Jointing and Fracturing in the

Marcellus Shale
A discussion of natural fractures, or joints, present in the Marcellus Shale
and the hydraulic fractures that are induced during unconventional gas
drilling to extract natural gas.
PA L E O N T O L O G I C A L R E S E A R C H I N S T I T U T I O N
T H E S C I E N C E B E N E A T H T H E S U R F A C E
M A R C E L L U S S H A L E I S S U E N U M B E R 5 A U G U S T 2 0 1 1
1259 Trumansburg Road
Ithaca, New York 14850
www.museumoftheearth.org/marcellusshale
MUSEUM OF THE EARTH
AT T H E PA L E O NT O L O G I C A L R E S E A R C H I N S T I T U T I O N
D I D Y O U K N O W ?

Fractures are commonly

known by geologists as
joints.

Te diference between
joints and faults, like the
San Andreas system of
faults, is that faults have
experienced marked direc-
tional movement along the
fracture, whereas joints only
experience separation.

Even though water is

used, the hydro in hy-
drofracking doesnt refer to
water. It references the
hydraulic pressures used
to fracture the rock.
Introduction
Te Marcellus Shale is a natu-
ral gas-bearing rock found beneath
the surface of the Earth in parts of
Pennsylvania, Ohio, West Virginia,
and New York. It was deposited
in a shallow sea that covered these
states nearly 400 million years ago.
Shale is a type of sedimentary rock
formed from very tiny, fat grains
packed together like decks of cards
strewn across a table. Grains found
in the Marcellus Shale accumulated
together as mudsalong with a large
quantity of organic materialin the
shallow sea. Some of the organic
material trapped in the mud has ma-
tured (changed chemically with heat
and pressure) into natural gas. Te
natural gas now present in the Mar-
cellus Shale remains mostly trapped
between the grains and in natural
fractures already present in the shale.
Economically extracting the natu-
ral gas tightly trapped by the shale
grains requires a process called hy-
draulic fracturing. Tere are various
types of hydraulic fracturing technol-
ogy, but all hydraulic fracture jobs
aim to create or extend a network of
joints, or fractures, in the shale that
allows the trapped natural gas to fow
into a natural gas well.
Marcellus Shale Issue Number 5 July 2011 / 1
Figure 1. Taughannock Falls
Aerial view of the creek near Taughannock Falls near Ithaca, NY. Note the joint pattern in the
rocks. Te J1 and J2 joint sets intersect at roughly 90, creating a pattern of repeating corners
in the rock. In these joints, the other side of the rock has fallen away, but beneath Earths sur-
face, the joints are planar cracks that exist within the rock mass. Tis same jointing pattern is
exhibited at depth in the Marcellus Shale.
Tere are environmental con-
cerns about hydraulic fracturing in
the Marcellus Shale, in part because
the shales surrounding and including
the Marcellus have already experi-
enced natural fracturing, or jointing,
as a part of their geologic history. It
has been suggested that stimulating
the Marcellus with hydraulic fractur-
ing may cause the pre-existing frac-
tures to connect and create a pathway
that leads drilling mud, hydraulic
fracture fuid, formation water, and
methane gas to drinking water aqui-
fers or water sources.
The Role of
Hydraulic Fracturing
At present, hydraulically frac-
turing a well requires 3-5 million
gallons of water, mixed with sand
grains of diferent sizes and a num-
ber of chemicals, to be pushed into
the rock at very high pressures. Te
chemicals perform various functions
in the fracturing process, like killing
bacteria and reducing the viscosity of
the hydraulic fracturing fuid. While
some of the chemicals are considered
benign, others can be considered
toxic: even when diluted with the
large quantities of water used in the
process. Te fuid is injected into the
well, and when the pressure of the
fracturing fuid exceeds the pressure
of the rock at depth, the rock breaks
at weak points, usually along planes
of weakness. Te direction of these
planes of weakness can be predicted
to some extent based on the physi-
cal characteristics of the rock and
the orientation of stresses the rock is
under. During a hydraulic fracture,
the rock breaks, creating fractures,
or joints, which would, without
something to prop them open, natu-
rally close when the water pressure
is subsequently lowered. Te sand
or sand-like grains in the fracturing
fuid act as propping agents (prop-
pants) remaining in the joints after
the water is removed and the pressure
is lowered to keep the joint networks
open for natural gas extraction.
In sum, hydraulic fracturing
2 / Marcellus Shale Issue Number 5 July 2011
operates to open and maintain path-
ways for fuid to fow where it other-
wise could not, and the process uses
chemicals that would be unsafe if
they entered drinking water aquifers
or streams. If hydraulic fracturing
were conducted near drinking water
aquifers or if the induced fractures
were sufciently long or connected
to long natural fractures, it is con-
ceivable that the process could also
connect new fracture networks with
pre-existing ones to create a pathway
for the natural gas, and the water and
chemicals used to release the gas, to
fow into drinking water aquifers.
Finally, there are other substances
trapped in the rocks around the Mar-
cellus Shale, like brines and natural
gas, which could contaminate drink-
ing water aquifers if a pathway was
created during the hydraulic fractur-
ing process. For these reasons, it is
important to understand how the
naturally occurring joints and the
fractures stimulated by oil and gas
companies could interact in order
to assess the risk of drinking water
contamination in areas where drilling
is done.
Natural Fractures
Te Marcellus Shale is a naturally
fractured rock because of the com-
bination of the quantity of organic
matter trapped in the rock and the
historical plat tectonic activity that
occurred in the area. Te conversion
from organic material to natural gas
in the Marcellus created pressure in
the fuids trapped in the rocks, which
helped created natural fractures in
the rock, called joints. Tese joints
were further exacerbated by the colli-
sion of plates during the Alleghanian
Orogeny, a mountain-building event
that began around 350 million years
ago. Tis combination of forces cre-
ated the majority of the joints in the
Marcellus Shale. Because the joints
were formed from the same processes,
most of the joints in the Marcellus
Shale run in roughly parallel sets in
one of two directions.
Tese two sets of joints are called
J1 and J2, and each has unique
characteristics. Joints in the J1 set
run east-northeast throughout the
Marcellus Shale. Te J2 joint set
runs generally north-northwest and
cuts across the J1 joint set. Te J1
joint set happens to coincide with
the modern underground stresses
found throughout the Northeast.
Joints in the J1 joint set are spaced
more closely together than those in
the J2 joint set. Joints in the J2 set
are more likely to have been healed,
which means the joint has been flled
in with minerals and is no longer a
pathway for gas fow. Tese healed
joints are still relatively weak because
the mineral cement gluing them
together is weaker than the original
bedding planes of the rock, and are
therefore places where new fractures
could grow.
In addition, the Marcellus Shale
was buried by numerous other rock
layers in the last 400 million years,
some of which have been eroded
since the Alleghanian Orogeny
(in part by the numerous glacia-
tions that have sculpted the current
landscape of the Northeast). Tese
erosional forces have brought the
Marcellus Shale closer to the surface
of the Earth and thereby decreased
the downward pressure provided by
overlying rocks. As overlying weight
that had compressed the ground was
released, much like a memory foam
mattress, the decrease in pressure
resulted in the J3 joint set. Tey, too,
run roughly parallel, and are referred
to as release joints and unloading
joints.
1
Release and unloading joints cre-
ated by the various unloading of rock
and ice are only present at or near
Earths surface. Tis is why J3 joints
are not found at the depth of Marcel-
lus Shale natural gas extraction.
2

