40 Oileld Review

Shale Gas Revolution

Around the globe, companies are aggressively pursuing shale resource plays, hoping
to nd the next Barnett Shale. But developing and producing from these enigmatic
resources require more than just nding organic-rich shales and hydraulically
fracturing them. As the shale gas revolution gains momentum globally, exploration
companies are discovering that an integrated approach is essential to success.
Learning from past experiences and continually improving methodologies may not
guarantee success, but its likelihood is greatly improved.
Tom Alexander
Southwestern Energy Company
Moncton, New Brunswick, Canada
Jason Baihly
Sugar Land, Texas, USA
Chuck Boyer
Pittsburgh, Pennsylvania, USA
Bill Clark
George Waters
Oklahoma City, Oklahoma, USA
Valerie Jochen
College Station, Texas
Jol Le Calvez
Houston, Texas
Rick Lewis
Camron K. Miller
Dallas, Texas
John Thaeler
Southwestern Energy Company
Houston, Texas
Brian E. Toelle
Denver, Colorado, USA
Oileld Review Autumn 2011: 23, no. 3.
Copyright 2011 Schlumberger.
For help in preparation of this article, thanks to Doug
Bentley, Warsaw, Poland; Damian Friend, Joanne Hresko,
Reese Mitchell, Brad Sylvester and Steve Thomson,
Southwestern Energy Company, Houston; Vincenzo De
Gennaro, Pau, France; Hariharan Ramakrishnan, Oklahoma
City, Oklahoma, USA; and Leah Hogarth, Houston.
DSI, ECLIPSE, EcoScope, ECS, ELANPlus, FMI, geoVISION,
HiWAY, HRA, Petrel, Platform Express, Sonic Scanner,
SonicScope, SpectroLith, StimMAP, StimMAP LIVE, TRA
and VISAGE are marks of Schlumberger.
From the advent of the modern oil and gas indus-
try, petroleum geologists have followed a conven-
tional route for exploration: look for hydrocarbon
source rocks, nd reservoir quality rocks where
hydrocarbons can accumulate, identify a trap-
ping mechanism and then drill a well. But a revo-
lution is taking place in the E&P industry. Rocks
that in the past were of little interest, other than
as potential source rocks, are today being actively
pursued as potential reservoirs. When considering
unconventional resource plays, the focus is on
nding organic shales.
The Barnett Shale of central Texas, USA, is
recognized as the play that initiated the recent
interest in developing shales as producing reser-
voirs. This development represents a fundamen-
tal shift in the way exploration companies
consider resource plays. The two main enabling
technologies that have made shale plays econom-
ical are extended-reach horizontal drilling and
multistage hydraulic fracture stimulation.
However, operators have discovered that there is
much more to producing gas from source rocks
than drilling horizontal wells and hydraulically
fracturing them.
Engineers and geologists studying shale gas
resources nd that having a greater under-
standing of a reservoir can lead to process
adaptation and renement of techniques. It is
important to integrate data from many sources
and at many scales to optimally drill, complete
and stimulate wells to produce hydrocarbons
from their source rocks.
This article reviews lessons learned from 30
years of shale development and looks at some of
the tools used to analyze shales. Taking an inte-
grated approach to developing its Fayetteville
Shale play, one operator was rewarded for its
efforts, as evidenced by increased gas produc-
tion and improved operational efciencies. In
another example, a large production log study,
using data from six major shale basins in the US,
uncovered results that may impact drilling and
stimulation practices.
Shale Reservoir Characteristics
Shales are ne-grained rocks that form from the
compaction of silt and clay-sized particles.
1
Sixty
percent of the Earths sedimentary crust consists
of shale, and it is the primary source rock for
most of the conventional hydrocarbon deposits in
the world.
2
Because shales are formed from mud,
they are often referred to as mudstones. Shales
are differentiated from other claystones and
mudstones in that they are laminatednely
layeredand ssile, which means they can be
1. Geologists generally consider sedimentary particles
smaller than 4 microns to be clay sized. Silt particles
range from about 4 to 62.5 microns.
2. Blyth FGH and de Freitas MH: A Geology for Engineers,
7th ed. Burlington, Massachusetts, USA: Butterworth-
Heinemann, 1984.
43223schD7R1.indd 1 11/17/11 7:44 PM
Autumn 2011 41 Autumn 2011 41
43223schD7R1.indd 2 11/17/11 7:44 PM
42 Oileld Review
broken or split into sheets along their lamina-
tions (left). Depending on their maturity and
mineralogy, they may also be easily fractured.
For all their abundance, few shale deposits
can be developed as hydrocarbon resource plays.
The targets for gas shale exploration are organic-
rich sediments that were deposited in such a
manner as to preserve a signicant fraction of
the original organic matter that eventually serves
as the feedstock for hydrocarbon generation.
Once generated, much of the hydrocarbon
remains trapped in the ultralow-permeability
rock matrix, unable to migrate.
Organic-rich shale formations form under spe-
cic conditionshigh levels of organic matter
and low levels of oxygendifferentiating them
from other shale deposits.
3
These conditions were
prevalent during various geologic ages, including
the Devonian Period when the climate was warm,
sea levels were high and much of the Earth was
covered by tropical seas. But the Devonian Period
was not the only time when thick deposits of
organic-rich sediments formed. Organic-rich
shales from the Precambrian Period through
modern times have been identied (below left).
4

