**Abstract**

Natural fractures seem to be ubiquitous in shale gas plays. It is often said that their presence is one of the most critical factors in defining an economic or prospective shale gas play. Many investigators have presumed that open natural fractures are critical to gas production from deeper plays such as the Barnett, as they are for shallower gas shales such as the Dev‐ onian shales of the northeastern US and for coal bed methane plays. A common view on production mechanisms in shales is "because the formations are so tight gas can be pro‐ duced only when extensive networks of natural fractures exist" [6]. However, there is now a growing body of evidence that any natural fractures that do exist may well be filled with calcite or other minerals and it has even been suggested that open natural fractures would in fact be detrimental to Barnett shale gas production [9].

Commercial exploitation of low mobility gas reservoirs has been improved with multi-stage hydraulic fracturing of long horizontal wells. Favorable results have been associated with large fracture surface area in contact with the shale matrix and it is here that the role of natu‐ ral fractures is assumed to be critical. For largely economic reasons hydraulic fracturing for increasing production from shale gas reservoirs is often carried out using large volumes of slickwater injected at pressures/rates high enough to create and propagate extensive hy‐ draulic fracture systems. The fracture systems are often complex, due essentially to intersec‐ tion of the hydraulic fractures with the natural fracture network. After hydraulic fracturing operations the injected water is flowed back. Typically, only a small percentage (on the or‐ der of 20 to 40%) is recovered.

In this paper we investigate the role played by natural fractures in the gas production proc‐ ess. By applying a new model of the production process to data from many shale gas wells across a number of shale plays in North America, we can for the first time begin to sort out assertion from inference in the role that these fractures play. Specifically, we are able to esti‐

© 2013 Walton and McLennan; licensee InTech. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

mate the magnitude of the fracture surface through which gas is actually produced. We are able to demonstrate that although it may be commensurate with the expected surface area of open natural fractures for the ultra-low permeability shallow gas shales, it is in fact com‐ mensurate with a very much smaller area for the deeper gas shales such as the Barnett. Fur‐ thermore, given a typical value of the matrix permeability, almost all the gas between the fractures would have been produced in an uncharacteristically short period of time unless the producing fractures are 100s of feet apart. The implications of these findings for comple‐ tion and stimulation strategies will be discussed.

state (PSS) regime is applicable (see the later discussion for definitions of this flow regime),

The Role of Natural Fractures in Shale Gas Production

http://dx.doi.org/10.5772/56404

329

then gas production rate should decline exponentially.

**Figure 1.** Stages in gas and water production from coal (after [1]).

few cm.

There are three essential elements to a model of the CBM production process:

**1.** Transport in the coal matrix, modeled as a diffusive process using Fick's diffusion law. In principle the gas concentration in the coal matrix satisfies a diffusion equation, but it is common to use a pseudo-steady-state (PSS) approximation similar to that proposed by Warren and Root [2] in their dual porosity formulation of production from naturally fractured reservoirs. For example, King et al [3] used the PSS simplification to reduce computing time and because after a period of time the numerical accuracy was deemed to be quite acceptable. We have estimated from King's data that the PSS solution is valid beyond about 40 days, which is much shorter than the typical duration of the production process. We note, however, that this time scale depends on the assumed values of the diffusivity in the matrix and on the spacing of the cleats, assumed to be of the order of a

**2.** Desorption at the cleat/matrix interface as characterized by the Langmuir isotherm

**3.** Transport of water and free gas in the cleat system. To avoid difficulties in defining the configuration of the cleat system, it is common to adopt a dual porosity description in which the cleat system is treated as a continuum with system characteristics analogous to those of a porous medium. Two-phase flow in this system can for the most part be

### **1. Introduction**

One of the main premises of our investigation of the production processes in shale gas plays is that the industry's mental picture of the process remains very much influenced by the concepts developed in the 1990s of the production processes in coal bed methane resources and in shallow shale gas plays. It is appropriate therefore that we begin our discussion of shale gas production characteristics by reviewing this early work.