How Stimulated
Hydraulic Fractures Form
Hydraulic fractures form when
the fuid pressure within the rock
unit exceeds the external pressures
pushing on the rock unit. To under-
stand how hydraulic fractures could
propagate (be stimulated) in Mar-
cellus Shale natural gas drilling, drill-
ing engineers must understand the
pressures acting upon the rock.
Beneath the surface, there is
vertical pressure from the weight of
the overlying rock. Tere are also
horizontal pressures in all directions
pressing on the Marcellus Shale.
Tese pressures are the result of
plate tectonic forces interacting. Te
Earths outer shell is made of very
large plates, like puzzle pieces, which
slowly jostle each other. Te interac-
tions of these plates can build moun-
tains and cause earthquakes, but in
the Marcellus Shale region, they are
weaker than the vertical pressure
caused by the weight of the overly-
ing rocks. Hydraulic fractures grow
perpendicular to the direction (the
plane) of minimum principle stress.
In the Marcellus Shale region, frac-
tures made by hydraulic fracturing at
depth will have a vertical orientation
Marcellus Shale Issue Number 5 July 2011 / 3
and will separate against the mini-
mum horizontal stress.
Because the vertical stress changes
with depth, the orientation of the
hydraulic fractures also changes as the
fractures move closer to the surface. In
sufciently shallow rock, the direction
fractures propagate can leave the verti-
cal plane and propagate on an angle
and could even approach the horizon-
tal plane. So as fractures propagate
into shallower rock above, they be-
come less likely to propagate vertically
and more likely to extend laterally
into the rock. Te depth at which this
occurs varies by rock type and thick-
ness and by the change in related pres-
sures and is difcult to estimate.
While much of the Devonian
rock surrounding the Marcellus Shale
is also shale, there are many layers
of silt, limestone, and sandstone, as
well. Even the diferent shales sur-
rounding the Marcellus are unique
in their composition. Tese diferent
kinds of rock layers are called litholo-
gies, and each lithology has its own
unique set of physical characteristics
that react diferently to the pressures
that cause hydraulic fractures. Each
lithology has been deposited under
slightly diferent marine and terres-
trial conditions, which can afect its
grain size and provenance (where the
grains came from and what types of
grains are present), composition of
the mineral glue cementing grains
together, and the quantity of organ-
ics present in the
shale. For example,
some layers contain
evidence of hurri-
cane-force storms,
and others record
times of little or no
sediment deposi-
tion. Some shales
are black due to a
high quantity of
organic material, while others with
less organic material are grey. Te
depositional conditions determine
the physical characteristics of the
lithology and thus how it may react
to hydraulic fracturing. In general,
limestones have higher fracture
thresholds than shales and can act as
barriers to vertical fracture growth.
3
As a fracture grows above or be-
low the Marcellus Shale, it encoun-
ters other lithologies, and this can
play an important role in determin-
ing the ultimate extent of a hydraulic
fracture. Lithologies often change
abruptly, represent-
ing events or chang-
ing environments in
Earth history; such
distinct boundaries
between litholo-
gies are often more
weakly cemented
together than the
layers, or beds, with-
in each lithology.
Because of this, these boundaries
are considered planes of weakness.
Layers, or beds, within a lithology
Africa
Eurasia
Indo-Australia
Antarctica
Nazca
Pacific
South
America
North
America
Cocos
Caribbean
Somalia
Arabia
Philippines
Scotia
Juan de Fuca
1
0
2
3
2
9
7
9
1
0
6
46
3
2
12
6
8
8
14
54
5
9
7
4
9
9
2
1
1
0
5
1
0
2
1
1
Figure 2. Map of current plate
boundaries throughout the world.
Arrows indicate the speed and direction of
plate movement, represented in millimeters
per year. Dots indicate the presence of earth-
quakes, which are concentrated along plate
boundaries. Notice how the dots concentrate
when two plates are moving toward each
other, and are less common where plates are
moving away from each other. Tis image
is courtesy of Incorporated Research Institu-
tions for Seismology (IRIS), and is part of
their educational series of pamphlets about
earthquakes that can be accessed online at
http://www.iris.edu/hq/publications/bro-
chures_and_onepagers/edu.
The stiffness of
the Marcellus Shale is
not uniform, so some
areas are more prone
to longer fractures
than others.
4 / Marcellus Shale Issue Number 5 July 2011
Relatively highly fractured rock
Minimum
principal
stress
Minimum
principal
stress
Relatively unfractured rock
1 2 3 3a
can also be planes of weakness, but
these are usually more cohesive than
boundaries.
Boundaries play an important
role in the growth of hydraulic frac-
tures, because individual fractures
tend to have greater difculty propa-
gating straight through them. Often
the boundary between two litholo-
gies acts as a bufer, and the some
of the pressure causing a hydraulic
fracture is diverted to planes of weak-
ness, radiating horizontally along the
boundary (because it is weaker than
the surrounding rock). As a result,
there is less pressure available to cre-
ate a fracture in the overlying lithol-
ogy. When the pressure creating the
fracture succeeds in extending into
the overlying lithology, the energy is
distributed in this new rock layer dif-
ferently based on the characteristics
of the rock, and the direction and
pattern of the fracture can change.
Field geologists commonly note that
a set of fractures end at a boundary
between two lithologies, and some-
times use this characteristic to orient
themselves at a new outcrop.
Hydraulic fracture characteristics
are dramatically impacted by the mea-
sure of the modulus (ratio of stress to
strain) of the rock. If the modulus is
large, the rock is considered stif. In
stif rock units, fractures grow long
and narrow away from the source of
additional pressure, and in less stif
materials, fractures are wider and
vertically shorter because the pres-
sures causing the fractures can pen-
etrate and dissipate further laterally
(into planes of weakness like bedding
planes and existing fractures) within
the rock unit. Shales are usually very
stif, but highly fractured black shales,
like the Marcellus, are generally much
less stif than unfractured grey shales.
Te stifness of the Marcellus Shale is
not uniform, so some areas are more
prone to longer fractures than others.
However, when a hydraulic fracture
hits the less stif, already fractured
materials in the Marcellus Shale, the
energy can move laterally within the
shale and, as a result, cause shorter,
wider fractures.
4,5,6
Understanding Current
Fracture Networks within the
Marcellus Shale
Te Marcellus Shale is a natu-
rally-fractured gas shale, and the body
of literature on the characteristics of
those fractures is expected to grow
signifcantly as additional data are
gathered. In some respects, the only
way to gather additional data is to
drill more wells and examine rock
carefully removed from the well (a
process called core sampling), so
some things could remain unknown
until after additional drilling has tak-
en place. While this pamphlet high-
lights much of the base level under-
standing of Marcellus Shale jointing,
some generalizations will be refned
only after engineers have collected
and analyzed data from numerous
wells and core samples throughout
the Marcellus.
Tere are a number of physical
Figure 3. Fracture Propagation
Scenarios 1, 2, and 3 represent theoretically likely directions of fracture propagation result-
ing from increasingly higher levels of fracture-inducing force. As force increases in relatively
unfractured rock, fractures propagate perpendicular to the direction of maximum principle
stress. When they encounter a rock layer boundary (which is naturally weaker), the energy
forcing the fracture dissipates laterally, making it harder for a fracture to continue across
boundaries. In more fractured rock (scenario 3a), even with the same pressures as seen in 3,
pre-existing fractures can cause energy forcing a fracture to dissipate in other directions. Be-
cause of the many boundaries between rock layers and pre-existing fractures in scenario 3a,
the fractures propagated are wider and shorter than they would normally be if the same rock
was relatively unfractured.
Marcellus Shale Issue Number 5 July 2011 / 5
characteristics of the Marcellus Shale
that are not homogeneous. For
example, some geologists have hy-
pothesized that the lower part of the
Marcellus Shale, sometimes called
the Union Springs shale, which has
a higher organic
content and
larger quantities
of pyrite (fools
gold) than the
rest of the Mar-
cellus, may have
more jointing
(as a result of
the maturing
organics in
concert with
the Alleghanian
orogeny). It
would thus have
a higher natural
interconnec-
tion of fractures
through which
gas can move
and a higher
natural gas yield. Because the eastern
part of the Marcellus Shale has been
subjected to greater plate tectonic
forces, jointing may be more abun-
dant in the east. But thinner units
are generally more fractured than
thicker units, and the Marcellus is
thinner in the west than it is in the
east, so jointing could be hypoth-
esized to be more prevalent in the
western portion of the Marcellus
Shale.
6