However, to meet thermal maturity criteria, most
gas shale plays focus on sediments from a geologic
time range that begins in the Ordovician and
extends through the Pennsylvanian Period.
Organic matter, which consists mostly of
remains of plants and animals, settles to the bot-
tom of lakes or oceans and becomes food for other
animals and bacteria. However, in anoxic environ-
ments, anaerobic bacteria, which are less ef-
cient consumers than their aerobic counterparts,
are the only bacteria able to consume organic
matter. Thus, the sediments may retain much of
their original organic material. Today, the Black
Sea is a close analog for conditions needed to cre-
ate shale resource playsanoxic conditions allow
sediments to accumulate with high concentra-
tions of organic matter (next page, top right).
5
As more material accumulates and underlying
ooze becomes compacted, the sediments are bur-
ied deeper and subjected to increased pressure
and temperature. Laminations also develop. In
deep marine environments, these processes can
take place very slowly and encompass very long
periods of time; accumulations of a few centime-
ters may take thousands of years. The weight of
the overlying sediments expels uids and com-
pacts the mudstone, which are steps in the pro-
cess of lithication. The organic material slowly
and partially cooks and is transformed into kero-
gen, an insoluble material from which hydrocar-
bons, both oil and gas, can be generated.
>
Organic shales. Shales are ne-grained ssile sediments found in layers. They may be fractured and
may break in sheets along their laminations (inset). Organic-rich shales are characterized by higher
levels of total organic carbon than other shalesin the range of 2% to 10%. Thermally mature
organic-rich shales are the primary targets for exploration.
>
Global distribution of organic marine shales by geologic period. Tectonics, geography and climatic
conditions contribute to the deposition of organic-rich sediments. Organic-rich marine shales are
found across the globe. (Black circles represent the number of occurrences for each age.)
Exploration companies have focused on marine sediments that have sufcient thermal maturity to
convert kerogen into hydrocarbons to develop as resource plays. Lacustrine shales from shallow,
freshwater deposits (not shown) are also targets of exploration but have not yet proved to be as
prolic as their marine counterparts. (Adapted from Tourtelot, reference 4.)
Oilfield Review
AUTUMN 11
Shale Fig. 2
ORAUT11-SHL 2
Quaternary
and Tertiary
65
135
190
225
280
320
345
395
435
500
570
2,500
Million
years ago
North
America
South
America
Siberia and
Central Asia
Australia
and Asia
Africa Period Europe
Cretaceous
Jurassic
Triassic
Permian
Pennsylvanian
Mississippian
Devonian
Silurian
Ordovician
Cambrian
Proterozoic
Archeozoic
43223schD7R1.indd 3 11/17/11 7:44 PM
Autumn 2011 43
Different organic material generates different
kerogen types.
6
When exposed to heat and pres-
sure, each kerogen type is more prone to generate
specic products: oil, wet gas, dry gas and nonhy-
drocarbons. During the processes of burial and
maturation, kerogen passes through a range of
temperatures and pressures. First is the oil win-
dow, in which liquid petroleum may be generated
from oil-prone kerogen, or wet gas from gas-prone
kerogen (middle right). This stage of maturation is
referred to as catagenesis. With deeper burial, the
kerogen passes into the dry gas window. Through
the process of metagenesis, gas is generated by the
conversion of remaining kerogen and the transfor-
mation of heavier hydrocarbons created during
catagenesis. Shales that are rich in organic materi-
als and that have been subjected to temperatures
and pressures in the dry gas window are prime
targets for gas shale exploration.
However, just because sediments pass through
the maturation stages does not necessarily mean
they are reservoir quality rock. Using geochemi-
cal, petrophysical and geomechanical properties
derived from a variety of sources, geologists and
engineers determine the feasibility of proceeding
with gas shale exploration.
Geochemical Analysis
To identify shales that have production potential,
geologists look for specic geochemical proper-
ties, which are typically derived from core data.
Some of the properties can be measured with
downhole sensors; however, petrophysicists
rene and characterize downhole measurements
by calibrating log data to core data. Geochemical
properties needed to adequately characterize
shale resources include total organic carbon
(TOC), gas volume and capacity, thermal matu-
rity, permeability and mineralogy.
TOCAn organic shale, by denition, must
have organic carbon, and the TOC governs the
resource potential of a shale. Rocks with higher
TOC values are organically richer. Exploration
targets have TOC values in the general range of
2% to 10% (bottom right). Rocks with TOC above
10% are usually too immature for development.
3. Some geologists also believe that if the sediments are
deposited faster than oxic fauna can consume them, high
levels of organic matter can be preserved in sediments
that are not oxygen poor.
4. Tourtelot HA: Black ShaleIts Deposition and
Diagenesis, Clays and Clay Minerals 27, no. 5
(October 1979): 313321.
5. Lueschen H: Black Sea Sediments, http://www.
mbgc.icbm.de/HolgerL/BlackSea.html (accessed
September 20, 2011).
6. For more information on kerogen type: Boyer C,
Kieschnick J, Suarez-Rivera R, Lewis RE and Waters G:
Producing Gas from Its Source, Oileld Review 18,
no. 3 (Autumn 2006): 3649.
>
Modern analog for organic-rich shales. Decay of organic material is a bacterial process that occurs
under aerobic conditions; limited anaerobic bacterial activity can also occur under anoxic conditions.
The Black Sea is stratied with an upper oxidized layer and a lower anoxic zone. Freshwater (green
arrows) ows in from rivers, and denser seawater (blue arrow) ows in from the Mediterranean Sea
via the Bosporus Strait. Because of the different salinities and densities, mixing is limited to the
uppermost 100 to 150 m [330 to 490 ft]. The mixing between surface water and bottom water is strongly
restricted; the water at the bottom is exchanged only once in a thousand years. Black, organic-rich
sediments accumulate on the bottom. Anaerobic bacteria strip oxygen from sulfates and give off
hydrogen sulde [H
2
S] as a waste product. The hydrogen sulde may react with iron in the sediments
to form pyrite [FeS
2
], which is frequently observed in organic-rich shale deposits. (Adapted from
Lueschen, reference 5.)
Oilfield Review
AUTUMN 11
Shale Fig. 1
ORAUT11-SHL 1
Oxidized layer
Freshwater
Anoxic zone depleted of oxygen
Black organic sediments
S
a
l
t
w
a
t
e
r
Bosporus Strait
Black Sea
>
Maturation stages in hydrocarbon generation. The processes of burial,
conversion of organic matter and generation of hydrocarbons can be
summarized in three steps. Diagenesis: characterized by low-temperature
below 50C [122F]conversion of organic matter to kerogen. Bacteria may
digest and convert some of the organic matter into biogenic methane.
Catagenesis: generally occurs as further burial results in more pressure and
increased heat in the range of 50C to 150C [122F to 302F], which causes
chemical bonds to break within the shale and the kerogen. Metagenesis: the
nal stage, in which heat and chemical changes transform kerogen to
carbon. During this stage, late methane, or dry gas, evolves, along with other
gases, including CO
2
, N
2
and H
2
S. Hydrocarbons produced in earlier stages
eventually convert to methane, as well. Temperatures range from about
150C to 200C [302F to 392F] and higher.
Biomarkers
Oil
Wet gas Dry gas
H
y
d
r
o
c
a
r
b
o
n
s

g
e
n
e
r
a
t
e
d
Biogenic methane
Increasing depth and temperature
Metagenesis Catagenesis Diagenesis
Immature zone Oil window Gas window
>
The relationship between total organic carbon and resource potential.
Oilfield Review
AUTUMN 11
Shale Fig. Table 2
ORAUT11-SHL Table 2
Total Organic Carbon, Weight % Resource Potential
< 0.5 Very poor
Poor
Fair
Good
Very good
Unknown > 10
0.5 to 1
1 to 2
2 to 4
4 to 10
43223schD7R1.indd 4 11/17/11 7:44 PM
44 Oileld Review
The total carbon in a shale sample includes
both inorganic and organic carbon. To quantify
organic carbon, engineers use a combustion
technique.
7
A small portion of the rock sample is
rst treated with phosphoric acid to remove
inorganic carbon. Sediments are dried and then
combusted at 1,350C [2,462F] in an oxygen-
rich environment. The organic carbon is oxidized
to form CO
2
, which ows through a nondisper-
sive infrared detection cell tuned to respond to
CO
2
. The measured gas volumes are converted to
a TOC measurement and recorded as a weight
percent of the rock.
TOC values may vary greatly across a reser-
voir section. But because it is not feasible to
acquire and then analyze cores over long inter-
vals, petrophysicists commonly use downhole
data from geochemical and conventional logging
suites to quantify the volume of kerogen in the
rock and then compute TOC values from these
data. To validate the models used to measure
TOC, scientists calibrate petrophysical data to
core-derived values.
Gas volume and capacityGas is adsorbed
on the surface of the kerogen in the shale and is
also freely distributed in the primary and second-
ary porosity. The total gas in place (GIP) is the
combination of adsorbed and free gas. Depending
on the initial pressure of the reservoir, as free gas
is produced and the pore pressure falls, adsorbed
gas will be liberated, or desorb, from the surface of
the kerogen. However, recent research indicates
that desorption is also a function of the shale pore
size, which scientists must consider when estimat-
ing resource potential.
8

Scientists sometimes use canister desorption
tests to determine the total GIP from cores.
Immediately upon retrieval, freshly cut core sam-
ples are sealed in canisters and sent to the labo-
ratory for testing. The gas is removed from the
canister, volumetrically measured and composi-
tionally analyzed as a function of time. A plot of
gas produced over time can be used to estimate
the GIP for the core sample at reservoir condi-
tions. This analysis is sensitive to the amount of
time it takes to retrieve the core from downhole.
To determine adsorbed gas volume for shales,
engineers use pressure relationships that estimate
the sorptive potential of the rock. Samples are pul-
verized to maximize surface area and then heated
to drive off any adsorbed gas. Samples are then
exposed to methane at increasingly high pressure
while held at a constant temperature. The volume
of gas adsorbed by the rock sample, presented
in units of standard cubic feet/ton (scf/ton), is
described by a Langmuir isotherm curve (below).
Once an isotherm is established, the storage
capacity of the rock can be determined by refer-
encing the pore pressure of the formation, which is
representative of the in situ reservoir pressure.
Engineers use the Langmuir isotherms from
core data to compute the adsorbed gas from log-
derived TOC data. They compute free gas volumes
from log-derived effective porosity and gas satu-
ration, after subtracting the computed pore vol-
ume occupied by the adsorbed gas. Proper
evaluation requires geochemical and petrophysi-
cal input including clay content and type, matrix
density, formation water and bound water resis-
tivities, effective porosity and gas saturation.
Thermal maturityThermal maturity is a
function of depositional history. As kerogen is
exposed to progressively higher temperatures
over time, vitrinitecell-wall material and
woody plant tissue preserved in the rockunder-
goes irreversible alteration and develops
increased reectance. The measurement of
vitrinite reectance (R
o
) was originally devel-
oped to rank coal maturity.
R
o
is determined by microscope measure-
ments of the reectivity of at least 30 grains of
vitrinite from a rock sample: Values typically
range from 0% to 3%. Measurements in excess of
1.5% are a sign of dry gasgenerating source
rocks, a positive indicator for gas shales. R
o
ranges
of 0.6% to 0.8% indicate oil and ranges of 0.8% to
1.1% indicate wet gas. Initially, oil and conden-
sate were considered negative indicators for
shale development; however, some operators have
had success producing oil and condensate from
shale, and lower R
o
values can be considered a
positive indicator in these cases. A reectance
value below 0.6% is indicative of kerogen that is
immature, not having been exposed to sufcient
thermal conditions over adequate time for con-
version of the organic material to hydrocarbons.
>
Langmuir isotherms and gas storage capacity. The Langmuir isotherm (gold, left) is derived from crushed rock samples and quanties a rocks adsorbed
storage capacity. The Langmuir volume, V
L
(red line), is the theoretical limit for gas adsorption at innite pressure. Storage capacity at a given pressure, p,
can be determined from the plot. The Langmuir pressure, P
L
(vertical blue line), is the pressure at half the Langmuir volume. Using the Langmuir isotherm, the
total GIP (magenta, right) for a specic reservoir can be determined as a function of pressure. The total gas is the gas adsorbed to kerogen (gold) and free
gas stored in pores (blue). At low pressures, adsorption is an effective gas storage mechanism. As the pressure increases, pore gas correspondingly
increases. Productivity of most of the organic shale reservoirs being developed today is driven by volume of pore gas. Desorption becomes important as the
bottomhole owing pressure declines.
Oilfield Review
AUTUMN 11
Shale Fig. 7
ORAUT11-SHL 7
100
120
1,000 2,000 3,000 4,000 5,000
0
20
40
60
80
0
G
a
s