#### **1.1. Coal bed methane**

Coal is a heterogeneous and anisotropic porous medium, characterized by two distinct porosity systems:


Gas is stored essentially by adsorption in the coal matrix; very little is stored as free gas in the micropores or as free gas or dissolved gas in the connate brine in the cleats. In the subsurface the cleats are usually filled with water, some of which must be produced to the surface to facilitate gas production.

The conventional view of the production process divides it into three stages (see Figure 1):


state (PSS) regime is applicable (see the later discussion for definitions of this flow regime), then gas production rate should decline exponentially.

**Figure 1.** Stages in gas and water production from coal (after [1]).

mate the magnitude of the fracture surface through which gas is actually produced. We are able to demonstrate that although it may be commensurate with the expected surface area of open natural fractures for the ultra-low permeability shallow gas shales, it is in fact com‐ mensurate with a very much smaller area for the deeper gas shales such as the Barnett. Fur‐ thermore, given a typical value of the matrix permeability, almost all the gas between the fractures would have been produced in an uncharacteristically short period of time unless the producing fractures are 100s of feet apart. The implications of these findings for comple‐

One of the main premises of our investigation of the production processes in shale gas plays is that the industry's mental picture of the process remains very much influenced by the concepts developed in the 1990s of the production processes in coal bed methane resources and in shallow shale gas plays. It is appropriate therefore that we begin our discussion of shale

Coal is a heterogeneous and anisotropic porous medium, characterized by two distinct

**•** Macropores or cleats, slot-like with spacing of the order of 2 cm and width of the order of microns; permeability is stress-dependent but far in excess of the micro-pore permeability. They are often formed by shrinkage of the coal matrix due to dewatering during the

Gas is stored essentially by adsorption in the coal matrix; very little is stored as free gas in the micropores or as free gas or dissolved gas in the connate brine in the cleats. In the subsurface the cleats are usually filled with water, some of which must be produced to the surface to

The conventional view of the production process divides it into three stages (see Figure 1):

**1.** Cleat dewatering, lasting of the order of several years; the water production rate gradually falls as water is removed from the cleats. At the same time more and more gas is produced at increasing rates and the relative permeability to gas in the cleats increases leading to

**2.** Stabilized flow: eventually, most of the water in the cleats has been removed, the cleat fluid pressure bottoms out and the relative permeability to gas levels off. Over this period

**3.** Decline: there is then no more increase in drawdown available to sustain gas production and gas production declines. If the cleat pressure was constant and the pseudo-steady-

**•** Micropores of diameter of the order of nanometers with almost zero permeability.

tion and stimulation strategies will be discussed.

gas production characteristics by reviewing this early work.

**1. Introduction**

328 Effective and Sustainable Hydraulic Fracturing

**1.1. Coal bed methane**

coalification process.

facilitate gas production.

lower pressures and more gas production.

the gas rate slowly peaks.

porosity systems:

There are three essential elements to a model of the CBM production process:


adequately described by Darcy flow. In narrower cleats it may be necessary to include capillary pressure and slippage effects especially at low pressure.

reduce the need for a large fracture surface area and ultimately for the need to propose the

The Role of Natural Fractures in Shale Gas Production

http://dx.doi.org/10.5772/56404

331

Thus there are three essential elements to a model of the production process in the Devonian