Te Marcellus Shale is the base
rock unit in a group of shales and
limestones called the Hamilton
Group. Te Hamilton Group is well
known in New York State for the
abundant fossils present in some of
the rock units. While the Marcellus
is not the only shale in the Hamil-
ton Group, it is the only uniformly
thick, gas-bearing, black shale. Black
shales have some unique distinctions
over grey shales, including a higher
quantity of organic material and a
higher amount of
naturally-occur-
ring radioactive
material. Because
of the organic
content of black
shales, the num-
ber of joints in
a square unit of
rock (joint den-
sity) is usually
much higher than
in grey, lower
organic content
shales, with spac-
ing usually less
than one meter
apart in black
shales.
7
Because
of this and other
shale character-
istics, many times the fractures in
black shales consistently tend not to
extend beyond lithologic boundaries
into overlying (or underlying) grey
shales.
Te Marcellus Shale represents
a dynamic system that has been
changing for almost 400 million
years. While there is consensus on
basic fracture patterns likely to occur
in the Marcellus Shale as a result of
stimulation, the shale is not homoge-
neous, and diferent regions will react
diferently to the same hydraulic frac-
turing treatment. Tere will be areas
with concentrated fractures and areas
almost devoid of them, and while
we can predict where those might be
based on our understanding of how
fractures form, there is not yet much
observational data to support those
predictions.
Tracking the Results of
Hydraulic Fracturing
Because hydraulic fracturing is
occurring beneath Earths surface,
the only way to track fractures result-
ing from a hydraulic fracture is to
use proxy evidence in the form of
models, feld tests, and monitoring
production of wells, and incorporat-
ing that information into a body of
knowledge that grows as more wells
are drilled in the Marcellus region.
M O D E L I N G
Models are used to predict the
efectiveness of hydraulically frac-
turing wells in shale gas layers like
the Marcellus, and to inform future
hydraulic fracturing in nearby wells.
Tese models are based on data gath-
ered in the feld which are imported
into software programs to predict the
characteristics of the rock at depth.
Field data usually incorporated into
hydraulic fracture models include
seismic and microseismic data, well
log data, expected pressures through-
out the subsurface, and other rock
characteristics. Models predict how
fracture fuid (and the associated
pressures and fractures already pres-
ent in the rock) will act during hy-
draulic fracturing. When measured
values difer signifcantly from the
modeled values, the fuids are not
behaving in the way the model pre-
dicted they would and one or more
characteristics of the rock have not
been adequately modeled. Tis may
occur if the model itself needs to be
revised or if estimates of rock proper-
While there is
consensus on basic
fracture patterns
likely to occur in the
Marcellus Shale as a
result of stimulation,
the shale is not homo-
geneous, and differ-
ent regions will react
differently to the same
hydraulic fracturing
treatment.
6 / Marcellus Shale Issue Number 5 July 2011
ties where the hydraulic fracturing
is done were not accurate. Tus, to
understand what is happening in the
rock below the surface, modeling is
used in concert with feld testing to
assess most efective hydraulic frac-
turing treatment for diferent regions
in the Marcellus Shale.
8,9,10

