Frac Facts

HOUSTON–A major challenge to efficiently completing a horizontal well in shale plays is determining the optimal stage spacing and number of perforation clusters per stage to effectively drain the reservoir.

Because the Eagle Ford Shale play in South Texas is still in its development infancy compared with other North American shale plays, it is lacking scientific studies that address these key lateral coverage questions. In developing the Eagle Ford, industry practices have been based largely on available capital or best estimations. Even though the Eagle Ford has different formation characteristics and mineralogy content, the tendency on initial wells was to apply the same drilling and completion designs used in more established shale plays.

El Paso Corporation undertook a project to study stage spacing and numbers of perforation clusters per stage in four of its first Eagle Ford horizontal wells in an effort to optimize completion design and improve reservoir performance. The wells were completed with typical 4,000-foot lateral lengths with four to six perforation clusters and 300,000-350,000 pounds of proppant per stage, with a total of 14-16 stages.

Radioactive tracers (antimony, scandium and iridium) were used to tag the proppant, chemical tracers marked the stage frac fluid, and flow-through composite bridge plugs isolated the treated zones. After stimulating all the zones, the plugs were drilled out and a spectral gamma ray tool was run in the hole with coiled tubing to log the lateral.

Tracer log results from the first well in the study showed that most of the perforated zones were treated, with the exception of the toe stage, which showed little or no stimulation. A close observation of the tracer log output indicated that the treated lateral area (the area with tracer signature) was less than half of the entire lateral length. A linear count of the exact lateral length covered by the tracer mark compared with the entire lateral length revealed less than 36 percent coverage.

After drilling the first well, changes were made to the completion methodology to improve the stimulated lateral coverage and reservoir performance (the well productivity index). These changes resulted in obtaining 100 percent improvement in stimulated lateral coverage, with a corresponding positive impact on productivity index.

Understanding The Reservoir

FIGURE 1

Eagle Ford Dry Gas,
Condensate and Oil Windows

The hydrocarbons produced from the Eagle Ford formation range from predominant dry gas in the south (Duval County), to condensate-rich liquid (La Salle County), to predominant oil (Zavala and Frio counties). Figure 1 shows the general areas where the reservoir produces oil (green), high liquids (dark green) and predominately dry gas (brown). Moreover, some parts of the play are overpressured, and the basic structural characteristics of the shale vary significantly across the play. For example, gross height ranges from 20 to 500 feet in thickness, and depth varies from 2,500 to 14,000 feet. The variability makes understanding the local reservoir critical to successfully developing and stimulating the Eagle Ford.

The first exploratory Eagle Ford wells were drilled in the dry gas window in LaSalle County in late 2008. The early completion designs emulated Barnett-style water frac stimulation treatments and had varying degrees of success. Determining what spacing to apply and how many perforation clusters to use per stage are issues that many operators have struggled with. With Eagle Ford wells costing $8 million-$10 million, it is important to achieve efficient lateral coverage without spacing fractures too close together. Hydraulic fracturing is the largest single completion cost component, and spacing determines how many fracs will be placed in a given lateral length (the denser the spacing, the higher the cost).

The effectiveness of the fracture network created depends on multiple factors. Given the same fracture design (job size, rate, fluid type, etc.), the major variables are spacing between composite bridge plugs, the number of perforation clusters per stage, and the space between the clusters. Considering the ultralow permeabilities of shale formations, the amount of hydrocarbons drained from the reservoir are proportional to the stimulated rock volume (SRV), making placing multiple, closely-spaced fractures critical to establishing commercial production.

On the other hand, spacing multiple fracs too closely can be counterproductive, since the stress field generated between competing fractures can impact the propagation of adjacent fractures within the same stage, resulting in less-than-optimal fracture geometry and SRV, reducing production and recovery.

In Situ Stress Field

The in situ stress field has a major influence on hydraulic fracture creation, impacting fracture initiation, propagation, orientation and geometry. Creating a hydraulic fracture also alters the stress field in the surrounding area. The stress field induced from multiple fractures will influence the propagation of neighboring fractures within the same stage and adjacent stages to varying degrees.

Stress concentration occurs between fractures. The created fracture alters both the in situ minimum horizontal and the maximum horizontal stress, and studies show the stress increase is greater for minimum horizontal stress than for maximum horizontal stress. The width of the center and subcenter fractures are influenced strongly by fracture spacing (perforation cluster spacing), while the width of the edge fractures are relatively insensitive to fracture spacing.

For multistage fracturing treatments, the stress concentration resulting from the previous fractured stage may have a significant effect on the initiation and propagation of fractures in the subsequent stage when cluster spacing is too close. However, this effect may be negligible by the time the next stage is ready for fracturing, since the stress concentration created during fracturing dissipates with time (until fracture closure). In ultralow-permeability shale formations, it is not uncommon for stress dissipation to take longer between stages.

The need for adequate lateral coverage in horizontal shale completions has been well documented. In some instances, additional sets of perforations have been added between existing perforations a few years after the original completion and then refractured, with well production exceeding the rates after the initial completion.

An example is Bakken Shale refracs. In one case, the original completion design used fewer perforated clusters in an attempt to divert treatment along the lateral, but post-frac tracer logs revealed long intervals of unpropped sections near the well bore. To increase contact with the reservoir, additional perforation clusters were added between existing clusters and the well was fracture treated again (using a similar job size as the original frac treatment). A post-frac tracer log revealed more lateral coverage with higher production after refracturing than what had been achieved with the initial frac. This highlights the importance of adequate lateral coverage in effectively draining a shale reservoir.

Increasing perforation clusters in one stage can lead to the resulting fractures competing for space (width), while having too few clusters in one stage results in inadequate drainage of the reservoir. This necessitates finding an optimal spacing per 1,000 feet of lateral and number of perforation clusters per stage to adequately drain the reservoir and ensure the resulting fractures have enough width.

Eagle Ford Case Studies

To determine the optimal spacing and number of perforation clusters in the Eagle Ford, El Paso’s study analyzed four of its horizontal wells: the first two it drilled in the dry gas window and the first two it drilled in the oil window. The Briscoe Nunley GU No. 1H was the company’s first Eagle Ford completion. It is located in the dry gas window, where the formation produces mainly gas without condensate.

FIGURE 2A

Open-Hole Data In Vertical Pilot Hole

To adequately understand the formation and reservoir characteristics, core sample analysis with X-ray diffraction was conducted. A cased-hole pulsed neutron log (PNL) was run on electronic coil consisting of a natural gamma-spectrum log for shale and clay type identification, and an inelastic and capture spectrum log with which a mineralogy model was built by determining the volume fraction of clay, sand and carbonate.

The PNL log was converted using the cased-hole interpretation (CHI) model to pseudo open-hole data (density, neutron and resistivity), which simplified future well-to-well comparisons. The resulting synthetic data, together with the PNL data, were analyzed in the petrophysical shale log model to determine the brittleness factor of the shale along the lateral. Formation brittleness and ductility influence fracture initiation and propagation (hence stimulated rock volume and coverage in the lateral section of the formation), since fractures in more brittle areas tend to initiate and propagate easily compared with those in ductile sections.