**1.** Desorption of gas in the matrix (as characterized typically by the Langmuir isotherm)

**2.** Transport in the matrix towards the fracture network, modeled as Darcy flow even though

The Antrim Shale is a shallow, under-pressured, naturally-fractured shale reservoir with characteristically low matrix permeability, and with adsorbed gas, free gas and mobile water co-existing in the reservoir. A typical Antrim well will produce considerable quantities of water early in life, and as dewatering of the reservoir progresses, water production rates decline and a corresponding increase in gas production is normally observed (as a result of gas desorption with reduced reservoir pressures), similar to a CBM well. In fact, the Antrim shale is often considered to be a hybrid of productive dry gas shale and CBM plays. It has characteristics which are similar to these other unconventional reservoirs, but it is also different in many ways. The Antrim shale is more intensely fractured than the Devonian Shales of the Appalachian Basin, with fracture spacing as close as 1 to 2 ft. Kuuskraa et al [8] have noted that the "intensity and interconnection of the fractures govern the shale's

The typical depth of the Antrim shale is less than about 2000 feet, pore pressure is a few hundred psi and more than 70% of the gas in place is adsorbed, the remainder being stored essentially as free gas in the matrix pores. Peak gas may occur as late as 3 years into production. Production data has been history-matched using similar software to that used for CBM [8]. It was found that fracture spacing of the order of a few inches facilitated a good match with production data. It was stated that if a fracture spacing of 3-6 ft was used (which is compatible with observations from cores and logs), then production would be an order of magnitude lower than observed in existing wells. One possible resolution of this conundrum may lie, as the authors suggest, in detrital silt layers within the matrix that could provide conductive flow paths. (An alternative explanation that remained unconsidered by the authors lies in the use of the PSS approximation for matrix transport, which may be completely invalid in this

This description leaves many issues unresolved, but importantly places the Antrim shale as an intermediate between CBM and the other shales, in that the natural fractures appear to be conductive and initially water-filled, but has free pore gas in addition to adsorbed gas.

existence of a large open fracture network.

the permeability is extremely small.

**3.** Transport of free gas in the fracture system.

**1.3. Devonian Antrim Shale of the Michigan basin**

shales:

natural producibility."

context.)

It is apparent that the cleat or natural fracture system plays a very important part in the production process. The density of the cleats plays two critical roles: first, the close spacing of the cleats reduces the time required for the gas to diffuse to the cleats and, second, it is associated with a high cleat/matrix surface area without which economical gas production would be unlikely. The width of the cleat is a primary influence on the pressure drop in the whole system and therefore on the water and gas production rates. In situ the cleats are usually water-filled and presumably kept open by the pressure of the fluid they contain. The cleats may close somewhat as the pressure falls during production, though this may be more than offset by matrix shrinkage as the gas is desorbed.

#### **1.2. Devonian shales of the appalachian basin**

Gas production from Devonian shales received a great deal of attention in the 1980s and early 1990s as a result of US DoE initiatives. This is well documented in many GRI reports and industry publications. The consensus view is that these reservoirs are highly fractured containing a substantial number of fractures with spacing of the order of 1-10 cm (see, for example, [4]). Luffel et al [5] measured the matrix permeability at less than 0.1 nd. Water content of the Devonian Shale averages 2.5 to 3% of bulk volume and appears to be at irredu‐ cible water saturation. Typical depth is a few thousand feet, pore pressure is less than about 3000 psi and about 50% of the gas in place is adsorbed; there is little or no water production.

Carlson and Mercer [6] summarized the consensus view of the production process as "because the formations are so tight gas can be produced only when extensive networks of natural fractures exist." The extent to which this statement holds for other gas shale plays is debatable, but it has certainly been influential in developing the industry's vision of what is happening downhole.

Gatens et al [7] used a dual-porosity model similar to that formulated by Warren and Root [2] but extended to use the unsteady-state equation instead of the pseudo-steady-state (PSS) equation for matrix flow. Analysis of hundreds of Devonian wells showed that most of the production data fell into the linear transient regime (as we discuss later in this document). Luffel et al [5] obtained a good history match with data by assuming an open fracture spacing of a few feet, while Carlson and Mercer [6] needed a fracture spacing of about 80 ft, both with matrix permeability of less than 0.1 nd. An issue that does not seem to have been addressed however, is whether these fractures, if present, are in fact open and if so how they are main‐ tained open against closure stress. If, like coal cleats, they are initially water-filled, is water production observed? An implicit assumption seems to be that they are open and gas-filled.