F I E L D T E S T S
Common feld tests used initially
to inform and eventually to truth
fracture propagation models include
observing well logs, seismic testing,
sonic logs, gamma ray logs, resistivity
logs, and monitoring fuid injection
and downhole pressures of the well
during hydraulic fracturing, among
other techniques.
11
Seismic testing is
a technique that can be run on the
surface of the Earth, but most tech-
niques require an observational well
to be drilled for monitoring purposes.
A few feld techniques will be dis-
cussed below.
Seismic testing uses seismic
waves, created by thumping the
ground with heavy weights, to read
the subsurface. Te waves travel
through the ground, hit diferent
layers of rock, fractured zones, and
other variations in density beneath
the surface of the Earth, and return
to the surface at diferent times. Te
waves are recorded and analyzed
to create an image of the rock lay-
ers beneath the surface. Tese tests
are usually run along roads or other
straight paths, and interpreted with
other seismic tests to search for exist-
ing fractures and other structures in
the subsurface that could potentially
interact with a hydraulic fracturing
treatment. Because seismic testing
can be done from Earths surface,
it is sometimes used after hydraulic
fracturing to explore the accuracy of
the model of fracturing in predicting
actual fracture patterns.
Microseismic testing works in
much the same way as seismic testing,
but small seismic monitors are placed
at the surface, or shallowly buried,
above the area undergoing hydraulic
fracturing. Tese monitors record
small seismic changes and relay
them, frequently in real time, to the
engineers conducting the hydraulic
fracturing treatment. Tis is the most
efective way to visualize the fractures
that are created during a hydraulic
fracture and evaluate model efective-
ness.
12
Well logs are descriptions of the
rock layers as observed by the drillers
that document the rock type, joint-
ing, and other visual characteristics
of the rock. Sonic logs are similar in
principle to seismic tests, but use son-
ic waves instead of sound waves. Also,
sonic logs are run down a well so
that, instead of from the surface, the
sonic logs record information near
the well bore. Gamma ray logs also
run down the well bore, but these
look specifcally for natural radioac-
tivity, which is much higher in black
shales than in other rocks. Gamma
ray logs help drillers determine the
thickness and depth of the reservoir
rock in the region. Resistivity logs use
various methods to extract the poros-
ity and fuid content of diferent rock
layers.
Observation wells are sometimes
drilled near a well undergoing hy-
draulic fracturing while a region is
being established to collect data on
the efectiveness of hydraulic fractur-
ing. A similar suite of feld tests and
observations can be conducted on an
observation well to more objectively
test the efectiveness of a hydraulic
fracturing treatment, the model used
to predict the fracture behavior, and
other parameters.
13,14
Existing Fracture Networks
Natural gas has been rising to-
ward areas of low pressure since it
formed in the Marcellus Shale, which
is part of the reason for the existing
fractures in the unit. Natural gas has
reached near-surface layers along
fracture networks where the Marcel-
lus Shale (and other shales and lime-
stones) comes within several hundred
feet of the surface. Some of these
seeps were sources for natural gas for
towns during the 19
th
century. Tis is
why methane can also be found ris-
ing from some abandoned, uncapped
wells.
Te NYS DEC has recom-
mended that stimulation more than
2,000 feet beneath the surface is deep
enough to avoid impacts on potable
water sources.
15
Fluid and gas migra-
tion through fracture networks from
below 2,000 feet to the surface is not
theoretically impossible. A recent
study documented the presence of
thermogenic gas in water wells within
close proximity of active gas wells.
16