FIGURE 2B

CHI Model Pseudo
Open-Hole Data in Lateral

The data retrieved from the PNL log were then calibrated with the open-hole log data from the vertical pilot hole. Figure 2B shows that the CHI-modeled pseudo open-hole data correlate well with that of the vertical pilot shown in Figure 2A. The shale log model was used to calculate the rock brittleness factor to grade the shale reservoir into potential completion intervals. Figure 3, displays the CHI model of PNL shale log analysis in the lateral well bore. The more ductile rock is indicated by a green color spectrum in the brittleness factor track, while the more brittle rock is colored in red.

FIGURE 3

CHI Model of PNL Shale
Log Analysis in Lateral

 

Completion Design

The lateral length of the Briscoe Nunley well is 4,000 feet (16,643 feet measured depth and 12,513 feet total vertical depth). The well was completed with 250-foot spacing between plugs, four perforation clusters (three feet with six shots per foot), and 50 feet between perforation clusters. The lateral section was divided evenly up to a total of 16 stages.

The vertical pilot hole was drilled, logged and evaluated to determine where to drill the lateral. The well bore configuration used 4.5-inch, 15.1-pound/foot casing. The job was designed with a series of three 100-mesh sweeps using 10-pound linear gel followed by linear pad and main proppant stages ranging from 0.25 to 2.0 pounds. The plug-and-perf completion method was used to isolate the stages.

Each stage consisted of 350,000 pounds of total proppant (50,000 pounds of 100-mesh and 300,000 pounds of 40/70-mesh resin-coated sand) staged from 0.5 to 3.0 pounds/gallon with 12,800 barrels of a “hybrid” system with 3 percent potassium chloride (KCl) treated water. The average surface treating pressure was 12,000 psi at 80 bbl/minute.

The pump schedule was designed to accommodate the complexity of the frac system in the near-well-bore region. The hybrid design schedule included 30,000 gallons of prepad, which created fracture geometry ahead of three sets of 100-mesh sweeps pumped with 5.0-centipoise linear gel fluid, a 15,000-gallon main frac linear pad, and the 40/70 resin-coated proppant. Proppant was staged from 0.5 to 2.0 pounds/gallon, and at the 1.25 pounds/gallon stage, the fluid was cross-linked with a borate cross-linker and cross-linked gel was used to transport the heavier proppant-laden fluids and create frac width for the tail end of the treatment.

Frac design was finalized on location, the stimulation treatment was performed, and both chemical and radioactive tracers were run as part of the post-frac diagnostics to evaluate the treatment. During each stage, a different chemical tracer was added to the proppant and frac fluid. Proppant was traced with alternating isotopes, and fluid was traced with a fluid tracer. Water samples were collected during flow back, and production logs were run in the lateral about six weeks after the well was completed.

FIGURE 4

Radioactive Tracer Log
(Briscoe Nunley GU No. 1)

The radioactive tracer log (Figure 4) shows that all 16 fracs had communication with the formation. It also appears that the majority, if not all, of the perforation clusters had communication between the well bore and the reservoir. However, the log also reveals a large interval of unpropped section near the well bore. The section of the lateral without tracer signature (unpropped) was counted, and only 32 percent of the lateral section was propped based on the tracer signature. Also, the tracer signatures near the well bore were not distinct, which may be a result of low proppant concentration in this area.

Nunley Traylor No. 1H

The Nunley Traylor No. 1H was El Paso’s second well drilled in its Eagle Ford acreage. It is located about one mile from the Briscoe Nunley No. 1, again in an area where the Eagle Ford produces mainly gas without condensate. Both wells share identical petrophysical properties.

The lateral length of the Nunley Traylor well is 4,000 feet (16,838 feet MD and 12,500 feet TVD). The well was completed with 279 feet of spacing between plugs, six perforation clusters (three feet with six shots per foot), and 37 feet between clusters. The lateral section was divided evenly into 14 stages. The job was designed with a series of three 100-mesh sweeps with 10-pound linear gel, followed by a linear pad and main proppant stages ranging from 0.25 to 4.0 pounds. The well was designed with 14 stages, with the plug-and-perf method used to isolate stages.

Each stage again consisted of 50,000 pounds of 100-mesh sand and 300,000 pounds of 40/70-mesh resin-coated sand staged between 0.5 and 4.0 pounds/gallon and 12,800 pounds of a hybrid system with 3 percent KCl treated water. Proppants and frac fluids were traced again. Average surface treating pressure was 10,200 psi at 80 barrels/minute.

FIGURE 5

Radioactive Tracer Log
(Nunley Traylor No. 1)

The radioactive tracer log (Figure 5) showed that all 14 stages had communication with the formation. It also appears that the majority, if not all, of the perforation clusters had communication between the well bore and the reservoir. Furthermore, the log showed more coverage in the lateral than was seen in the Briscoe Nunley. In fact, 69 percent of the lateral section was propped, which was more than twice the section propped in the Briscoe Nunley GU No. 1. The tracer signature also was more distinct. Overall, the coverage was much better in the Nunley Traylor well.

A production index comparison of both wells after 30 days of production shows the Briscoe Nunley and Nunley Traylor wells had the same PI (0.13 cfd/psi2), despite the fact the Briscoe Nun­ley had two more stages and pumped 700,000 pounds more proppant. The cost of the two extra stages in the Briscoe Nunley GU No. 1 was $487,450.

Lateral communication with the well bore increased by more than 100 percent by increasing the perforation clusters from four to six. A more defined tracer indication also was noticed on the log by increasing proppant concentration to four pounds of proppant added (ppa) on most of the stages and to five ppa in the last stage, even though the well was overflushed. The increase in proppant concentration at the tail end of the job was still very effective in establishing a larger open pathway (propped width) to the formation.

A zero-time plot of the production profile between the Briscoe Nunley GU No. 1 (4,000-foot lateral with 16 stages spaced 250 feet apart and 50 feet between clusters using four clusters and a maximum 2 ppa proppant concentration) and the Nunley Traylor No. 1 (4,000-foot lateral with 14 stages spaced 279 feet apart and 27 feet between clusters using six clusters and a maximum proppant concentration of 4 ppa) shows little to no difference, despite the fact the Nunley Traylor well was flowing at a more restricted rate. The choke size for the Nunley Traylor was 24⁄64, while the Briscoe Nunley was flowed on a 38⁄64 choke.

From day 12, when the Nunley Traylor well came on line, the production was higher than the Briscoe well. This trend continued for another 40 days, at which point the choke was opened up to 46⁄64 on the Briscoe Nunley well, causing its production to jump above the Nunley Traylor. The Briscoe Nunley remained slightly higher for another 60 days until day 120, but from there on, production from both wells was exactly the same.

Hixon No. 1H

The Hixon No. 1H was El Paso’s first well completed in the Eagle Ford’s liquid-rich condensate window. The net pay thickness in the Hixon area is 180-220 feet. The lateral length is 4,000 feet (14,161 feet MD and 9,950 feet TVD). The well was completed with 285 feet of spacing between plugs with four perforation clusters (three feet with six shots per foot) and 65 feet between clusters. The lateral section was divided into 14 stages.