c
o
n
t
e
n
t
,

s
c
f
/
t
o
n
Pore pressure, psi
Langmuir volume (V
L
) =
gas volume at infinite pressure
Gas content = V
L
p/ (p + P
L
)
Langmuir pressure (P
L
) =
pressure at 0.5 V
L

Pressure, psi
350
250
150
100
50
0
0 1,000 2,000 3,000 4,000 5,000
300
200
G
a
s

c
o
n
t
e
n
t
,

s
c
f
/
t
o
n
Free gas
Adsorbed gas
Total gas
43223schD7R1.indd 5 11/22/11 11:11 PM
Autumn 2011 45
PermeabilityOne of the most difcult
properties to quantify when characterizing
shales is permeability to gas, which can range
from 0.001 to 0.0000001 mD. Permeability is a
function of effective porosity, hydrocarbon satu-
ration and mineralogy. Conventional reservoirs
have permeabilities in the hundreds of millidar-
cies, several orders of magnitude greater than
that observed in shales. Engineers measure per-
meability of conventional rocks by forcing uid
through cores and measuring the volume and
rate of uids as they pass through the sample.
Shale permeabilities in the nanodarcy range pre-
clude this conventional approach.
TerraTek developed the TRA tight rock analy-
sis pyrolysis technique to quantify ultralow per-
meability in unconventional formations.
9
In
addition to permeability, the TRA technique pro-
vides bulk and grain density, total and effective
porosity, water and hydrocarbon saturation, gas-
lled porosity, bound hydrocarbon saturation and
clay-bound water volume.
Reservoir permeability can also be estimated
from short duration nitrogen-injection falloff
tests. These tests provide system permeability and
take into account not only the matrix permeability
but also the inuence of natural fractures.
MineralogyShales can have complex mix-
tures of minerals, and the relative concentrations
of the constituents have the potential to make or
break a potential resource play. Core samples can
provide a wealth of information about the geo-
chemistry and mineralogy, but are limited to the
specic location where the sample was retrieved.
Mineralogy is more often determined from petro-
physical data acquired from downhole logging
tools, which are calibrated to core data.
Petrophysical Data
The primary data used for petrophysical analysis
of shale formations are the same as those used
for conventional reservoir analysisgamma ray,
resistivity, porosity and acousticwith the addi-
tion of neutron capture spectroscopy data. Just
as conventional oil and gas wells have key indica-
tors for production, shales with hydrocarbon pro-
duction potential display specic characteristics
that set them apart from shales with little or no
potential (above right).
Petrophysical analysis of shales begins with
one of the most basic measurements: the gamma
ray log. It may provide one of the rst indications
of the presence of organic-rich shale. Organic
matter commonly contains higher levels of natu-
rally occurring radioactive materialsthorium,
potassium and uraniumthan do conventional
reservoir minerals. Because they have a higher
concentration of organic matter than other sedi-
ments, organic-rich shales often exhibit gamma-
ray counts in excess of 150 gAPI. Petrophysicists
use high gamma ray counts to identify organic-
rich shale formations; however, some formations
of Cretaceous, Mesozoic and Tertiary age may not
display this artifact.
Triple-combo toolstrings, such as the Platform
Express integrated wireline logging tool, provide
resistivity and porosity measurements. They also
provide petrophysical characteristics to help log
analysts identify potential gas-bearing shales.
For instance, the resistivity measurements in
gas-bearing shales are usually higher than those
in surrounding shales that have no gas potential.
7. Bernard BB, Bernard H and Brooks JM: Determination
of Total Carbon, Total Organic Carbon and Inorganic
Carbon in Sediments, College Station, Texas, USA,
TDI-Brooks International and B&B Laboratories, Inc.,
http://www.tdi-bi.com/analytical_services/environmental/
NOAA_methods/TOC.pdf (accessed October 21, 2011).
8. Ambrose RJ, Hartman RC, Diaz-Campos M, Akkutlu IY
and Sondergeld CH: New Pore-Scale Considerations for
Shale Gas in Place Calculations, paper SPE 131772,
presented at the SPE Unconventional Gas Conference,
Pittsburgh, Pennsylvania, USA, February 2325, 2010.
9. Boyer et al, reference 6.
>
Prospecting with existing logging data. Petrophysicists use conventional
triple-combo logging data to identify potential organic shale deposits. The
bottom 50 ft of this log spans a conventional shale (blue shading). The gamma
ray (Track 1) reads less than 150 gAPI. The resistivity (Track 2) is low and has
a consistent value. The density and neutron porosities (Track 3) are separated
and exhibit little variability. Compared with the conventional shale, the
organic shale section (red shading) has a higher gamma ray reading, higher
and variable resistivity values and higher density porosity; the neutron
porosity has a higher degree of variability. The neutron porosity is often lower
in organic shales because they have a lower volume of bound water.
Gamma Ray
400 to 600 gAPI
Gamma Ray
200 to 400 gAPI
Washout
Crossover
Bit Size
in. 16.3 6.3
Caliper
in. 16.3 6.3
Gamma Ray
gAPI 200 0
10-in. Array
ohm.m 2,000 0.2
20-in. Array
ohm.m 2,000 0.2
30-in. Array
ohm.m 2,000 0.2
60-in. Array
ohm.m 2,000 0.2
Neutron Porosity (Limestone)
% 10 40
Density Porosity (Limestone)
% 10 40
Photoelectric Effect
20 0
90-in. Array
ohm.m 2,000 0.2
Resistivity
43223schD7R1.indd 6 11/17/11 7:44 PM
46 Oileld Review
Porosity measurements also have distinct
characteristics in gas-bearing shales. In general,
conventional shales exhibit a uniform separation
between the density porosity and neutron poros-
ity measurements. Organic-rich shales with
hydrocarbon production potential, however,
exhibit more variability, higher density porosity
and lower neutron porosity. This response is
partly a result of the presence of gas in the rock,
which lowers the hydrogen index and the result-
ing neutron porosity. A lower neutron porosity
may also occur in organic shales because of the
lower clay-mineral content in organic shales
compared with typical shales.
As for the density porosity measurement, the
constituent materials that make up shales gener-
ally have higher bulk densities than those of con-
ventional reservoir rocks such as sandstone or
limestone. In contrast, kerogen has a much lower
bulk density (1.2 g/cm
3
) than sandstone or lime-
stone, and its presence can lead to a higher com-
puted porosity. To accurately compute the density
porosity of a shale, engineers must know the
grain density of the rock in question. The grain
density is primarily derived from the ECS ele-
mental capture spectroscopy tool. The ECS tool
also provides a kerogen estimation for correcting
the grain density.
Formation evaluation to characterize uncon-
ventional reservoirs depends heavily on under-
standing the mineralogy of rocks. The
characterization can be done by analyzing cores,
but this method is neither efcient nor cost-
effective over long depth intervals. Continuous
measurements from logging tools such as the ECS
tool provide elemental yields that lead to the esti-
mation of weight percent for various minerals
that are common to organic shales.
The primary outputs from the spectroscopy
tool include silicon [Si], calcium [Ca], iron
[Fe], sulfur [S], titanium [Ti], gadolinium [Gd]
and potassium [K]. Schlumberger engineers use
SpectroLith lithology processing of spectra from
neutron-induced gamma ray spectroscopy tools
to compute the mineralogy and geochemical
data. They then calibrate SpectroLith outputs
with empirical relationships derived from an
extensive core chemistry and mineralogy data-
base. Matrix grain density can be determined
from these data and used for the porosity com-
putation. Certain types and quantities of miner-
als may indicate rocks that break or fracture
more easily. Log analysts use this information to
identify intervals for placing both vertical and
horizontal wellbores and initiating hydraulic
fracturing. Spectroscopy data can also be
acquired while drilling using the EcoScope mul-
tifunction logging-while-drilling service.
Along with mineralogy data, spectroscopy