Termogenic gas, the kind of gas
found in the Marcellus Shale, is the
result of temperature and pressure
changes deep beneath Earths sur-
face that change carbon matter into
methane. It has a diferent chemical
signature than biogenic natural gas,
which is produced by the metabolic
decay of organisms close to Earths
surface. Tis study correlates a high
percentage of thermogenic gas (as op-
posed to a mix of thermogenic and
biogenic gas) contamination with ac-
Marcellus Shale Issue Number 5 July 2011 / 7
tive gas drilling, but does not explain
how the migration from deep rocks
has occurred. Migration pathways are
difcult to trace, and no conducted
studies have been able to confrm or
deny a fracture connection between
deep gas and drinking water aquifers.
Tis study did not record the
presence of fracturing fuid contami-
nation or brine contamination in any
water well in close proximity to an
active gas well, which suggests that
faulty well casings are a more likely
pathway than deep fracture connec-
tions.
17
Te presence of natural gas
and other fuids at the depth of the
Marcellus Shale and in surrounding
layers suggests that existing fracture
networks have not allowed these sub-
stances to escape.
Te study also noted that some
water wells that were not located
near active gas wells had methane
contamination. However, the amount
of methane in these wells was far
less than those water wells near ac-
tive gas drilling, and the chemical
makeup of the methane was a mix
of thermogenic and biogenic natural
gas. Te reason gas and other fuids
may migrate from fractures near the
surfaceleaving a small amount of
mixed methane contaminationbut
not from a mile beneath the surface
is likely related to the number of rock
units acting as barriers to migration
and the degree of discontinuous frac-
tures passing from the Marcellus to
the surface.
In considering existing fracture
networks, it is also important to note
that Marcellus Shale drilling is not the
only type of drilling that has occurred
in NY and throughout the northeast.
Conventional oil and gas drilling has
commonly occurred throughout NY
for decades, and some of these wells
were improperly abandoned. At least
70,000 wells have been drilled in
NYS and only 30,000 are accounted
for by the DEC.
18