Carlson and Mercer [6] proposed that molecular diffusion is the dominant transport mecha‐ nism in the matrix in these extremely tight reservoirs, in which case a matrix diffusion coefficient should be used instead of the matrix permeability. They did not evaluate the consequences of this hypothesis. It remains a possibility that the use of such a coefficient would reduce the need for a large fracture surface area and ultimately for the need to propose the existence of a large open fracture network.

Thus there are three essential elements to a model of the production process in the Devonian shales:


adequately described by Darcy flow. In narrower cleats it may be necessary to include

It is apparent that the cleat or natural fracture system plays a very important part in the production process. The density of the cleats plays two critical roles: first, the close spacing of the cleats reduces the time required for the gas to diffuse to the cleats and, second, it is associated with a high cleat/matrix surface area without which economical gas production would be unlikely. The width of the cleat is a primary influence on the pressure drop in the whole system and therefore on the water and gas production rates. In situ the cleats are usually water-filled and presumably kept open by the pressure of the fluid they contain. The cleats may close somewhat as the pressure falls during production, though this may be more than

Gas production from Devonian shales received a great deal of attention in the 1980s and early 1990s as a result of US DoE initiatives. This is well documented in many GRI reports and industry publications. The consensus view is that these reservoirs are highly fractured containing a substantial number of fractures with spacing of the order of 1-10 cm (see, for example, [4]). Luffel et al [5] measured the matrix permeability at less than 0.1 nd. Water content of the Devonian Shale averages 2.5 to 3% of bulk volume and appears to be at irredu‐ cible water saturation. Typical depth is a few thousand feet, pore pressure is less than about 3000 psi and about 50% of the gas in place is adsorbed; there is little or no water production.

Carlson and Mercer [6] summarized the consensus view of the production process as "because the formations are so tight gas can be produced only when extensive networks of natural fractures exist." The extent to which this statement holds for other gas shale plays is debatable, but it has certainly been influential in developing the industry's vision of what is happening

Gatens et al [7] used a dual-porosity model similar to that formulated by Warren and Root [2] but extended to use the unsteady-state equation instead of the pseudo-steady-state (PSS) equation for matrix flow. Analysis of hundreds of Devonian wells showed that most of the production data fell into the linear transient regime (as we discuss later in this document). Luffel et al [5] obtained a good history match with data by assuming an open fracture spacing of a few feet, while Carlson and Mercer [6] needed a fracture spacing of about 80 ft, both with matrix permeability of less than 0.1 nd. An issue that does not seem to have been addressed however, is whether these fractures, if present, are in fact open and if so how they are main‐ tained open against closure stress. If, like coal cleats, they are initially water-filled, is water production observed? An implicit assumption seems to be that they are open and gas-filled.

Carlson and Mercer [6] proposed that molecular diffusion is the dominant transport mecha‐ nism in the matrix in these extremely tight reservoirs, in which case a matrix diffusion coefficient should be used instead of the matrix permeability. They did not evaluate the consequences of this hypothesis. It remains a possibility that the use of such a coefficient would

capillary pressure and slippage effects especially at low pressure.

offset by matrix shrinkage as the gas is desorbed.

330 Effective and Sustainable Hydraulic Fracturing

**1.2. Devonian shales of the appalachian basin**

downhole.

#### **1.3. Devonian Antrim Shale of the Michigan basin**

The Antrim Shale is a shallow, under-pressured, naturally-fractured shale reservoir with characteristically low matrix permeability, and with adsorbed gas, free gas and mobile water co-existing in the reservoir. A typical Antrim well will produce considerable quantities of water early in life, and as dewatering of the reservoir progresses, water production rates decline and a corresponding increase in gas production is normally observed (as a result of gas desorption with reduced reservoir pressures), similar to a CBM well. In fact, the Antrim shale is often considered to be a hybrid of productive dry gas shale and CBM plays. It has characteristics which are similar to these other unconventional reservoirs, but it is also different in many ways. The Antrim shale is more intensely fractured than the Devonian Shales of the Appalachian Basin, with fracture spacing as close as 1 to 2 ft. Kuuskraa et al [8] have noted that the "intensity and interconnection of the fractures govern the shale's natural producibility."