measurements provide information on clay types.
Engineers use clay type to predict sensitivity to
fracturing uids and to understand the fracturing
characteristics of the formation. Contact with
water will cause some clays to swell, which inhib-
its gas production and creates numerous opera-
tional issues. Smectite is the most common
swelling clay. Fluid sensitivity may be inferred
from clay typing, but tests on cores from the res-
ervoir provide the most accurate information.
In addition to indicating uid sensitivity, clay
type is an indicator of rocks that are ductile, thus
do not fracture easily. Ductile shales are more
likely to embed proppant. Other shale types may
be brittle and are more easily fractured. The
presence of illite is preferred for hydraulic frac-
turing because it is often indicative of brittle
rocks that are not reactive with water. The pres-
ence of smectite usually indicates ductile clay.
Acoustic measurements, especially those
that provide mechanical properties for anisotro-
pic shale media, are also a signicant need for
understanding the long-term productivity of
shale gas wells. The Sonic Scanner acoustic
scanning platform provides data that are used
to enhance mechanical earth models and opti-
mize drilling and stimulation. Mechanical prop-
erties that can be derived from acoustic tools
include bulk modulus, Poissons ratio, Youngs
modulus, yield strength, shear modulus and
compressive strength. These values are com-
puted from compressional-, shear- and Stoneley-
wave measurements.
In highly laminated, argillaceous shales, the
Youngs modulus and Poissons ratio are a func-
tion of the orientation of the measurement rela-
tive to that of the formation bedding planes.
These anisotropic mechanical properties inu-
ence the closure stress and therefore the hydrau-
lic fracture height. Sonic Scanner data are used
to estimate anisotropic mechanical properties
and thus provide a realistic determination of the
in situ rock stresses.
When a large difference occurs between the
vertically and horizontally measured Youngs
moduli, the closure stress will be higher than
that in isotropic rocks. These anisotropic inter-
vals are normally associated with rocks that have
higher clay volume as well as with expandable
clays. These clay-rich rocks are poor candidates
for both horizontal well placement and hydraulic
fracture stimulation. Stresses in these intervals
will be higher, and it will be more difcult to
retain fracture conductivity during production
because the proppant is more likely to embed
into the ductile formation.
Sonic porosity is another acoustic measure-
ment that is benecial in shale analysis. For
shales, sonic porosity is usually much lower than
neutron porosity. This is a function of the high
clay-bound water volume that is common in
shales. When the sonic porosity is much higher
than the neutron porosity, this may indicate that
gas, rather than water, is present in the pore
spaces. When the sonic and neutron porosity val-
ues are similar, the shale may be oil prone.
Log analysts also use wireline borehole image
logs, such as those from the FMI fullbore forma-
tion microimager tool, to identify the presence of
natural and drilling-induced fractures and to
dene their orientation and concentration.
Interpreters can determine from these data
whether the natural fractures are closed (miner-
alized) or open. Engineers use this information to
optimize lateral placement and select perfora-
tion cluster locations along a horizontal wellbore.
By analyzing drilling-induced fractures, they can
also infer the state of near-wellbore stresses.
Although there are methods for acquiring pet-
rophysical data in lateral wellbores, most wells
are drilled vertically and logged with a full suite
of tools prior to drilling the lateral section. Data
can be acquired using LWD tools, which have the
added benet of allowing directional drillers to
optimally steer the bit into potential sweet spots.
The geoVISION imaging-while-drilling service
can provide resistivity along with images for bed-
ding and fracture identication; the SonicScope
multipole sonic-while-drilling tool can provide
acoustic measurements.
The measurements from these various tools
can be combined in an integrated display such as
the shale montage log provided by Schlumberger.
The formation properties are presented using a
common platform, and geologists can directly
compare the quality of the rocks (next page).
Free and adsorbed gas are computed and pre-
sented in units of scf/ton, a common unit of mea-
surement in coal mining operations. Some
operators prefer adsorbed, free and total gas to
be presented in Bcf/mi
2
. Gas resource concentra-
tion presented in a Bcf/section helps quantify the
total potential for a prospective shale reservoir.
43223schD7R1.indd 7 11/17/11 7:44 PM
Autumn 2011 47
>
Shale montage log. Water saturation and porosity are computed from conventional resistivity (Track 2) and porosity (Track 3) logging tools. The ECS tool
provides mineralogy (Track 5) and matrix density (Track 4) for improved porosity computation and differentiates TOC from porosity (pink shading). The grain
density of the rocks can also be computed and used to correct the density porosity (Track 4). Data from core samples provide Langmuir isotherms for gas
storage capacity and conrm computed data to ensure model-based outputs such as matrix and bulk density, water saturation, total porosity and TOC
(circles, Tracks 4, 6 and 7) are valid. Fluid saturations, corrected for lithology, are presented in Track 6. Geologists use total GIP, adsorbed gas and free gas
(Track 8), to determine the potential for the reservoir. Additional features of the shale montage log are the numerical outputs in Tracks 4, 6, 7 and 8, which
allow the geologist to read values directly from the log. For example, at XX,350 ft, the effective permeability (red numbers, Track 7) is 313 nD and the
cumulative free gas volume (blue numbers, Track 8) is 32 scf/ton. In this interval, contrary to most organic shales, the gamma ray (Track 1) is not reading in
excess of 150 gAPI.
XX,300
Depth,
ft
Hole Rugosity
1/ft 0 1
XX,350
XX,400
XX,450
10-in. Array
ohm.m 0.2 2,000
20-in. Array
ohm.m 0.2 2,000
30-in. Array
ohm.m 0.2 2,000
60-in. Array
ohm.m 0.2 2,000
Sonic Porosity
% 40 10
Corrected Neutron Porosity
% 40 10
Neutron Porosity
(Limestone)
% 40 10
Density Porosity
(Limestone)
% 40 10
90-in. Array
ohm.m 0.2 2,000
Matrix Density
Gas Matrix Difference
g/cm
3
Matrix Density (Core)
g/cm
3
2.0 3.0
Bulk Density (Core)
g/cm
3
2.0 3.0
ECS Matrix Density
g/cm
3
2.0 3.0
ElanPlus Matrix Density
g/cm
3
2.0 3.0
Corrected Bulk Density
g/cm
3
2.0 3.0
Density Correction
g/cm
3
0.25 0.25
Bulk Density
g/cm
3
2.0 3.0
Effective Porosity
%
Gas Saturation
%
Water Saturation (Core)
% 100 100
Water Saturation
% 100 100
Total Porosity (Core)
% 40 0
Bound Water
% 40 0
Free Water
% 40 0
Effective Porosity
% 40 0
Total Porosity
% 40 0
ElanPlus TOC
%
Permeability
nD
TOC (Core)
% 0 25
Shale Permeability
mD 1 0.00001
Intrinsic Permeability
mD 1,000 0.01
Gas Permeability
mD 1,000 0.01
TOC Schmoker Method
% 0 25
TOC ELANPlus Method
% 0 25
TOC TerraTek Method
% 0 25
Free Gas
scf/ton
Total GIP
Bcf/mi
2
Adsorbed GIP
Bcf/mi
2
0 60
Total GIP
Bcf/mi
2
0 60
Adsorbed Gas
scf/ton 0 400
Total Gas
scf/ton 0 400
Resistivity
Bit Size
in. 4 14
Caliper
in. 4 14
Gamma Ray
gAPI 0 150
Water
Moved Hydrocarbon
Gas
Water Saturation
Bound Water
Gas Free Gas
Water
TOC
Moved Water
Illite
Montmorillonite
Bound Water
Quartz
Mineralogy
Pyrite
Calcite
Kerogen
Chlorite
43223schD7R1.indd 8 11/17/11 7:44 PM
48 Oileld Review
The integrated product, in addition to charac-
terizing the petrophysical and geomechanical
properties of the reservoir, helps engineers deter-
mine the depth to land the lateral for horizontal
drilling (below). The preferred strategy is to drill
in the minimum horizontal stress direction,
which is perpendicular to the maximum horizon-
tal stress.
Quality Rocks
A study in 2007 concluded that fewer than 30% of
Barnett Shale wells would be protable at com-
modity pricing levels that existed at that time.
10

