition, the RQ
N
10,000 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
Anisotropic Isotropic 0 m 3,000 results to the first 90 days of production from
0 ft 12,000 wells across the field. Areas of low production
0.8
Slow/fast shear velocity ratio
1.0 correlated to those having weak P- and S-wave
anisotropy, and areas of high production were
> Production sweet spots. A seismic horizon through the middle Bakken member shows the slow-to-
associated with strong anisotropy (left).
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 Anisotropy was weak to the west and strong in
azimuth from the inversion. The colored circles mark the average location of long horizontal wells and the east, which helped explain why the eastern
show the initial 90 days of oil production within the mapped area. To the west, the production was low part of the field was more productive than the
to moderate, and S-wave velocity anisotropy is weak (blue to purple); the fast S-wave direction tends western part. Along with production improve-
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 ment from west to east in the area of interest, the
in situ principal compressive stress direction. Initial production tends to be higher where the anisotropy orientation changed from NW–SE in
anisotropy is stronger. Analysts interpret the anisotropy to be associated with production sweet spots the west to NE–SW in the east. An improvement
that are potential targets for drilling. (Adapted from Johnson and Miller, reference 41.)
in matrix properties is one explanation for this
change; in addition, geophysicists speculate that
this change in anisotropy direction represents a
Top of the change in natural fracture orientation from one
upper
Bakken side of the field to the other. In the east, NE–SW
Top of the Isotropic oriented fractures would be parallel to the pres-
middle 1.0 ent-day maximum principal stress direction and
Bakken
would tend to be open (left).
Top of the
In addition, anisotropy appeared strong in
Slow/fast shear velocity ratio

lower
Bakken close proximity to source and reservoir rock con-
tacts. From bottom to top, strong anisotropy
occurred around the upper Three Forks–lower
N Bakken contact, the lower Bakken–middle
Bakken contact, the middle Bakken–upper
Bakken contact and through the upper Bakken
1,5 00 ft
m

member into the lower Lodgepole Limestone

00
4,9

4,900 ft 0.8 (next page). This result indicates that anisotropy

1,500 m Anisotropic from 3D multiazimuth surface seismic data may
be used to delineate the areal and depth distribu-
tion of sweet spots and future drilling targets.
45. Hall SA and Kendall JM: “Constraining the Interpretation
> Volumes of high anisotropy. This view of S-wave velocity anisotropy within the middle Bakken of AVOA for Fracture Characterization,” in Ikelle L and
member is looking down and north. The orange and red clouds are volumes of low ratios of slow-to-fast Gangi A (eds): Anisotropy 2000: Fractures, Converted
Waves, and Case Studies. Tulsa: Society of Exploration
S-wave velocity, equivalent to high anisotropy, extracted from 3D seismic data between the upper and Geophysicists (2001): 107–144.
lower Bakken members. The anisotropy is strong in the east and south and weaker in the northwest. 46. Sturm and Gomez, reference 38.
The blue surface underneath the clouds is from within the lower Bakken member and shows the
47. Sturm and Gomez, reference 38.
ant-tracking seismic attribute (black to white), which accentuates traces of faults and fractures.
(Adapted from Johnson and Miller, reference 41.)

28 Oilfield Review
 Bulk Density
1.5 g/cm3 3

Depth, ft
Slow-to-fast shear P-Wave Slowness
Slow-to-Fast Shear
velocity ratio 120 m 147 μs/ft 24 Velocity Ratio
1.0 Isotropic P-Wave Impedance
0.98 15,700 (ft/s)×(g/cm3) 60,500 Anisotropic Isotropic
X,235
0.96

Lodgepole
False Bakken False Bakken

Lower
0.94 Scallion
30 m Scallion
0.92
X,240 Upper Bakken Upper Bakken
0.90
0.88
Middle Bakken
0.86 Middle Bakken
0.84
Middle X,245
0.82 Anisotropic Bakken

Bakken
X,250 Lower Bakken Lower Bakken

X,255

Sanish Sand Sanish Sand

Three Forks Three Forks

> 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.)

The Value of Seismic Data
These 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

Winter 2013/2014 29