The job consisted of only one sweep stage with 10 pounds of linear gel, followed by a linear pad and the main proppant stages ranging from 0.25 to 4.0 ppa. The plug-and-perf method was used for isolation. Each stage consisted of 300,000 pounds of proppant (80,000 pounds of 100-mesh sand and 220,000 pounds of 40/70 resin-coated sand) staged from 0.5 to 4.0 ppa with 7,500 barrels of a hybrid system with 2 percent KCl substitute treated water. The proppant and fluid were traced with alternating isotopes and a fluid tracer. The average surface treating pressure was 9,400 psi at 80 bbl/min.

Knowing that the Hixon well was in a liquid-rich condensate area and that multiphase flow would require greater conductivity to adequately drain the reservoir, the pump schedule was designed with a more aggressive proppant ramp to help generate conductivity.

FIGURE 6

Radioactive Tracer Log (Hixon No. 1)

The post-frac log indicated that all 14 stages had communication with the formation. The majority, if not all, of the perforation clusters had communication between the well bore and the formation. However, a count of the interval of propped section near the well bore, as indicated by the tracer signature, revealed that only 37 percent of the lateral section was propped. It also should be noted that the tracer signature was quite distinct compared with the Briscoe Nunley (Figure 6).

Hixon No. 4H

The Hixon No. 4H was El Paso’s second well completed in the liquid-rich condensate window and is located 1.5 miles from the Hixon No. 1H. The lateral length is 4,950 feet (15,033 feet MD and 9,950 feet TVD). The well was completed with 275 feet of spacing between plugs, six perforation clusters (three feet with six shots per foot) and 37 feet between perforation clusters. The lateral section was divided evenly into 18 stages.

The per-stage job design was the same as the Hixon No. 1, except that the maximum proppant concentration reached 5 ppa for most of the stages and larger 30/50-mesh proppant was pumped on stages 1, 2, 7 and 13, where a report from Horizontal Solutions International, a geonavigation/geosteering consulting and reporting services provider, showed faults going through these sections of the lateral. All 18 stages were pumped to completion.

FIGURE 7

Radioactive Tracer Log (Hixon No. 4H)

After logging the well post-frac, the tracer log (Figure 7) showed that all 18 stages were in communication with the formation. The tracer also showed more coverage in the lateral than was seen in the Hixon No. 1H. The interval of propped section near the well bore, as indicated by tracer signature, showed that 67 percent of the lateral section was propped, similar to the results obtained in the Nunley Traylor well. In addition, the tracer signatures were quite distinct (it is worth noting that the propped width indicated by tracer signatures is larger in the Hixon No. 1H than in either the Hixon No. 4 or Nunley Traylor No. 1).

From the tracer logs, it is evident that the Hixon No. 1H has more spaces between the propped sections compared with the Hixon No. 4H, and consequently, less SRV near the well bore. The propped width also is pronounced compared with the propped width of the Hixon No. 4H well. Propped width in the No. 4H well is strongly influenced by spacing, since there seems to be competition for space among the individual fractures compared with the No. 1H. The spacing in the Hixon No. 1H seems to allow for creating larger individual fracture widths.

After six months of data collation, the Hixon No. 4H (4,900-foot lateral with 18 stages spaced 275 feet apart and 37 feet between clusters using six perforation clusters and 5 ppa maximum proppant concentration) had yielded 38 percent more production than the Hixon No. 1H (4,000-foot lateral with 14 stages spaced 285 feet apart and 65 feet between clusters using four perforation clusters and 4 ppa maximum proppant concentration).

FIGURE 8

Production from Bakken Shale Wells
with Different Stage Spacing

Figure 8 shows cumulative production results from multiple wells with varied spacing in the Bakken Shale play. In this case, it is evident that wells with 250-300 feet of spacing perform better than wells with greater spacing. The four wells in the Eagle Ford study indicate that wells with 270-280 feet of spacing perform better than wells with more than 280 feet of spacing between stages.

Main Conclusions

The main conclusions that can be drawn from the analysis of the four Eagle Ford Shale horizontal wells include:

• Too much space between perforation clusters (greater than 40 feet) results in less than 40 percent of propped section near the well bore.
• Reducing space between perforation clusters to within 35 feet doubles the percentage of the propped section near the well bore.
• Placing 3.5 stages per 1,000 feet of lateral seems to give the best results.
• Final proppant concentrations of less than 3 ppa do not give adequate propped width near the well bore (this is even more critical since all plug-and-perf completion methods overflush after each stage, which means the actual final proppant concentration near the well bore may be significantly less than the final pumped proppant concentration).
• Given the choice between propped width and stimulated rock volume/coverage, SRV seems to be more critical than width near the well bore (this is not surprising, considering the nano-Darcy nature of shale permeability and the fact that any area of the formation not fractured or propped does not contribute to production).
• Significant cost savings can be realized by placing the optimal number of stages necessary to adequately drain the reservoir.

AUSTIN, TX.–All geoscientists operate with a paradigm or working hypothesis on the nature of the subsurface in their project areas. Having a working model is imperative for predicting the nature and presence of oil and gas reservoirs. The concept of this stratigraphic framework would allow professionals to incorporate it into their own interpretations and use it to further apply their understanding of where to drill or how best to exploit oil and gas reserves.

This model potentially would include virtually all oil and gas basins in North America, and allow users to focus on unconventional resource plays such as the Eagle Ford Shale in South Texas, the Bakken Shale in the Williston Basin, the Niobrara in the Denver-Julesburg Basin, the Wolfberry/Wolfbone play in the Permian Basin, or the Barnett Shale in the Fort Worth Basin.

A new approach has been developed to use interpreted, digitized well log data and sequence stratigraphy to build high-resolution subsurface models. The high-density interpretations are generated from the well log data according to allostratigraphic methodology.

Allostratigraphy is a methodology defined by the North American Stratigraphic Code that uses bounding discontinuities to subdivide the sedimentary section into mappable units. Bounding discontinuities include unconformities, disconformities, discontinuities and omission surfaces. The bounding surfaces may relate back to sequence stratigraphic boundaries, such as marine flooding surfaces.

It is important in high-resolution subsurface modeling to understand the difference between allostratigraphy and lithostratigraphy. Lithostratigraphy maps sedimentary rocks solely on the basis of their lithology, and does not necessarily consider that these rocks may have accumulated over different periods (diachronous lithostratigraphic units). Lithostratigraphy often ignores significant breaks in the sedimentary section, including those caused by unconformities, omission surfaces, ravinement surfaces and flooding surfaces.

Allostratigraphy maps rock units on the basis of the timing of their accumulation. Allostratigraphy (sequence stratigraphy) uses a framework based on surfaces of erosion and nondeposition (sequence boundaries), and flooding (transgressive surfaces and/or maximum flooding surfaces) that can be recognized in 2-D and 3-D seismic, well log data and outcrops.

FIGURE 1

Lithostratigraphic (Top) versus
Allostratigraphic (Bottom) Interpretation

More Accurate Interpretation

Figure 1 shows the stratigraphic interpretation that results from a conventional lithostratigraphic approach (top) in comparison with an allostratigraphic interpretation approach (bottom). Although the subsurface rocks are the same, the allostratigraphic interpretation more truly represents the genetic significance, timing and relationships among the various depositional units.