When a well is no longer pro-
ducing commercial quantities of oil
or natural gas, a company abandons
the well. Abandoning a well properly
involves removing some of the casing
and other surface equipment, plug-
ging the well with cement so that
it is no longer a conduit for fuids
migrating from beneath the surface,
and then reclaiming the surface. If
wells are improperly abandoned, they
can act as a connection between the
surface and the depth of the well. In
most cases, improperly abandoned
wells are older and relatively shallow,
but when the Marcellus Shale is be-
ing exploited at shallower depths, like
in central NY, improperly abandoned
wells could potentially connect a
fracture network to a groundwater
source. Improperly abandoned wells
were considered when the DEC sug-
gested that stimulation of Marcellus
Shale should be at least 3,000 feet
beneath the surface to sufciently
decrease risk of contamination of po-
tential groundwater sources.
Fracture networks are not the
only consideration of how fuids can
travel through rock layers to the sur-
face. Because of the greater pressures
beneath the surface, fuid naturally
fows toward areas of lower pressure
if a pathway (through fractures or be-
tween connected pore spaces) exists.
Permeability (amount of connected
pore space) is measured in mil-
liDarcies (after Darcys Law, which
describes how fuids fow through po-
rous media). Te rate at which a fuid
can fow through a rock, based on
the amount of fuid present, pressure,
fracture presence, and rock perme-
ability, is called conductivity.
Te conductivity of the Marcel-
lus Shale is relatively low, but varies
between 0.000011 and 0.00059 feet
per day (just over 7/1000
th
of an inch
of movement per day at its fastest),
and the surrounding units also have
variable but low conductivities.
17
It is
more common for fuid to fow hori-
zontally than vertically in the Marcel-
lus Shale because of the pressures at
depth and the existence of horizontal
bedding planes. However, it is not
impossible for pressures at depth to
create a gradient angled to some de-
gree toward Earths surface in some
regions of the Marcellus or surround-
ing units. As the rock unit changes,
the direction of fow and conductiv-
ity can change.
Very little data is available on the
range of conductivities in the rocks
surrounding and including the Mar-
cellus Shale in locations being exam-
ined for natural gas drilling, although
more is gathered by gas companies
as additional wells are drilled.
18
Te
impacts of the short-term increase in
pressure caused by hydraulic fractur-
ing conductivity from the depth of
the Marcellus Shale to the surface is
not well understood, and could pos-
sibly play a role in the migration of
gas and fracture fuids to adjacent
rock units
19
although are not likely to
provide sufcient pressure change to
migrate gas and fracture fuids all the
way into drinking water aquifers.
Te conductivity of the Mar-
cellus Shale and its overlying units
does not currently allow signifcant
methane migration from depths being
considered for natural gas extraction
to the surface. If that were the case,
8 / Marcellus Shale Issue Number 5 July 2011
the quantity of natural gas stored
throughout the Marcellus Shale
would have been greatly diminished
by now. We might assume, therefore,
that the conductivity provided by
existing fracture networks connecting
the Marcellus to much shallower units
is very low. Based on this reasoning, it
appears unlikely, though not theoreti-
cally impossible, that gas and fracture
fuid could migrate to potential sourc-
es of drinking water in this way.
Although it is expected that hy-
draulic fractures would propogate
upward little further than the top of
the Marcellus, one ought to consider
the potential impact if there were
sufcient pressure from hydrofractur-
ing to connect to and open existing
fractures up to very shallow depths,
or to connect them with shallow, im-
properly abandoned wells. Fractures
remain open during pressurization
from hydrofracturing, for a few hours
at a time for up to a few days, then
remain open if proppant sand has
entered the fractures. If fractures did
connect to existing fracture networks,
they would not have sufcient prop-
pant to keep them open after hydrau-
Table 1. Geological Characteristics of Coal and Marcellus Shale
depth below surface
Aquifer location
Surrounding rock
lithologies
Thickness of
desired unit
Original deposition
of rock unit
Depth of targeted
formations for
hydraulic fracturing
Pre-existing fractures
in desired rock unit
Rock Characteristics
Coal Bed
Methane Extraction
06,000 ft; most between 4004,000 ft
mostly within 450 ft
variable; other coal seams, shale, sandstone
inches to 250 ft thick lenses
in terrestrial settings, among braided streams,
wetlands, and deltas
06,000 ft; dependent upon target depth
cleating present; degree of cleating
primarily based on coal type and depth
below surface
3,00010,000 ft
01,000 ft (most less than 700 ft)
mostly shale, some silt and limestone lenses;
varies regionally
one contiguous layer between 50250 ft
thick; regionally variable
bottom of shallow, continental sea
3,00010,000 ft; dependent upon
depth of unit
fractures present; degree of fracturing
based on organic content, rock unit
thickness, and proximity to tectonic
action; varies regionally
Comparison of the geological characteristics of coal bed methane extraction and Marcellus Shale methane extraction.
Coal bed information compiled from data from Powder River, San Juan, Raton, Piceance, and Uinta Basins.
21