Much of the data were taken from wells that were
completed while operators were still learning
how to properly exploit shales. Production log
(PL) data from several Barnett Shale wells indi-
cated that 30% of the perforations provided 70%
of the total gas ow, and in some wells, 50% of the
perforations were not owing at the time of log-
ging.
11
Taking into account all the basins, the PL
study showed that approximately 30% of perfora-
tion clusters were not contributing to production.
These statistics illustrate that, when log data are
lacking, drilling blindly and hydraulically fractur-
ing geometrically staged intervals may not be the
optimal strategy.
Prior to drilling, geologists and engineers
should identify layers that have superior reser-
voir and geomechanical characteristics, then
drill and complete within these high-quality
intervals. Shale resource plays typically cover
large geographic area and their log characteris-
tics may not change much laterally across the
basin. However, subtle, and sometimes not so
subtle, lateral heterogeneity within these
sequences results in areas with characteristics
that promote better production and hydraulic
fracture stimulation. These sweet spots include
zones with high gas potentialgood reservoir
quality (RQ)and those that can be optimally
stimulatedgood completion quality (CQ).
Geologists build detailed models to simulate the
reservoir and attempt to identify the parts of the
reservoir with the best RQ and CQ. These models
can be rened as wells are drilled and more data
become available.
Geologic features, especially naturally occur-
ring fractures, inuence well productivity.
Knowledge of fracture density and orientation
and in situ stress properties can help engineers
make decisions on well placement and spacing,
as well as optimize a fracture stimulation pro-
gram. Conventional reservoirs can be drained
across long distances, but recovery from resource
plays depends on optimal well spacing and maxi-
mizing the fracture stimulated volume.
The local stress regime is important for both
drilling and stimulation. Stresses along the
wellbore are a function of tectonic forces, depth
and formation thickness in addition to changes
resulting from previous stimulation and produc-
tion of nearby wells. For optimal wellbore stabil-
ity, drilling should be oriented in the direction of
minimum principal stress. It is important to
understand how the reservoir, including existing
natural fracture networks, will react as it is
stressed during drilling, stimulation and produc-
tion. Changes in the reservoir from production
and induced stresses will inuence stimulation
stage organization, perforation placement and
well spacing decisions. This type of information
can be quantied in geomechanical studies.
Geomechanics is a branch of engineering that
applies solid mechanics, mathematics and phys-
ics to predict how rocks respond to external
forces.
12
Following the lead from mining and civil
engineeringdisciplines that have long used
geomechanics to predict and avoid catastrophic
consequencesdrilling and production engi-
neers are more frequently applying these con-
cepts to reservoir development. These practices
call for measuring and estimating stress and
examining how materials respond to stress.
A regions initial state of stress results from
its tectonic and sedimentary history. Stresses
are also induced by a variety of processes com-
mon to oileld operations, including the effects
of drilling and changes in uid pressure caused
by injection and production. All of these effects
can be simulated using 3D and 4D earth models,
allowing engineers to predict behavior of the res-
ervoir in response to drilling, stimulation and
future production.
Geophysicists and engineers at Schlumberger
have built a mechanical earth model in the Petrel
seismic-to-simulation software that integrates
data from ECLIPSE reservoir simulation soft-
ware. Engineers use these models for well plan-
ning and to determine initial stress states.
By coupling the ECLIPSE 3D simulation with
the VISAGE Modeler, geophysicists can create a
4D model, which simulates changes in the magni-
tude and orientation of downhole stresses that
>
Landing the lateral. Characterizing the reservoir is a crucial step in developing gas shales, but drilling
the lateral relies on more than petrophysics and geochemistry. Key parameters that control completion
quality (CQ) must be taken into account. The bulk modulus is one indicator of CQ. Rock of higher
modulus is more likely to possess lower clay content and therefore have a lower closure stress along
with a lower propensity for proppant embedment during production. This gure shows a well prole
against a projection of geologic structure and bulk modulus. The rst two-thirds of the lateral were in
the target interval. The toe section is in rocks of lower bulk modulus, which are poorer producers. Not
only was the stimulation of the toe section more difcult than the rest of the lateral, production from
the low bulk modulus intervals was less than that from the better quality rock.
3.4 10
6
3.2 10
6
3.0 10
6
2.8 10
6
2.6 10
6
2.4 10
6
2.2 10
6
2.0 10
6
Bulk modulus, psi
43223schD7R1.indd 9 11/17/11 7:44 PM
Autumn 2011 49
occur over time (above). The full production his-
tory for a single well, multiple wells or an entire
eld can be simulated and visualized using the
combination of ECLIPSE and VISAGE software.
The identication of potential drilling, stimu-
lation and production hazards is another crucial
piece of information gleaned from seismic data.
Existing faults can be especially troublesome
when an operator is hydraulically fracturing a
shale interval. A fault can effectively dominate
the fracture growth and redirect all the energy of
the treatment into the fault system and out of the
target zone. Subsequent fracture stages may grow
into previously stimulated faults, adding little to
the total stimulated volume. Faults also serve as
conduits that direct the hydraulic fracture treat-
ments into zones that produce water, which can
potentially kill or greatly reduce gas production.
Based on knowledge gained from 30 years of
Barnett Shale development, engineers have
learned to account for several considerations
when developing resource plays. These include
knowing the present-day maximum horizontal
stress direction to determine optimal wellbore
direction; quantifying natural fracture density,
nature and orientation relative to the maximum
horizontal stress direction; having sufcient
knowledge of geomechanics to design comple-
tions that promote maximum hydraulic fracture
surface area and complexity; and understanding
the fracture-to-fracture interference from both
multiple stages and multiple wells.
13
Completion
engineers must balance the cost of stimulation
and drilling against the addition of wells or frac-
ture stimulation stages. These decisions are
greatly aided by the 3D and 4D models.
Hydraulic Fracturing
Success in developing the Barnett Shale has been
attributed in part to the use of cost-effective
slickwater fracture treatments.
14
However, slick-
water is not the only type of treatment utilized
during development and may not be the correct
uid choice for some shale types.
15
Gelled, gas-
assisted and hybrid fracture stimulations have all
been tested in the Barnett Shale as well as in
other shale plays.
16

There is no single solution for all shale reser-
voirs. For instance, although slickwater has been
an effective technique in the Barnett Shale, the
lower sand-carrying capacity of these systems, as
compared with gelled systems, and the propen-
sity for proppant settling can limit ow capacity
for induced fractures. The limited contact or loss
of conductivity within the stimulated area can
also cause production rates to plummet.
10. Berman A: Whats New in Exploration, World Oil
Online (November 2007), http://208.88.130.69/Article.
aspx?id=38918 (accessed October 11, 2011).
11. 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,
June 1216, 2011.
>
Geomechanics in well design. The VISAGE Modeler uses seismic data (top right), 3D models created in the Petrel program (bottom left, bottom right) and
data from well logs and core samples (top left) to construct a geomechanical model that includes present-day stresses and mechanical properties (center).
Once the 3D model has been built, the element of time can be added, creating a 4D model. Changes in stresses affect drilling and impact the effectiveness
of the stimulations, which greatly inuences productivity.
Oilfield Review
AUTUMN 11
Shale Fig. 11
ORAUT11-SHL 11
Well Logs, Cores, Cluster Analysis,
Mechanical Earth Model
Structure and Faults Horizons, Gridding
Seismic Inversion, ECLIPSE Data,
Properties Modeling with HRA
Heterogeneous Rock Analysis Processing
12. Zoback MD: Reservoir Geomechanics. New York City:
Cambridge University Press, 2007.
13. King GE: Thirty Years of Gas Shale Fracturing: What
Have We Learned? paper SPE 133456, presented at the
SPE Annual Technical Conference and Exhibition,
Florence, Italy, September 1922, 2010.
14. Slickwater fracturing uids are composed of water
and a polymer (usually polyacrylamide) for lowering
the friction pressure when pumping the uid
through tubulars.
15. King, reference 13.
16. Hybrid fracture stimulation often begins with low-
viscosity uids to create a complex network. At the end
of the stimulation, uids capable of carrying heavy
concentrations of proppant are used to open the
near-wellbore region.
43223schD7R1.indd 10 11/17/11 7:44 PM
50 Oileld Review
Another concern with slickwater fracturing is
shale uid sensitivity because some clay types
have a tendency to swell when they come into
contact with water. Gas and foam uids seem to
offer an ideal alternative to slickwater and gelled
systems; however, they are not extensively used
because of their higher cost, limited availability
and the fact that they are traditionally deployed
for smaller jobs. The conundrum faced by opera-
tors is to nd the best system for the specic shale
while minimizing the expense of trial and error.
17