In developing an allostratigraphic interpretation, it is necessary to define a set of allomembers to be used as stratigraphic surfaces. The types of data that can be useful in the allomember definition process may include:

• Well logs;

  FIGURE 2

  Well Log Curves Used to Define Allomembers and Their Sequence Stratigraphic Significance

  
• Core descriptions;
• Outcrop analysis;
• Paleontology;
• Sample logs/cuttings;
• Geochemistry; and
• Petrophysics.

Ideally, allomembers should be regionally correlative and have genetic significance.

Figure 2 illustrates how allomembers are defined from regionally correlative log responses and how they can be tied to a sequence stratigraphic framework. In this case, detailed core analysis was used to provide the necessary geologic facies information required to confidently assign the allomembers to the appropriate system tracts.

FIGURE 3

Traditional Well Log Display of Four Wells

High-Density Interpretation

Traditional well log displays, as shown in Figure 3, have been used for years by geologists as a means to correlate subsurface depositional units. This method is still in use today by a large majority of working geologists. Although it is now more common to see scanned well logs being manipulated on a computer screen rather than the shuffling of paper logs on a drafting table, the process between the two is essentially the same. The high-density well log display is a relatively new and patented method that facilitates visualizing stratigraphic relationships present in the well logs.

FIGURE 4

High-Density Cross Section
(Lower Wilcox, Webb County, Tx.)

The high-density well log display in Figure 4 is composed of gamma ray, spontaneous potential and deep resistivity logs. In this case, the display shows a high-density cross section of the Lower Wilcox formation from Webb County, Tx. The wells are equally spaced over a distance of 65 miles. Facies shading has been applied to the gamma ray and spontaneous potential curves to emphasize the presence of sand. Many regionally significant depositional patterns can be discerned easily in this display. It would be virtually impossible to use paper or scanned well logs to create this kind of cross-sectional display, let alone try to visualize the stratigraphic relationships.

By carefully correlating key allostratigraphic surfaces in a basin and understanding their relationships to identified sequence boundaries, the sequence stratigraphic model for a basin can be developed. Major first- and second-order sequence boundaries frequently correspond to major flooding surfaces, unconformities and named formation boundaries. The third- and fourth-order sequence boundaries are mapped where they can be identified in well logs.

FIGURE 5

Example of Framework Correlations
Matching Interpreted Seismic Data

Figure 5 details how these framework interpretations generate regionally significant correlations over a large area that can be integrated with seismic data or used independently to understand local geology in a regional context, as well as to easily visualize productive facies, their lateral extent, and potentially identify analogs.

Structure Maps

Based on the interpreted results of thousands of well logs, it is possible to create highly accurate and very detailed structure maps of the subsurface. Because these basinwide correlations focus on flooding surfaces and sequence boundaries, the structure maps can tie back to 3-D and 2-D seismic data and be used to better control the time/depth relationships.

Structure maps are generated for each correlated surface in a basin or trend. Many areas potentially could have 10 to 20 structure maps detailing the unique characteristics associated with each depositional layer. Because of the correlation methods and data density, the presence of even relatively minor faults frequently can be seen by their effect on the structure maps.

However, perhaps the most remarkable and valuable derivation from these basinwide correlations are the gross isopach maps of the intervals between the interpreted surfaces. Gross isopach maps can be extremely valuable tools in defining the paleo-structure in a basin. They have the ability to identify a thicker region that may indicate a zone of subsidence, allowing for greater accumulation of sediments. This ability is largely a function of the large-scale correlations representing sequence boundaries, and therefore, is more likely to capture the significant depositional packages.

Taking the vertical difference between two adjacent surfaces in each interpreted well creates the gross interval isopach maps. Gross isopach maps then are created for each interval of the interpretation. Additionally, isopachs are created by multiple intervals. These combined isopachs are necessary to visualize larger deposition systems.

Because this stratigraphic framework is dynamic and is derived from interpreting an ongoing and ever-increasing collection of digitized well logs, it becomes immediately possible to calculate and map petrophysical properties for each interpreted interval.

Common well log derived property maps such as feet of gamma ray above or below a certain cutoff point provides insight into the distribution of a sand, limestone or organic shale in a basin. The value of this information is even greater when interpreters can flip through a series of layers and observe how the depositional thickness may migrate over time.

More advanced petrophysical properties, or even geochemical measurements, can be associated with an interpreted framework interval. The trend of these properties can be mapped and followed through each succeeding depositional sequence.

Integrating Production Data

The ability to match interval data with well completion and production information is an important aspect of any interpretation. Ideally, all well completions would be matched to known production intervals and reservoir boundaries that are consistent throughout the entire basin or trend.

With the ability to match completions to a consistent set of intervals, it becomes possible to relate the interpreted framework-derived property maps to known production trends. This functionality, in turn, is extremely useful in generating “opportunity maps” for each potentially productive interval.

FIGURE 6

Framework Interpretation Integrated with
Published Production Data from the Haynesville/Bossier Zones

Figure 6 provides an example of the types of visualizations that can be created using data and information from the framework interpretation. This figure combines data derived from integrating framework interpretation and published production data from the Haynesville/Bossier zones in East Texas and northern Louisiana. Peak production is mapped and defined by the area in red, which is superimposed over a series of gross isopach maps of the Lower and Upper Haynesville, and Lower Bossier zones.

A comprehensive geological model for each of the primary oil and gas basins and trends can be created by applying a patented process that utilizes high-density digital well log displays. This framework interpretation potentially can generate a large and expanding array of reservoir characterization tools, including structure maps, isopach maps, various petrophysical property maps and high-density well log displays.

Integrated with available production and completion data, the concept of a nationwide stratigraphic interpreted framework would provide an invaluable tool for identifying and validating opportunity trends, and would lead to a better understanding of the relationship between geology and known production.

NATRONA HEIGHTS, PA.–Multiple technologies are available to assist operators and engineers in optimizing stimulation treatments in horizontal shale wells to enhance production rates, increase ultimate recoveries and maximize economics.

This article describes an integrated approach to enhancing production in the Marcellus Shale while optimizing economics using advanced analysis and modeling techniques to optimize multiple transverse vertical hydraulic fractures intercepting horizontal well bores.

The procedure integrates technologies and techniques such as hydraulic fracture analyses, numerical simulations, minifrac analysis with diagnostic fracture injection tests (efficiency, instantaneous shut-in pressure, closure pressure, net pressure, etc.), after-closure analysis (permeability and reservoir pressure), replay pressure history matching, complex fracture geometry and aerial extent, stimulated reservoir volume (SRV) modeling, and comparisons of numerical fracture propagation simulations with microseismic imaging results.

The advantages of technology integration for multiple transverse fractures in horizontal wells include utilizing fundamental engineering principles and production optimization, and providing a foundation for technical analyses to predict the production behavior and associated economics of multistage fracturing scenarios.