Marcellus Shale information from Hill, et. al., 2002.
12
Marcellus Shale
Methane Extraction
Marcellus Shale Issue Number 5 July 2011 / 9
lic fracturing was complete. To date
migration through such fractures of
fracturing fuids and produced water
has not been documented.
20
Tis may
be because such vertical migration
paths do not or rarely exist, or be-
cause the amount of fracturing fuid
and proppant that can migrate verti-
cally over such a distance is small.
Understanding
Fracture Networks in
Different Rock Units
Diferent rock types, like sand-
stones, shales, and coal seams, have
diferent physical characteristics that
impact the likelihood of the ability of
fractures to grow upward toward the
surface from various depths. When
they are sought after for fossil fuel
extraction, the depth of the rock unit
below the surface, the stifness of the
rock unit being targeted and of the
surrounding lithologies, the regional
stresses on the rocks, the depth of po-
table water, and many other consider-
ations are used to predict the impact
hydraulic fracturing will have on a
particular lithology.
Recently, water well contamina-
tion in coal bed methane extraction
using hydraulic fracturing has been
used to compare the potential for
water well contamination in the
Marcellus Shale. It is important to
learn from the environmental prob-
lems experienced in other active gas
plays; comparisons among diferent
reservoir rocks, however, must also
consider the important geologic
similarities and diferences between
them.
While the organics in the Mar-
cellus Shale were deposited at the
bottom of a shallow sea basin and
were comprised primarily of marine
algae, coal beds are deposited in ter-
restrial environments and are usually
comprised of woody and leafy plant
matter. Coals generally form around
braided streams, wetlands, or delta
environments, and, as a result, an in-
dividual coal layer usually cannot be
traced laterally for a long distance. In-
stead, the environment conducive to
coal formation moves and migrates,
like sand bars in a stream. Te result
is that a coal bed can sometimes be a
series of horizontal lenses interspersed
in a vertical unit of rock. Coal can
also form in one distinct layer, but
even then the thickness of the coal is
highly variable. Coal deposits usually
have a much smaller regional extent
than the Marcellus Shale. Tis is be-
cause they are deposited in areas near
streams, lakes, and rivers, which are
geographically smaller than shallow
continental seas.
As a result of these geologic dif-
ferences, the target depth for coal bed
methane extraction extends from very
shallow depths that sometimes inter-
sect drinking water aquifers to depths
that are similar to those for Marcellus
Shale gas drilling. Te target depth
for gas extraction has a narrow range,
changing verticallyat mostaround
500 ft. in a 25 mile radius. However,
in a coal bed over the same 25 mile
radius, the target depth for methane
extraction can range from around
400 to 4,000 feet, and multiple
depths may be targeted. Shallow coal
beds (less than 450 ft deep) are not
only home to sources of methane, but
sometimes are the source of potential
drinking water.
21
Shallowly buried coal is usually
highly permeable, because as the
organic material matures fractures
form called cleats, which cause mined
coal to appear blocky. Cleats increase
coal permeability dramatically. Tis
is why fresh water can frequently
be found in shallow coal beds, and
why coal bed methane extraction
frequently involves what is called
dewatering the coal. Tis is also
why methane can fow easily within
the coal, especially at shallow depths.
When coal is more deeply buried,
however, the pressure of the overlying
rocks can close the cleats and lower
the permeability of the coal. Natural
gas companies prefer to extract gas
from shallower coal units because
of this. Because coal bed methane is
sought from a wide range of depths,
and because coals have a higher per-
meability, stimulated fractures are
usually intended to connect between
multiple coal layers.
Some similarities can be drawn
between coal bed methane extraction
and Marcellus Shale methane extrac-
tion, but in detail, the similarities are
few. Because of the diferent physical
characteristics that dictate how and
where natural gas is stored in the rock
types, the permeability of the rock
units, and their proximity to poten-
tial drinking water sources, the likeli-
hood of fractures propagating beyond
their target zone(s) in coal bed meth-
ane extraction is much higher than in
the Marcellus Shale.
Summary
Tere has been signifcant com-
munity concern that high volume
hydraulic fracturing of the Marcellus
Shale may allow fuidsincluding
natural gasto rise from connected
fracture networks several thousands
of feet below aquifers. In addition,
10 / Marcellus Shale Issue Number 5 July 2011
the Marcellus Shale has been deeply
fractured by several geologic events
since deposition of the shale. Where
the Marcellus Shale occurs near the
surface, for example in western New
York, natural gas has been known to
seep naturally into drinking water
aquifers and to the surface. Available
evidence suggests that deep shale
(several thousand feet) is not natural-
ly connected to drinking water sourc-
es through fracture networks, likely
because of the discontinuity of frac-
tures across numerous lithologies.
22

It is not known to be impossible for
fuids to migrate to drinking water
sources from deep hydraulic fractur-
ing, but the likelihood of signifcant
migration appears to be very low,
especially when compared to the risk
of faulty casings or surface spills as a
potential source of contamination.
Finally, the relative risk of hydraulic
fracturing varies substantially by local
geological context, including the na-
ture and depth of source rock, lithol-
ogy of overlying rocks, and the nature
of existing fractures and fault net-
works. Any comparison between the
Marcellus Shale and another source
of natural gas must be well framed
and consider these parameters.
23

References
1. Hill, David, Tracy E. Lombardi, and
John P. Martin. 2002. Fractured shale
gas potential in New York. TICORA
Geosciences, Inc., Arvada, Colorado,
U.S.A.
2. Engelder, Terry, Gary G. Lash, and
Redescal S. Uzcategui. 2009. Joint sets
that enhance production from Middle
and Upper Devonian gas shales of the
Appalachian Basin. AAPG Bulletin
v.93, 7, 857-889.
3. Jarvie, D.M., R.J. Hill, T.E. Ruble, and
R.M. Pollastro, 2007, Unconventional
shale-gas systems: Te Mississippian
Barnett Shale of north-central Texas as
one model for thermogenic shale-gas
assessment, AAPG Bulletin, v.91:4, pp.
475-499.
Potable H
2
O

level
Potable H
2
O

level
GAS BEARING COAL
MARCELLUS
SHALE
Tully Limestone
Moscow Shale
Ludlowville Shale
Skaneateles Shale
Onondaga Limestone
Generic Coal Bed Methane
(Modeled in part after Powder River Basin)
Typical Marcellus in NYS
H
A
M
I
L
T
O
N
G
R
O
U
P
G
E
N
E
S
S
E
E