The typical depth of the Antrim shale is less than about 2000 feet, pore pressure is a few hundred psi and more than 70% of the gas in place is adsorbed, the remainder being stored essentially as free gas in the matrix pores. Peak gas may occur as late as 3 years into production. Production data has been history-matched using similar software to that used for CBM [8]. It was found that fracture spacing of the order of a few inches facilitated a good match with production data. It was stated that if a fracture spacing of 3-6 ft was used (which is compatible with observations from cores and logs), then production would be an order of magnitude lower than observed in existing wells. One possible resolution of this conundrum may lie, as the authors suggest, in detrital silt layers within the matrix that could provide conductive flow paths. (An alternative explanation that remained unconsidered by the authors lies in the use of the PSS approximation for matrix transport, which may be completely invalid in this context.)

This description leaves many issues unresolved, but importantly places the Antrim shale as an intermediate between CBM and the other shales, in that the natural fractures appear to be conductive and initially water-filled, but has free pore gas in addition to adsorbed gas.

#### **1.4. Deeper gas shales**

Most modern commercial gas shale plays are similar in many respects to the Barnett shale, though there are of course many differences and variations in the values of the parameters that control the gas production process. In these relatively deep and high pressure reservoirs, most of the gas is stored as pore gas, but the production process is still similar to that described above for shallower shale gas plays.

A common argument for the necessity for open natural fractures in shale gas plays is that a large surface area is necessary for economic gas production from these very tight rocks. Later in this paper we analyze production data to estimate the magnitude of the fracture surface through which gas is actually produced. We are able to demonstrate that although it may be commensurate with the expected surface area of open natural fractures for the ultra-low permeability shallow gas shales, it is in fact commensurate with a very much smaller area for the deeper gas shales such as the Barnett. Furthermore, given the typical permeability of the Barnett shale (some 100 times that of the shallow gas shales), almost all the gas between the fractures would have been produced in an uncharacteristically short period of time unless the fractures are 100s of feet apart. These issues and conclusions will be discussed at length later

The Role of Natural Fractures in Shale Gas Production

http://dx.doi.org/10.5772/56404

333

**2. Production mechanisms, production modeling techniques and**

standing of their geology and the underlying geophysics.

two methods of characterizing a multiply-fractured reservoir:

their production characteristics, such as permeability and conductivity.

plays and the deeper shale plays.)

of flow characteristics.

Having outlined the pertinent characteristics of unconventional gas reservoirs, we now document the likely production mechanisms in the various shale plays based on our under‐

The matrix permeability of shale gas reservoirs is extremely small, probably on the order of one tenth of a microdarcy or 100nd. It is virtually impossible to produce gas from these reservoirs in commercial quantities unless the wells are hydraulically fractured and even then, or so it is commonly believed, production is really only possible because a network of natural fractures is opened up. (It is interesting to note that gas has been produced from the ultra-tight Devonian shale plays of the North Eastern USA from more conventionally-fractured vertical wells, which implies that multi-stage hydraulic fracturing was unnecessary for these plays. This is the first hint that the role of the natural fractures may be quite different for the Devonian

An essential element of a mathematical model of gas production from shales is therefore the ability to describe flow in a very tight rock matrix and flow in a network of fractures. In most gas shale reservoirs most of the reservoir fluid is stored in the matrix and the primary flow path is from the matrix into the fractures and thence into the wellbore. There are essentially

**•** Discrete fracture network (DFN) model, in which the fractures are defined explicitly in terms of their location in the reservoir, their connectivity to one another and to the wellbore and

**•** Dual porosity/dual permeability models in which the fracture network is treated as a continuum in much the same way as a porous medium is treated as a continuum for analysis

in the paper.