Integrating minifrac analysis, hydraulic fracturing, microseismic, production and economic evaluation technologies provides a methodology to perform more reliable engineering analyses and economic optimization of horizontal wells. The end result is a hydraulic fracture model categorized by reservoir quality, production potential and economics, giving operators a systematic approach to designing, analyzing and optimizing multistage and multicluster transverse hydraulic fractures in the Marcellus Shale.

Integrated Work Flow

Among the key elements of the technology integration workflow for Marcellus wells are:

• Petrophysical analysis and log evaluation;
• Minifrac analysis;
• Microseismic analysis;
• Discrete fracture network modeling;
• Production history matching;
• Production simulation/prediction; and
• Production and economic optimization.

Petrophysical analysis and log evaluation is used as a base line to determine rock and reservoir characteristics to estimate pay and permeability as well as rock properties. Pay is necessary for after-closure analysis (ACA) and production simulation. Rock and fluid properties are used in the fracture model to define in situ stress, leak-off profiles and fracture extent.

Minifrac analysis helps define stress, net pressure and fluid efficiency. These variables are necessary to calibrate the discrete fracture network model. If pseudo-radial flow has been achieved, the reservoir pressure and reservoir flow capacity can be determined. Resulting estimates of permeability should be compared to log analysis.

Microseismic data collected during treatment are very useful diagnostic tool for determining fracture geometry, including calibrating a discrete fracture network (DFN) model by inferring the areal extent of each stage, fracture height and half-length, and plane orientation. Integrating minifrac analysis and microseismic with the production response for multiple transverse vertical fractures provides a methodology to improve stimulation programs and enhance gas production.

DFN modeling is used to define the stimulated area. Log analysis, minifrac data and microseismic are tools that increase confidence in the geometry solution and resulting stimulated reservoir volume. Production history matching defines fracture length, conductivity, skin and permeability. Fracture characteristics resulting from the history match are used as a calibration tool to define “cutoff” values in the DFN simulator. Permeability results should be compared with petrophysics and ACA analyses.

With a calibrated model, parametric studies focusing on fluid types and volumes, proppant type and mass, and fracture geometry can be performed and used in the production simulator to forecast production responses to changes in stimulation design. In addition, economics can be run on the fracture design and flow streams to determine optimal net present value and discounted return on investment.

Marcellus Case Study

A Marcellus Shale horizontal well (Well A) with a 2,100-foot lateral was completed in seven fracture stages with five perforation clusters per stage. Each stage was isolated with a composite bridge plug. The treatment design consisted of 450,000 gallons of slick water (water, friction reducer and surfactant) and 300,000 pounds of 100-mesh sand and 100,000 pounds of 40/70-mesh resin-coated sand per stage. The treatment design rate was 100 barrels a minute, although higher-than-expected surface treating pressures limited the average rate to 85 bbl/minute.

Following the seven-stage treatment, the well was flowed back and the plugs were drilled out with coiled tubing. A production log was run to determine early-time flow contributions from each stage. The results of the production log showed a total gas flow rate of 2,941 Mcf/d and a total water flow-back rate of 2,541 bbl/d. A minifrac analysis was performed on a nearby Marcellus vertical well (Well B) to determine the closure stress gradient and a Marcellus horizontal well (Well C) to provide estimates of reservoir pressure and flow capacity.

Petrophysical and geomechanical analyses were conducted on Well A to determine reservoir characteristics and mechanical rock properties. Values for stress, Young’s modulus, Poisson’s ratio and fracture toughness were defined. Table 1 shows average values for the geological formation intervals.

TABLE 1

Figure 1 shows the petrophysical analysis, which characterizes the effective porosity, permeability, clay and quartz content, and mechanical rock properties (calculating the stress gradient, Young’s modulus and Poisson’s ratio). The petrophysics show an average permeability of 377 nanoDarcy (nD) over the gross formation pay interval of 162 feet, including both the upper and lower Marcellus.

FIGURE 1

Marcellus Well A Reservoir Characteristics

A method to define the pay interval, or net effective pay (NEP), is not a straightforward process in an unconventional shale reservoir. Traditionally, NEP determination is based on some predetermined cutoff for water saturation, clay volume, permeability, resistivity, porosity, etc. As a result, the pay zone height used in the ACA and production modeling was 162 feet with a reservoir flow capacity of 0.061 milliDarcy-foot.

Offset Minifrac Analysis

A minifrac analysis was performed on both Marcellus offset wells to provide estimates of closure stress gradient, reservoir pressure and reservoir flow capacity for Well A. Minifrac analysis in unconventional resources is challenging, principally because of the extensive amount of time necessary to see fracture closure. A minifrac in the vertical B well was used to determine instantaneous shut-in pressure (ISIP), closure stress gradient, time to closure, net pressure, fluid efficiency, and nonideal leak-off characteristics. The minifrac was designed for 3,500-gallon slick-water injection and the pressure was monitored for approximately one hour.

The minifrac results indicated a bottom-hole closure pressure of 7,006 psi, with a closure stress gradient of 0.88 psi/foot, time to closure of 15 minutes, net pressure of 434 psi, and fluid efficiency of 75.5 percent. The ISIP for Well B was 3,980 psi, resulting in a fracture gradient calculation of 0.93 psi/foot, which is consistent among vertical Marcellus wells.

Unconventional reservoirs often are described with “fracture gradients” of more than 1.0 psi/foot on horizontal well stimulations based on the pressure at the end of pumping. Because of this, the propagation of horizontal fractures enters the discussion as the dominant fracture plane. While we believe there may be horizontal components in the near-well-bore region, microseismic data indicate that far-field fractures are vertical. We submit that fracture gradients greater than 1.0 psi/foot are a result of shale storage and extended well bore effects, and not a physical characteristic of the fracture extension pressure. This is not to say that the fractures in the network do not have a horizontal component, but it is not a dominant characteristic.

Well C had approximately 470 gallons of water pumped into the toe stage of the lateral. The pressure was monitored with a surface gauge for 375 hours (16 days). The minifrac was designed with a shut-in for this length of time to reach pseudo-radial flow in order to determine reservoir pressure and reservoir flow capacity. The ISIP for this well was estimated at 3,825 psi.

The purpose of ACA analysis is to determine reservoir pressure and reservoir flow capacity. ACA is only valid when pseudo-radial flow has been achieved during the pressure decline period. The reservoir pressure indicated from ACA for Well C was 5,031 psi with a pore pressure gradient of 0.625 psi/foot. The reservoir flow capacity was 0.0945 mD-foot with an estimated system permeability of 583 nD, based on 162 feet of formation thickness.

DFN Numerical Simulation

Although many conventional fracture treatments result in biwing fractures, some naturally fractured formations provide the geomechanical conditions that enable the initiation and propagation of hydraulically induced discrete fractures in multiple planes, as indicated by microseismic mapping. In these cases, planar hydraulic fracture models may not be the proper tool to predict propagation. Well A was designed and a post-job fracture treatment pressure match was performed using a commercial pseudo 3-D discrete fracture network simulator. It is designed to model multiple, cluster/complex/swarm, and discrete fractures in shale and coalbed methane formations, and to predict fracture propagation and extent in fractured and naturally fractured reservoirs. The areal extent also was based partly on microseismic mapping.