G
R
O
U
P
SURFACE
1000'
2000'
3000'
4000'
5000'
6000'
COAL BED
CLEATING
(detail)
Figure 4. Comparing
Coalbed and Marcellus Shale
Methane Extraction
Comparing a generalized depiction of the
environment from which coal bed meth-
ane is extracted to a generalized depic-
tion of the Marcellus Shale in New York
State. 1) Gas-bearing coal deposits can
commonly be found at multiple depths
in the same region, including as shallow
as within the water table, and all of the
deposits are of interest to gas companies.
Coal lenses are cleated, which increases
their permeability. 2) In NYS, the Mar-
cellus Shale averages around 100ft in
thickness, andin areas presently being
considered for natural gas drillingis
located between 3,000 and 5,000ft be-
neath the surface. Potable water deposits
in NYS are not deeper than 850ft, but
sources that can be treated and become
potable extend to 1,000ft beneath Earths
surface. See Table 1 for additional detail.
Marcellus Shale Issue Number 5 July 2011 / 11
Authored by Trisha A. Smrecak and the PRI Marcellus Shale Team
Illustrations by J. Houghton unless otherwise attributed.
Partnering Organizations include
Cornell Cooperative Extension (naturalgas.cce.cornell edu),
New York State Water Resources Institute
(wri.cornell.edu), Cornell University Department of
Earth and Atmospheric Sciences, and
Cornell University Agricultural Experiment Station.
Funded by NSF GEO No. 1016359.
Any opinions, ndings, and conclusions or
recommendations expressed in this material
are those of the author(s) and do not neces-
sarily reect the views of the National
Science Foundation.
4. OConnor, K.M. and J.A. Siekmeier,
2009. Infuence of rock mass stifness
variation on measured and simulated
behavior. Te 37
th
U.S. Symposeum
on Rock Mechanics (USRMS), June
7-9, 1999, Vail, CO.
5. National Research Council, 2010,
Management and Efects of Coalbed
Methane Development and Produced
Water in the Western United States,
Appendix A Hydraulic Fracturing
White Paper, National Academies
Academic Press, Washington, D.C.,
217 p.
6. Hill, David, Tracy E. Lombardi, and
John P. Martin. 2002. Fractured shale
gas potential in New York. TICORA
Geosciences, Inc., Arvada, Colorado,
U.S.A.
7. Sankaran, S., M. Nikolaou, and M.J.
Economides, 2000, Fracture Geometry
and Vertical Migration in Multilayered
Formations from Inclined Wells, SPE
63177, 8 pp.
8. National Research Council, 2010,
Management and Efects of Coalbed
Methane Development and Produced
Water in the Western United States,
Appendix A Hydraulic Fracturing
White Paper, National Academies Aca-
demic Press, Washington, D.C., 217 p.
9. Carter, B.J., J. Desroches, A.R. In-
grafea, and P.A. Wawrzynek, 2000,
Simulating fully 3D hydraulic fractur-
ing, in Modeling in Geomechanics, Ed.
Zaman, Booker, and Gioda, Wiley
Publishers, 730 pp.
10. Jenkins, C., A. Ouenes, A. Zellou,
and J. Wingard, 2009, Qunatifying
and predicting naturally fractured
reservoir behavior with continuous
fracture models, AAPG Bulletin, v.93,
11, 1597-1608.
11. Chipperfeld, S., 2008, Hydraulic
Fracturing: Technology Focus, JPT
v.60, 3. 56-71.
12. Digital EnergyJournal, 2011, CGG
Veritas real time microseismic moni-
toring, http://www.fndingpetroleum.
com/n/CGG Veritas real time micro-
seismic monitoring/892015e8.aspx
(accessed June 10, 2011).
13. Hill, David, Tracy E. Lombardi, and
John P. Martin. 2002. Fractured shale
gas potential in New York. TICORA
Geosciences, Inc., Arvada, Colorado,
U.S.A.
14. Raymond, M.S. and W.L. Lefer,
2006,Oil and Gas Production in Non-
Technical Language, PennWell Corp.,
Tulsa, OK, 255 pp.
15. Hyne, N.J., 2001, Nontechnical
Guide to Petroleum Geology, Explora-
tion, Drilling, and Production, 2
nd
ed.,
PennWell Corp., Tulsa, OK, 598 pp.
16. New York State Department of Envi-
ronmental Conservation, 2009, Draft
supplemental generic environmental
impact statement on the oil, gas, and
solution mining regulatory program.
http://www.dec.ny.gov/energy/58440.
html (accessed March, 2011).
17. Raymond, M.S. and W.L. Lefer,
2006,Oil and Gas Production in Non-
Technical Language, PennWell Corp.,
Tulsa, OK, 255 pp.
18. Osborn, S.G., A. Vengosh, N.R. War-
ner, and R.B. Jackson, in press, Meth-
ane contamination of drinking water
accompanying gas-well drilling and
hydraulic fracturing, PNAS.
19. New York State Department of Envi-
ronmental Conservation, 2009, Draft
supplemental generic environmental
impact statement on the oil, gas, and
solution mining regulatory program.
http://www.dec.ny.gov/energy/58440.
html (accessed March, 2011).
20. ICF International, 2009, Technical
Assistance for the Draft Supplemental
Generic EIS: Oil, Gas and Solution
Mining Regulatory Program Well
Permit Issuance for Horizontal Drill-
ing and High-Volume Hydraulic
Fracturing to Develop the Marcellus
Shale and Other Low Permeability Gas
Reservoirs, Agreement #9679, NY-
SERDA, Albany, NY.
21. Myers, 2009, Review and Analysis of
DRAFT Supplemental Generic Envi-
ronmental Impact Statement On Te
Oil, Gas and Solution Mining Regula-
tory Program Well Permit Issuance for
Horizontal Drilling and High-Volume
Hydraulic Fracturing to Develop the
Marcellus Shale and Other Low-Perme-
ability Gas Reservoirs, Natural Resourc-
es Defense Council, New York, NY.
22. Osborn, S.G., A. Vengosh, N.R. War-
ner, and R.B. Jackson, in press, Meth-
ane contamination of drinking water
accompanying gas-well drilling and
hydraulic fracturing, PNAS.
23. National Research Council, 2010,
Management and Efects of Coalbed
Methane Development and Produced
Water in the Western United States,
National Academies Academic Press,
Washington, D.C., 217 pp.
12 / Marcellus Shale Issue Number 5 July 2011