Recently, scientists at the Schlumberger
Novosibirsk Technology Center in Russia devel-
oped a novel approach to hydraulic fracture
design. The HiWAY ow-channel hydraulic fractur-
ing technique (see Open-Channel Fracturing
A Fast Track to Production, page 4), developed
after years of modeling and testing, has been suc-
cessfully applied in a number of environments.
Initial results in Eagle Ford Shale test wells have
shown increased production rates of 32% to 37%
compared with those from similar nearby wells.
The often marginal economics of resource plays
can be greatly enhanced by such improvements
in production.
Another common practice in shale fracturing
programs is to divide the lateral wellbore into
equally spaced segments (above left). This pro-
cess ignores the vertical and lateral heterogene-
ity of the formation and has resulted in cases of
signicant waste of fracturing capital.
18
However,
completion engineers can design programs using
data acquired during drilling to improve the
stimulation program.
Pathway to Success
In large part because of success in the Barnett
Shale, E&P activity in North America has been
dominated by gas shale exploration for the past
decade. Initially, however, operators were not
certain that success in the Barnett Shale could
be replicated with other shale plays. The
Fayetteville Shale, which was discovered by
Southwestern Energy Company, was one of the
rst reservoirs where operators attempted to
expand gas shale horizons beyond the Barnett
Shale. The companys success led to a rush to
develop shale resources elsewhere.
The Fayetteville Shale is located in the north-
ern Arkoma basin and southern Ozark region in
the central US (left). The Southwestern Energy
acreage in northern Arkansas, USA, encompasses
a Mississippian-age marine shelf deposit that
ranges from 50 to 550 ft [15 to 168 m] in thick-
ness at depths of 1,500 to 6,500 ft [457 to 1,980 m].
The play features complex structural geology as
a result of Precambrian and Cambrian rifting
>
Microseismic monitoring of multiple-stage hydraulic fracturing across a lateral. Analysis of
microseismic data, as in this StimMAP hydraulic fracture stimulation diagnostics presentation, can
provide operators with information about the effectiveness of hydraulic fracturing treatments. In this
example, ve stages were pumped from the treating well (red line) while monitored from a second
well (green line with location of geophones shown as green circles). The rst stage (yellow dots)
at the toe of the lateral creates a complex fracture network; however, the second stage (blue dots)
overlaps the stimulated volume from the rst stage. The third and fth stages (red and magenta dots)
develop complex fractures. The fourth stage (cyan dots) has a preferential direction that leaves some
of the formation untouched. It is important to identify these variations in fracture quality to optimize
future stimulation designs, well placement and well spacing. Tools such as StimMAP LIVE real-time
microseismic fracture monitoring service may allow completion engineers the opportunity to adjust
operations during the execution of the job to improve the effectiveness of the treatment.
Z,200
Y,800
Y,400
Y,000
D
e
p
t
h
,

f
t
3,600
3,200
2,800
2,400
2,000
Horizontal departure, ft
1,600
1,200
800
400
0
X,600
X,200
>
Fayetteville Shale, Arkoma basin. Southwestern Energy developed a Fayetteville Shale play located
in north central Arkansas, USA. The companys acreage (red box) includes formations with complex
structure as a result of regional rifting and uplift.
Oilfield Review
AUTUMN 11
Shale Fig. 14
ORAUT11-SHL 14
Arkoma basin
Fayetteville
Shale
Caney Shale
Oklahoma
Arkansas
Louisiana
Texas
Barnett
Shale
Ouachita
Mountains
Fort
Worth
basin
U N I T E D S T A T E S

43223schD7R1.indd 11 11/17/11 7:44 PM
Autumn 2011 51
followed by compressional tectonics during the
Ouachita orogeny of the Late Paleozoic era. The
rifting generated large, northeast-striking nor-
mal faults, whereas the later tectonic events are
characterized by north-striking thrust faults.
The present-day maximum horizontal stress ori-
entation, a product of later relaxation, is north-
east to southwest.
19