The multidimensional DFN solution is based on a network grid system and has options for continuum theory and discontinuous (grid). The fracture characteristics, apertures and propagation in the x, y and z directions (the x-z, y-z and x-y planes) are calculated numerically. The interrelationship of fracture geometry and areal extent results were evaluated based on flow regimes, wall roughness and proppant settling options.

Mechanical rock property input values for the DFN simulator were generated from the geomechanical analysis. The input values include estimates for closure stress, Young’s modulus, Poisson’s ratio, and fracture toughness to define the rock properties. The average log-derived closure stress gradient in the lower and upper Marcellus is 0.86 psi/foot and 0.87 psi/foot, respectively.

Although there are no definitive or unique solutions, very different DFN simulation results were generated based on two alternative assumptions:

• Each of the five clusters creates a separate network that interferes with the other clusters (multicluster); and
• All five clusters interact, but propagate within the same secondary fracture network (that is, one large network of connected clusters created independent of which cluster takes most of the fluid).

These numerical simulations account for extended well bore storage and pressure loss caused by fracture initiation from a horizontal well bore. The fracture initiates horizontally through the extended near-well-bore region, and then twists and turns before finally reorienting itself vertically in the principal planes. This extended well bore phenomenon creates large fracture net pressures and ISIPs with a stress gradient of 1.0 psi/foot or greater. The resulting extended well bore excess pressure for these simulations was about 3,350 psi.

The stress gradients range from a maximum of 1.0 psi/foot to a minimum of 0.85 psi/foot. Therefore, the resulting maximum and minimum stress difference at true vertical depth of 7,800 feet is about 1,400 psi. It is not uncommon to have an ISIP gradient of 1.3 psi/foot or higher, which indicates extended well bore storage and pressure losses of near 3,500 psi.

FIGURE 2A

DFN Pressure History Match (Multicluster Simulation)

FIGURE 2B

Stress, Width and Length Profiles for Dominant Fractures (Five Clusters)

FIGURE 3

3-D Major Axis Partial View of Multicluster DFN Simulation (x-z Plane)

The discrete fracture network spacing was assumed to be approximately proportional to the zone thickness. Although the zones vary, a nominal value of 75 feet was used for both the minor and major axes. Horizontal fractures were not modeled discretely in the simulations, but rather accounted for in the extended well bore pressure loss/storage function.

There are numerous options for proppant settling and proppant distribution in the DFN. The numerical simulations were run with a modest settling velocity (0.25 feet/minute) and a uniform proppant distribution in the fracture network. If all proppant remained in the dominant fracture, it would pack off. There are, of course, abundant proppant placement possibilities and scenarios, but propped lengths in the dominant or primary fracture ranged from 100 to 600 feet, depending on the settling option (empirical, cluster settling, etc.).

Simulation Results

The multicluster simulation assumes that each cluster creates its individual DFN that interacts with the other clusters, but that each cluster also creates its own secondary network that may or may not coalesce with other secondary fracture cluster systems. This assumption is probably a very good limiting case, provided the clusters do not overlap greatly (they will, however, interact for stiffness and fluid loss within a cluster and with adjacent clusters).

The discrete fracture network is based on discrete fractures spaced 75 feet apart with a DFN aspect ratio (width-to-length) of 0.25 (little DFN overlap). Figure 2A illustrates the surface and bottom-hole treatment pressure match. The dominant fracture characteristics are shown in Figure 2B.

The fracture net pressure results showed that the inner fractures initiated, but started to close because of the stiffness interaction of the outer fractures. As the overall bottom-hole pressure increases, the discrete fractures once again begin to propagate. Figure 3 shows a 3-D major axis view (x-z plane) for the DFN.

The connected-cluster simulation assumes that the multiple clusters create a single DFN that interacts with the other clusters. That is, fractures from the separate clusters may coalesce with other secondary fractures created from the individual clusters. This assumption may be the best limiting case in that secondary fractures created by individual clusters can occupy the same secondary network system (rather than create separate interacting fractures in the principle planes).

The connected-cluster DFN is based on discrete fractures spaced 75 feet apart and a DFN aspect width-to-length ratio of about 0.5 for the entire DFN set of secondary fractures created by all clusters. Figure 4A illustrates the surface and bottom-hole treatment pressure match. The dominant fracture characteristics (width contours, height, length, etc.) are shown in Figure 4B. Figure 5 show 3-D views of the generated DFN.

FIGURE 4A

DFN Pressure History Match (Connected Clusters)

FIGURE 4B

Stress, Width and Length Profiles for Stage 2 Dominant DFN Fractures (Connected Clusters)

FIGURE 5

3-D Major Axis Partial View of Connected-Cluster DFN Simulation (x-z Plane)

The DFN simulations show that the multicluster DFN has a length (major axis) of 1,740 feet, a width (minor axis) of 600 feet, and a fracture height of 240 feet. The connected-cluster DFN has a length of 2,200 feet, a width of 1,100 feet, and a fracture height of 340 feet. The pressure history match is essentially identical for the multicluster and connected-cluster numerical solutions, illustrating the nonuniqueness of solutions.

The connected DFN results in greater height growth because of secondary fracture connectivity, more closely matching the results from microseismic mapping analysis, which showed a length of 1,800 feet, a width of 1,200 feet, and 400 feet of fracture height. As a first order, the DFN characteristics are comparable to the recorded microseismic events. Clearly, the DFN has a wide fairway, rather than biwing behavior.

Microseismic Mapping

The objective in deploying microseismic monitoring in Well A was to measure height, length and azimuth of the seven frac stages versus time. Figure 6 shows side and top views of microseismic events for all seven stages. The events represent double-couple events (slippage in the rock) located based on compressional and shear waves through a velocity model calibrated to the perforation shots.

FIGURE 6

Side and Top Views of Microseismic Events (Seven Stages)

Stage 1 exhibited growth parallel to the well bore on the near side, extending to a second trend that is also parallel. This trend grows past the area targeted for the stage 2 fracture treatment and also shows height growth (probably attributable to the fracture system in the rock). There is a lack of activity on the far side of the well bore, which may be a limitation of viewing distance, although events from other stages show up on the far side.

Stage 2 shows an extension of the stage 1 trends parallel to the well bore on the near side and some growth on the far side, plus some events far from the well bore on the near side toward the heel, which activates in later stages. This parallel growth extends past the stage 3 perforation locations. The same associated height growth is seen in this stage and ties with the vertical height growth predicted in the DFN connected-cluster simulation model.

Stage 3 shows more typical growth perpendicular to the well bore, but with connection to the previous trend, including upward growth, events far from the well bore toward the heel, and extension of the fracture into the stage 4 area.

Stage 4 growth can be seen initially within the network created by previous stages. As the stage progresses, it extends the network away from the well bore in the previous stage areas, and then grows on the far side of the well bore at the stage 4 areas as well as the stage 1 area, where no growth was originally seen.