**simulators**

The subject of open natural fractures is one of the most contentious within the community of Barnett workers. Many investigators have presumed that open natural fractures are critical to Barnett gas production, as they are for the shallower gas shales, even though there is now a growing body of evidence that any natural fractures that do exist may well be filled with calcite or other minerals (see Figure 2). There are also arguments that suggest that if there was an abundance of open natural fractures within the Barnett, there would be a much smaller gas accumulation present within the reservoir. Open natural fractures, if they existed, would have led to major expulsion and migration of gas out of the shale into overlying rocks, substantially decreasing pore pressure within the Barnett and, hence, the amount of gas in place. The Barnett would not be over-pressured (that is, over-pressured relative to the bounding strata) if copi‐ ous open natural fractures existed. Note that the Barnett is not just the gas reservoir, but also the source, trap, and seal for the gas; if the seal is fractured and inefficient, then the present gas in place would be reduced because the free gas would be lost, and only the adsorbed gas would remain in the shale (a similar situation to that of the Antrim Shale of northern Michigan). The hugeamountofgas inplace,inanover-pressuredandfully-saturated(intermsof sorption)state, is ultimately what makes the Barnett so prolific.

**Figure 2.** Mineralized natural fractures in a Barnett shale sample (adapted from [9]).

A common argument for the necessity for open natural fractures in shale gas plays is that a large surface area is necessary for economic gas production from these very tight rocks. Later in this paper we analyze production data to estimate the magnitude of the fracture surface through which gas is actually produced. We are able to demonstrate that although it may be commensurate with the expected surface area of open natural fractures for the ultra-low permeability shallow gas shales, it is in fact commensurate with a very much smaller area for the deeper gas shales such as the Barnett. Furthermore, given the typical permeability of the Barnett shale (some 100 times that of the shallow gas shales), almost all the gas between the fractures would have been produced in an uncharacteristically short period of time unless the fractures are 100s of feet apart. These issues and conclusions will be discussed at length later in the paper.

**1.4. Deeper gas shales**

332 Effective and Sustainable Hydraulic Fracturing

above for shallower shale gas plays.

is ultimately what makes the Barnett so prolific.

**Figure 2.** Mineralized natural fractures in a Barnett shale sample (adapted from [9]).

Most modern commercial gas shale plays are similar in many respects to the Barnett shale, though there are of course many differences and variations in the values of the parameters that control the gas production process. In these relatively deep and high pressure reservoirs, most of the gas is stored as pore gas, but the production process is still similar to that described

The subject of open natural fractures is one of the most contentious within the community of Barnett workers. Many investigators have presumed that open natural fractures are critical to Barnett gas production, as they are for the shallower gas shales, even though there is now a growing body of evidence that any natural fractures that do exist may well be filled with calcite or other minerals (see Figure 2). There are also arguments that suggest that if there was an abundance of open natural fractures within the Barnett, there would be a much smaller gas accumulation present within the reservoir. Open natural fractures, if they existed, would have led to major expulsion and migration of gas out of the shale into overlying rocks, substantially decreasing pore pressure within the Barnett and, hence, the amount of gas in place. The Barnett would not be over-pressured (that is, over-pressured relative to the bounding strata) if copi‐ ous open natural fractures existed. Note that the Barnett is not just the gas reservoir, but also the source, trap, and seal for the gas; if the seal is fractured and inefficient, then the present gas in place would be reduced because the free gas would be lost, and only the adsorbed gas would remain in the shale (a similar situation to that of the Antrim Shale of northern Michigan). The hugeamountofgas inplace,inanover-pressuredandfully-saturated(intermsof sorption)state,