The reservoir is subdivided into upper, mid-
dle and lower Fayetteville intervals. The upper
Fayetteville has streaks of high gas-filled
porosity and abundant natural fractures, which
can be stimulated at low treatment pressures.
The middle Fayetteville has relatively high
concentrations of illite and smectite clays and
is characterized by higher fracture gradients
and lower effective porosity. The lower
Fayetteville is subdivided into three zones; the
middle layer is the main target interval of the
three because of its low clay content and high
gas-filled porosity. Natural fractures, both
open and mineralized, are present throughout
the lower Fayetteville interval.
Because of a wide range of well production
performance, Southwestern Energy initiated a
multidisciplinary study to characterize the
Fayetteville Shale and identify the main produc-
tion controls.
20
The study covered three areas of
about 10 mi
2
[26 km
2
] each. A 3D earth model,
built using the Petrel seismic-to-simulation soft-
ware, was a crucial component in the analysis. To
build the model, geologists systematically inte-
grated data from a number of sources including
petrophysics, seismic surveys, geomechanics and
fracture models (above). In addition, a single-
phase uid model was used with the ECLIPSE
reservoir simulation shale gas module. The 3D
model included a dual porosity option for history-
matching and production forecasting.
To build the 3D earth model, geologists
started by constructing a geologic model,
following a workow that used data from core
samplesX-ray diffraction, scanning electron
microscope, thin sections, vitrinite reectance,
pyrolysis and TRA data. Next, they added cali-
brated log data that included triple-combo open-
hole logs, geochemical logs, Sonic Scanner and
DSI dipole shear sonic imager logs and FMI well-
bore images, which were acquired from vertical
pilot wells. Geochemical data from the ECS tool
provided key mineralogic information and were
calibrated with core data.
17. King, reference 13.
18. King, reference 13.
19. Ramakrishnan H, Peza E, Sinha S, Woods M, Ikeocha C,
Mengel F, Simon Y, Pearce P, Kiester J, McKetta S and
Jeffers J: Understanding and Predicting Fayetteville
Shale Gas Production Through Integrated Seismic-to-
Simulation Reservoir Characterization Workow,
paper SPE 147226, presented at the SPE Annual
Technical Conference and Exhibition, Denver,
October 30November 2, 2011.
20. Ramakrishnan et al, reference 19.
>
Living earth model workow and input. Engineers and geologists build the 3D earth model by integrating data from multiple sources. They use geologic data
from cores and well logs along with seismic data to construct a static model. Reservoir simulation software is used to populate the model with data from
vertical and horizontal wells. Engineers create well spacing plans and design hydraulic stimulation programs to maximize production. Microseismic data can
be used to validate the model and improve future fracture stimulation and well planning. The reservoir model is updated as new data become available.
Oilfield Review
AUTUMN 11
Shale Fig. 5
ORAUT11-SHL 5
Stimulation
Planning
Reservoir Model
Geologic Analysis
Seismic
Analysis
Stimulation
Monitoring
Reservoir
Simulation
Static Model
Construction
43223schD7R1.indd 12 11/22/11 11:14 PM
52 Oileld Review
Parameters commonly associated with shale
resource evaluation, such as Langmuir iso-
therms, TOC, initial GIP and adsorbed versus
free gas, were measured and then extrapolated
from a single-well scale to a basinwide scale
using 3D surface seismic surveys. Because lat-
eral changes in petrophysical properties across
shale basins are usually subtle, this extrapola-
tion was accomplished without sacricing the
variability of the ner resolution data. Geologists
were also able to correlate specic lithologies to
natural fractures. Using production logs, engi-
neers could correlate lithofacies to intervals
with natural fractures that exhibited higher ow
rates of gas after fracture stimulation.
Engineers next developed a structural model
using seismic data that was further rened using
log-derived geomechanical data. Downhole data
included fracture gradient, Poissons ratio, Youngs
modulus and natural fracture density. A fracture
model based on discrete fracture networks was
created from image logs, which were calibrated
with core data and microseismic events. Engineers
took a multiple pseudo-3D hydraulic fracture
model approach to predict fracture characteristics
for each stage of each well in the study. They
modeled half-length, fracture height, fracture con-
ductivity variations and elements of fracture
geometries and then used these data in the reser-
voir simulation model. Natural fracture orienta-
tions, based on interpretations from image logs,
were included in the model.
Engineers further developed the structural
model by populating it with reservoir properties
from the geologic model. The properties that
were used are similar to those of conventional
reservoirs and include porosity, permeability and
hydrocarbon saturation. However, for shale plays,
the porosity is calibrated to core data and may
come from the average of several different
sources. Permeability may be derived from core
data, although the ultralow permeabilities of
shales make a direct measurement practically
impossible. Thus, engineers applied a core-
derived porosity-permeability transform to down-
hole log data to compute permeability.
The 3D earth model included ow simulation
as well. Developing a ow simulation model for
ultratight rocks requires knowledge of the stress
regime because the orientation of the grids used
in the model can affect the computed results. In
the case of the Fayetteville Shale simulation, the
grid cells were created with one side parallel to
the maximum horizontal stress direction and the
other side parallel to the horizontal well paths.
This information was provided by the structural
model, which indicated that the present-day
maximum horizontal stress direction is northeast
to southwest.
The Bigger Picture
By integrating these various data, engineers cre-
ated the full 3D earth model, which helped them
characterize the Fayetteville Shale reservoir
(left). The model was used to develop better drill-
ing and completion programs, as, for example,
when it was used to analyze and improve hydrau-
lic fracture stimulations.
Hydraulic fracturing can be the most expen-
sive process in the completion and development
of resource plays, and it has the greatest impact
on effectively producing the well. Completion
engineers concluded, based on results of post-
stimulation production logs, that there was a
direct correlation between in situ stresses and
hydrocarbon production (next page, top right).
Thus, knowledge of the stress gradient along the
lateral wellbore provided completion engineers
with a tool for optimizing stimulation programs.
In addition, knowing the horizontal stress orien-
tation assisted drilling engineers in selecting bet-
ter wellbore trajectories for drilling laterals.
Borehole stability is improved when drilling in
the direction of minimum horizontal stress.
Because some wells had lateral lengths in excess
of 5,000 ft [1,500 m], signicant variations in reser-
voir properties were encountered along the lateral
wellbores. The 3D earth model provided engineers
with more-accurate design parameters for the frac-
ture stimulation program than would have been pos-
sible by projecting properties of a vertical borehole
section some distance away. Following each stimula-
tion treatment, the fracture model was updated using
production history-matching.
As the study progressed, engineers observed
that fracture height growth varied in different
parts of the basin. In the early stages of develop-
ment, microseismic data indicated that fracture
stimulation extended from the lower Fayetteville
Shale into the uppermost Fayetteville Shale
interval. However, they discovered that some
wells did not produce as expected because the
stimulation was not reaching the upper layer.
>
Fayetteville Shale 3D static geomodel. To understand the variation of properties across a gas shale
play, it is necessary to integrate a variety of different data types into the evaluation, acknowledging
the 3D spatial location of each data item. Within the context of a 3D Petrel earth model, seismic data,
well logs, faults and structural surfaces can be combined to develop a framework model into which
petrophysical, mechanical and natural fracture properties can be populated, such as the one shown
here that was developed for the Fayetteville Shale. The operator can compare property cubes, such
as those shown, to other geospatial information in 3D and optimize well placement and hydraulic
fracturing. Incorporating production results allows the intrinsic interconnectivity of these different
data types to be revealed and aids in the understanding of the impact of different production drivers
on well performance. These property cubes also form the basic building blocks for single- and
dual-porosity reservoir simulation models.
Oilfield Review
AUTUMN 11
Shale Fig. 15
ORAUT11-SHL 15
43223schD7R1.indd 13 11/22/11 11:15 PM
Autumn 2011 53
Reservoir engineers attributed the differ-
ences in fracture growth to areas with higher
clay content in the middle Fayetteville Shale
interval. The presence of more clay resulted in
higher in situ stresses, which inhibited fractures
from growing into the upper layers. Engineers
identied these anomalies by analyzing the 3D
earth model.
Tangible results from the optimization pro-
cess, which included drilling and stimulating lon-
ger lateral sections, performing optimized
fracture stimulations and increasing operational
efciencies, were evident in the continuous
improvements seen from 2007 to 2011 (below).
The number of days to drill a well decreased by
more than 52% even as the lateral length of the
average well increased by more than 84%. Average
production increased dramatically, almost seven-
fold, but well costs remained virtually unchanged
during the period.
Resource plays are capital intensive, but
because they typically cover large geographic
areas, operators benet from economies of scale
and operating exibility. Identifying and develop-
ing the sweet spots signicantly improve the eco-
nomics and ROI. The ultimate measure of success
is production: In October 2011, Southwestern
Energy reported gathering approximately
2.0 Bcf/d [56.6 million m
3
/d] in gas production
from the Fayetteville Shale play.
>
Production contribution obtained from PL analysis. The lateral well (blue
line) passed through low (red) and high (blue) in situ stress intervals. The
hydraulic fracture stimulation consisted of ve stages with three perforation
clusters per stage (green ovals). Poststimulation PL data were then acquired.
The red lines extending below each perforation cluster represent gas
production normalized to the maximum contributor. The length of each red
line represents normalized ow. The rst two stages (dashed red ovals), at
the toe of the well, were in the high stress zones. Only 16% of the ow
came from these stages. The other 84% of production came from the three
stages (dashed yellow ovals) located in the intervals of lower stress.
Engineers can use this type of information to identify sweet spots and avoid
expensive fracture treatments in zones with low production potential.
Oilfield Review
AUTUMN 11
Shale Fig. 17
ORAUT11-SHL 17
Stage 1
9%
1.05
1.00
0.95
0.90
0.80
0.70
0.85
0.75
Stage 2
7%
M
i
n
i
m
u
m