This implies that stage 1 was not limited by viewing distance on the far side of the well, but that the treatment preferentially grew in the natural fracture network on the near side. Events also occur perpendicular to the well bore in the stage 5 area. Unlike the first three stages (which grew toward the heel), stage 4 originally grew preferentially into the existing network. Extensions of the previous stages show growth above the well bore, but new growth is seen below the well bore in this stage on the near side.

Stage 5 shows extension of the growth of the previous stages, plus growth on the far side of the well bore, including some downward and upward growth on the far side of the well bore close to the stage 6 area.

Stage 6 shows a completely different character than the previous stages. It exhibits a large areal extent for the network on both sides of the well bore with only a small interaction with the stage 1, 2 and 3 trends. Events are seen far from the well bore and back toward the heel, but primarily away from the well bore. The far events in stage 2 on the near side of the well are part of this trend, and the activity seen on the far side of the well bore in stage 5 (where there is downward growth) also is consistent with stage 5 growth.

Stage 7 appears to be more contained by stage 6 growth and exhibits growth closer to the well bore, overlapping with the previous stage and demonstrating the same downward growth on the far side of the well bore.

From a total picture, there appears to be activity from the toe to the heel with good extension from the well bore on both sides. The far side events are limited to about 2,700 feet of growth viewing distance from the monitor well. The actual growth over time, however, shows trends that are different from the later stages, with activity occurring in trends parallel to and away from the well bore. Growth parallel to the well bore that dominates the earlier stages leads to questions of connectivity back to the well bore from the rock affected by these stages.

There is a definite difference in the growth between the earlier and later stages. Growth downward is seen below the well bore starting in stage 4 and is more pronounced in the later stages. Those downward events are the farthest from the well bore and have the largest uncertainty. However, the stage 4 events on the far side of stage 1 are 2,200 feet from the well bore and above it, while some downward events in stages 6 and 7 actually are closer.

The growth assumed to be dominated by the natural fractures that run parallel to the well bore act differently from those stages where perpendicular or very wide network growth is seen, and it is expected they will have a different production profile as a result of connectivity and proppant placement.

Production History Match

A new analytical solution for predicting the flow behavior of multiple transverse finite-conductivity vertical fractures intercepting horizontal well bores was utilized for the production history match on Well A. It provides a simple way to predict the production behavior accounting for production interference and associated economics of multistage/multicluster transverse fracture spacing along a horizontal well bore, and accounts for high initial production from multiple transverse fractures and matches the late-time production decline as a result of fracture interference. Figure 7A shows the production history match over 12 months, along with bottom-hole flowing pressure.

The history match is based on the tail-in of higher-conductivity proppant (40/70 resin-coated sand) compared with the initial 100-mesh proppant (results also were generated for no proppant tail-in). To better define the effective fracture length, the history match solution was run with a minimum number of created fractures (seven, or one per stage), a maximum number (35, or five per stage), and an average number (21, or three per stage).

FIGURE 7A

12-Month Gas Flow Rates versus Time
(Seven Multiple Transverse Fractures)

FIGURE 7B

100-Year Gas Flow Rates versus Time
(Seven Multiple Transverse Fractures)

While there were no unique history match solutions, a production log run on the well indicates that production was possibly from one dominant fracture in each stage. Although there is only one year of production, the predicted history match solutions seem reasonable based on the pressure match results, microseismic events and petrophysical analysis. With these solutions, the production prediction can be extrapolated. Data over a longer time will allow a more accurate predictor for production simulation.

Assuming that all fractures are contributing and the fractures near the toe of the lateral clean up, the predicted cumulative gas production from an 80-acre reservoir drainage area is 500 million cubic feet after one year, 1.5 billion cubic feet after five years, 2.4 Bcf after 10 years and 4.0 Bcf after 30 years (the multiple transverse fractures are assumed to be equally spaced throughout the reservoir, with a higher proppant conductivity tail-in).

The gas flow rate and cumulative production also were calculated for 100 years, using the history matched parameters for seven transverse fractures with a constant bottom-hole flowing pressure of 400 psi as a function of the number of transverse fractures (Figure 7B). The formation diffusivity is moderate, so it is possible for the transverse fracture spacing to be closer within the lateral without substantial interference in the near term while maximizing flow rates and cumulative production.

The spacing for an interference time of one year with a moderate diffusivity is about 300 feet (or 12 multiple transverse fractures in a 3,700-foot lateral). The cumulative recoveries after 30 years for seven, 20 and 40 equally spaced transverse fractures on an 80-acre drainage area are 4.2 Bcf, 5.4 Bcf and 5.5 Bcf, respectively.

The economic analysis to optimize the number of transverse fractures of specified fracture lengths to maximize the net present value and discounted return on investment (DROI) for Well A includes the total well cost (drilling, completion and facilities), as well as key economic inputs (gas unit revenue, inflation rate, revenue escalation rate, etc.).

The analysis shows the optimum number of transverse fractures is about 30 for NPV and 20 for DROI over 30 years. After three years, NPV is about $12 million and DROI is about $4.00 (i.e., for every $1.00 invested in drilling and fracturing a total return of $4.00 will be realized).

WASHINGTON–The United States is experiencing a dramatic shift in its energy fortunes. After years of declines, oil and gas production are on the rise, and some terminals that not long ago had been constructed to receive imports of liquefied natural gas now are being modified to handle LNG exports. These developments have been made possible by recent technological advances that have unlocked vast sources of natural gas as well as oil trapped in shale formations. The ability to economically produce these reservoirs, particularly shale gas, has been termed a “game changer” by a number of energy analysts, and offers the potential for significant impacts on America’s energy security and climate change initiatives.

Hydraulic fracturing is the key to unlocking shale gas and other unconventional oil and gas reserves. The technology has been used in the United States for the past 60 years to enhance production from oil and gas wells in a wide variety of formations. Experts have estimated that more than 90 percent of all producing U.S. oil and gas wells are hydraulically fractured. Without hydraulic fracturing, it simply would not be economically feasible to tap most unconventional reserves, and the “shale gas revolution,” in particular, would not be possible.

Because of its key role–and despite a long history of successful regulation of hydraulic fracturing operations by the states, and the safe and effective use of the technology in the United States–concerns among some members of the public about the increased production of shale gas has focused on hydraulic fracturing. Allegations have been made by environmental groups and others about drinking water well contamination and other purported impacts of hydraulic fracturing operations, but these allegations have not been scientifically confirmed.

Nevertheless, in response to these unconfirmed allegations, regulators and legislators at the federal, state and local levels as well as in a growing number of countries around the world have undertaken initiatives over the past few years to impose additional restrictions and requirements on hydraulic fracturing activities.

Legislation reintroduced in Congress (HR 1084 and S 587) would require the regulation of hydraulic fracturing under the federal Safe Drinking Water Act, resulting in additional permitting requirements and other restrictions on fracturing operations that could lead to significant administrative burdens, delays and costs for well operators. This proposed legislation, called the Fracturing Responsibility and Awareness of Chemicals Act (FRAC Act) also would require service companies to publicly disclose the chemical constituents of all fracturing fluids, a requirement that goes beyond any chemical disclosure requirement under any other applicable federal program in effect today.