i
n

s
i
t
u
s
t
r
e
s
s

g
r
a
d
i
e
n
t
,

p
s
i
/
f
t
Stage 3
25%
Stage 4
27%
Stage 5
32%
>
Continuous process improvement. Over a four-and-a-half year period, from 2007 to 2011, Southwestern Energy reduced days to drill (dark blue) by 52%,
even though the lateral length was increased by more than 84% (pink). Well costs (dark red) were at to slightly lower during the period but the companys
nding and development costs (F&D, light blue) were signicantly reduced during the period. Production (gold) and reserves (green) greatly increased
during the study period. (Data for 2011 are for the rst six months of the year.)
2007 2008 2009 2010 2011
17
14
12
11
8
Days to Drill
2,657
3,619
4,100
4,528
4,909
2007 2008 2009 2010 2011
Lateral Length,
ft
2.05
1.21
0.69
0.86
F&D Costs,
US$ per Mcf
2007 2008 2009 2010
53.5
134.5
243.5
350.2
Production,
Bcf
2007 2008 2009 2010
716
1,545
3,117
4,345
Reserves,
Bcf
2007 2008 2009 2010
2.9 2.9
2.8 2.8
3.0
Well Cost,
US$ millions
2007 2008 2009 2010 2011
Oilfield Review
AUTUMN 11
Shale Fig. 18
ORAUT11-SHL 18
43223schD7R1.indd 14 11/22/11 11:16 PM
54 Oileld Review
Postfracture Evaluation
The nal, and often neglected, component in
optimizing production from resource plays is pro-
duction analysis. Considerable effort goes into
determining reservoir qualities and developing
complex models to identify zones within the res-
ervoir with the greatest potential. Drilling engi-
neers analyze wellbore properties and use
geosteering to direct the well into the areas per-
ceived to have the best RQ and CQ. Completion
engineers design stimulation programs to maxi-
mize production by concentrating on rocks with
the best CQ. These efforts can identify most likely
candidates for production, but they rarely
address small-scale variations that exist within
the resource. PL data provide empirical proof of
production and offer the potential for identifying
reservoir characteristics that differentiate zones
with the greatest potential (below).
A recent large-scale study from six US gas shale
basins demonstrated the benets of PL data for
resource play development.
21
The study attempted
to highlight characteristics that engineers could
incorporate in development workows to improve
overall efciency. One disturbing nding was that
in only 20% of the wells were all the perforation
clusters contributing to production. In two Arkoma
basin Woodford Shale horizontal wells, only half of
the clusters were producing gas.
Shale resources are viewed by some as large
monolithic structures; however, heterogeneity
caused by variations in rock properties occurs
vertically at extremely small scales in these res-
ervoirs. Furthermore, the presence of natural
fractures can introduce large variations in rock
mechanical properties within a small area. If this
variability is not accounted for in stimulation
design, wells may not achieve expected results.
Engineers may be able to use PL data to cor-
relate gas production with differences in rock or
geomechanical properties. Wellbore geometry and
completion practices that may affect production
can be observed in PL data as well. The PL study,
which included data from more than 100 wells,
assessed various common practices used in shale
gas wells and evaluated their effects on production.
Wellbore trajectoryInitially, most horizon-
tal shale wells were drilled uphill with deviations
that exceeded 90 degrees. This was done to facili-
tate gravity drainage of fracture uids to the heel
of the lateral and help unload the uids more
quickly. In some shale plays, this practice has
been replaced by drilling laterals on structure,
regardless of trajectory. However, the ideal tra-
jectory is more than 90 degrees with minimal
sumps and doglegs while remaining in the target
zone. One trend evident from the data is that
wells with high ow rates can effectively unload
the fracture uids regardless of trajectory and
can overcome detrimental effects related to well-
bore geometry.
Fracture stagingWell productivity improves
with the number of fracture stages. An increase in
the number of stages often correlates with longer
laterals, thus contact with more of the reservoir.
However, the study indicated that stimulating
shorter sections of the lateral has a positive
impact on production, even when the data are
normalized for increased lateral lengths. Thus,
while the length of laterals has increased over
the last few years, the segment lengths that are
stimulated per stage have decreased.
The study analyzed the effects of fracture
stage spacingthe distance between stages.
Engineers observed that for most shale plays,
spacing in the range of 100 ft [30 m] resulted in
the best production. They concluded that any
increase in stress associated with previous stimu-
lation treatments did not adversely affect the
productivity of subsequent stages when this spac-
ing was utilized. An exception to this nding was
in the Barnett Shale, where there was no clear
correlation between stage spacing and productiv-
ity. Engineers attributed this difference to the
structural environment of the Fort Worth basin.
Because natural fractures in the Barnett Shale
tend to run orthogonal to the hydraulically
induced fractures, a complex fracture network
can be achieved during stimulation. Therefore,
closely spaced stages provide less benet com-
pared with those of other shale plays. This nding
illustrates the importance of understanding the
geologic setting of a reservoir and its impact on
completion quality. The practical result is that an
>
Production log comparison. These PL logs are from two different Woodford Shale wells in the Arkoma
basin. The red and gold vertical lines along the well trajectory indicate perforation cluster locations.
Red shading represents gas in the wellbore; blue shading indicates water. The PL data from Well A
(top) show variable production with only three clusters contributing signicantly to the total gas
produced and three clusters producing most of the water (measured rates for individual clusters
shown in top track). Three clusters (gold lines) are not contributing anything. Production in Well B
(bottom) is more uniform. Most clusters are contributing equivalent amounts, although the rst and last
clusters (gold lines) are not contributing. Although water is present at the toe, no water production is
indicated in Well B. (Adapted from Miller et al, reference 11.)
Oilfield Review
AUTUMN 11
Shale Fig. 19
ORAUT11-SHL 19
8,400 8,800 9,200 9,600 10,000 10,400
8,200 8,600 9,000 9,400 9,800 10,200 10,600 11,000
Measured depth, ft
Measured depth, ft
Water
Gas
Water
Gas
8,050
8,100
7,900
7,950
0
0
15,000
1,500
TVD, ft
0
3,000
0
1,000
Rate for
each cluster,
bbl/d equivalent
Cumulative
production rate,
bbl/d equivalent
Rate for
each cluster,
bbl/d equivalent
Cumulative
production rate,
bbl/d equivalent
TVD, ft
Well A
Well B
43223schD7R1.indd 15 11/22/11 11:18 PM
Autumn 2011 55
optimized stimulation methodology in one basin
may not transfer to another.
The PL study also analyzed fracture stage
designs by comparing theoretical production to
measured production. Theoretical production
was dened as the production rate if all stages
produced equally. The study found that for every
two wells completed, with an average of eight
stages per well, there was at least one stage that
contributed nothing. In addition, multiple
perforation clusters were not contributing at the
theoretical level. Again, these ndings had basin-
specic attributes. The percentage of stages pro-
ducing at least half of their theoretical rate
ranged from 18% for the Marcellus Shale in the
northeast US to 33% for the Haynesville-Bossier
Shale in north Louisiana and east Texas.
Perforation clusterAnalysis of perforation
cluster productivity provided few clear trends,
and results were often basin specic. Whereas
the Barnett Shale was effectively stimulated with
a single cluster per stage, Woodford Shale wells
with four clusters per stage outperformed those
with eight clusters per stage. Apart from these
basin-specic characteristics, the results illus-
trate that it is very difcult to effectively stimu-
late multiple perforation clusters per stage. For
example, almost half of the clusters in the wells
with six perforation clusters per stage were not
contributing at the time of production logging.
This contrasts with an average of 20% of the clus-
ters not contributing when only two perforation
clusters were included per stimulation treat-
ment. The current trend is toward fewer perfora-
tion clusters per stage.
Along with the number of perforation clus-
ters, the spacing of the clusters also affects pro-
duction. The cluster spacing for the wells in the
study ranged from 36 to 421 ft [11 to 128 m]. The
results suggest that a cluster spacing of less than
125 ft [38 m] produces superior results. Operators
have recognized the apparent correlation
between cluster spacing and productivity, and
more recently, developed shale plays are utilizing
closer cluster spacing. As before, however, this
trend did not hold true for the Barnett Shale,
where acceptable results were achieved even
with spacings in excess of 175 ft [53 m].
Applying the lessonsThe analyses indi-
cate some ndings are basin specic, others are
well specic, and some indicate that there is no
established trend. PL data provide information
about what is happening in the well at the time
of logging, but geologists and engineers may be
able to correlate the differences in production
with lateral variations in reservoir characteris-
tics. For example, image logs can provide infor-
mation about lateral variations, such as
fractures, mineralogy and changes of the stress
regime. However, these data are infrequently
available after the initial information-gathering
phase of development in shale plays as opera-
tors focus on operational efciencies and cost
reduction. This makes it difcult to correlate PL
results to formation properties.
In one horizontal well, engineers ran an FMI
tool in open hole and PL logs after completion
and stimulation (above). Geologists derived
microresistivity logs from the FMI image data
from which they could qualitatively determine
mineralogy. Low resistivity often corresponds to
high clay content and high resistivity corre-
sponds to better quality rocks. Lower clay con-
tent rocks have lower in situ stress and higher
Youngs modulus, and they are more likely to
retain fracture conductivity during production.
Of ve fracture stages in the well, the three
stages that were performed in zones identied
as low clay content outperformed the two in
clay-rich rocks. Engineers can optimize staging,
isolation of high stress intervals, cluster place-
ment and proppant scheduling when these data
are available.
Evolution or Revolution?
To develop resource plays, the oil and gas indus-
try faces challenges that go beyond technology,
and these challenges should not be minimized.
There are political, environmental and percep-
tual issues that have little to do with drilling and
producing hydrocarbons from the shale forma-
tions that are found around the globe. The indus-
try focuses primarily on the technical elements,
although the others are crucial considerations.
E&P companies have proved that, after three
decades of development, resource plays are via-
ble targets for exploration. Rocks that were once
considered practically worthless from a produc-
tion standpoint are now supplying the US with
abundant supplies of natural gas. In an environ-
ment with low natural gas prices brought on by
the success of organic shale development, opera-
tors will need to continue using innovation, tech-
nology and engineered solutions to improve
protability while developing resource plays.
What has been demonstrated over the last
decade is that the revolution that began in the
Barnett Shale has not stopped there. As technol-
ogy evolves, the revolution is poised to become
a global endeavor (see, Shale Gas: A Global
Resource, page 28). TS
21. Miller et al, reference 11.
>
Resistivity data, clay content and fracture initiation. FMI borehole microresistivity images (left) can provide qualitative mineralogy information. Clay-rich
intervals correspond to low resistivity (darker brown) in Stages 1 and 2 and have higher fracture initiation pressures (right). The resistive mineralogy
observed in Stages 3, 4 and 5 (lighter brown, left) corresponds to rocks with lower clay content. Production from all stages was normalized by the
production from Stage 5, which had the highest gas production, and the ow rates for the other four stages are presented as a fraction of this value. The
low resistivity, clay-rich shales underperformed the resistive shale intervals. These results illustrate the benet of stimulating horizontal shale wells in
intervals that have good RQ and CQ characteristics. (Adapted from Miller et al, reference 11.)
Shallow
Resistivity,
ohm.m
Fracture Initiation
Pressure,
psi/ft
Normalized
Production,
best producer = 1.0
Stage 1
Stage 2
Stage 3
Stage 4
Stage 5
8
5
632
581
724
1.2
1.1
1.0
0.9
0.9
0.28
0.22
0.80
0.86
1.0
F
r
a
c
t
u
r
e

i
n
i
t
i
a
t
i
o
n

p
r
e
s
s
u
r
e
,

p
s
i
/
f
t
Shallow resistivity, ohm.m
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0 100 200 300 400 500 600 700 800
Stage 3 Stage 4 Stage 5
Oilfield Review
AUTUMN 11
Shale Fig. 20
ORAUT11-SHL 20
Stage 1 Stage 2
43223schD7R1.indd 16 11/22/11 11:18 PM