EPA Activity

In the meantime, the U.S. Environmental Protection Agency has taken a number of steps to address various aspects of hydraulic fracturing. The EPA activity that has received the most attention involves a new study of the relationship between hydraulic fracturing and drinking water initiated at the direction of Congress. As part of this effort, the agency has released a proposed study plan that lays out a broad approach to its study of hydraulic fracturing and potential impacts on drinking water sources, pursuant to which EPA will study what it has described as the full “life cycle” of water associated with the hydraulic fracturing process.

For example, according to the draft study plan, EPA will review the withdrawal of water for use in hydraulic fracturing fluid, the management of frac fluids at the well site and injection of fluids into the subsurface, and the management and disposal of wastewater generated subsequent to the hydraulic fracturing process, such as flow back and produced water.

The draft study plan (http://wa­ter.epa.gov/type/groundwater/UIC/class2/hydraulicfracturing/index.cfm) has received an extensive review from EPA’s Science Advisory Board and has been subject to public comment. The initial results of EPA’s study are expected by the end of 2012, with a final report due in 2014.

At the same time, EPA has taken several other steps to increase its oversight and regulation of hydraulic fracturing. In June 2010, EPA posted statements on its website indicating the agency would be attempting, for the first time, to regulate hydraulic fracturing using diesel fuel or diesel-based additives under the Safe Drinking Water Act pursuant to the authority provided to the agency under the Energy Policy Act of 2005. Industry groups have challenged this website “rule making” based, among other things, on EPA’s failure to comply with key rule making requirements under the federal Administrative Procedure Act in imposing these new requirements.

In addition, EPA’s proposal to regulate wells being hydraulically fractured as Class II Underground Injection Control wells is inconsistent with the agency’s prior position that the regulations applicable to these wells are not suited to hydraulic fracturing operations. Notwithstanding this challenge, EPA is proceeding to develop guidance regarding how its regulations for Class II wells should be applied to hydraulic fracturing operations that use diesel fuel.

EPA also is addressing issues relating to other aspects of oil and gas well development. For example, the agency is taking steps to address discharges of flow-back and produced waters from oil and gas wells to surface waters. EPA has issued guidance concerning the regulation of discharges of flow back from Marcellus Shale wells to surface waters under the Clean Water Act, and is in the process of developing effluent limitation guidelines under the act for discharges of produced water from coalbed methane wells.

In the meantime, the Natural Gas Subcommittee of the Secretary of Energy Advisory Board is expected to issue its own report on hydraulic fracturing in the near future. This report is expected to include steps to “improve the safety and environmental performance” of hydraulic fracturing and recommendations for shale gas development practices to ensure the protection of public health and the environment.

State Initiatives

State regulators and legislators also are increasingly undertaking efforts to impose new or additional requirements on hydraulic fracturing. For instance, last year both Wyoming and Pennsylvania updated and amended their oil and gas regulatory programs to more specifically address hydraulic fracturing operations, while the Arkansas Oil & Gas Commission enacted regulations specifically requiring service companies to maintain master lists of chemicals used in hydraulic fracturing operations in the state, and to provide those lists to the commission. However, Arkansas regulators recognized that “trade secret” information warranted protection from public disclosure and provided alternative reporting options for any chemical that constituted a trade secret.

The legislative and regulatory activity at the state level has continued in 2011. Since the start of the year, legislation addressing the disclosure of chemicals used in hydraulic fracturing has been enacted in Texas and Indiana. Similar legislation was considered in states such as Illinois and West Virginia, and is currently being debated in California.

Some state legislatures, such as in Maryland and New Jersey, have gone so far as to consider outright bans on hydraulic fracturing, which would effectively preclude any oil and gas development in shales or other unconventional formations in those states.

A number of local jurisdictions, most notably in New York and Pennsylvania, likewise have moved to adopt ordinances purporting to prohibit hydraulic fracturing within their borders, despite significant questions about the authority of these municipal governments to regulate oil and gas development activities. At the same time, rule making processes addressing hydraulic fracturing are under way in Louisiana, Montana and Idaho.

As one of the key state efforts, the New York State Department of Environmental Conservation continues to consider revisions to its oil and gas regulatory program to address high-volume hydraulic fracturing operations in horizontal wells drilled in the Marcellus Shale. As part of this process, NYSDEC is assessing the potential environmental impacts of these activities by preparing a supplemental generic environmental impact statement.

After reviewing more than 13,000 comments from members of the public on its initial draft of the SGEIS, NYSDEC has issued a revised draft of the SGEIS in which it is proposing to allow hydraulic fracturing of Marcellus Shale wells on private lands, subject to certain restrictions.

Among other proposals in the revised draft, high-volume hydraulic fracturing would not be allowed in certain watersheds from which New York City and Syracuse obtain drinking water or near certain other aquifers, while operators who receive permits to drill wells would be subject to additional requirements concerning well casing and the use of tanks to manage wastewater. The Department of Environmental Conservation will provide a public comment period of 60 days on the revised draft of the SGEIS, effectively extending a de facto moratorium on Marcellus Shale gas development in the state for at least the near future.

Finally, much of the legislative and regulatory activity at both the federal and state levels has focused on the issue of transparency regarding the makeup of hydraulic fracturing fluids. In an effort to address this issue on a consistent, nationwide basis, the Interstate Oil & Gas Compact Commission and the Ground Water Protection Council have developed and launched a hydraulic fracturing chemical disclosure registry based on a website (www.fracfocus.org) that is designed to provide the public with ready access to information provided by well operators in a consistent format regarding the fluids used in fracturing individual wells. Many operators are participating voluntarily in this public disclosure and education effort, and states such as Texas are incorporating the registry in their disclosure requirements.

Potential Impact

Additional regulatory requirements at the federal, state and local levels could significantly disrupt hydraulic fracturing operations and dramatically impact domestic gas production. For example, the U.S. Department of Energy has estimated that regulating hydraulic fracturing under the Safe Drinking Water Act would add more than $100,000 to the cost of drilling and operating a natural gas well.

At the same time, proposed rules requiring the full disclosure of the formulas of proprietary products used in hydraulic fracturing would jeopardize service companies’ trade secrets. The risk posed by this type of disclosure creates substantial disincentives for service companies to engage in the necessary research and development of hydraulic fracturing fluids and to use their state-of-the-art technologies, which have proven critical to the economic development of U.S. domestic energy resources.

These potential impacts on oil and gas production will not be limited to the United States, as efforts to impose restrictions on or even ban hydraulic fracturing can be seen in Canada and in countries such as France, Australia and South Africa, based on many of the same fears and unfounded allegations that have characterized the debate in this country.

The need to impose these new requirements on hydraulic fracturing operations certainly remains subject to question. Hydraulic fracturing poses little to no threat to drinking water supplies. Contrary to popular allegations, there is no evidence that hydraulic fracturing has contaminated any drinking water supplies.

In fact, state regulators repeatedly have stated that they are not aware of any instances of contamination of drinking water aquifers as a result of hydraulic fracturing operations, a conclusion affirmed by Lisa Jackson, administrator of the EPA. A number of studies have confirmed that hydraulic fracturing poses minimal risk to drinking water aquifers.

These conclusions are logical, given the nature of hydraulic fracturing activities and the various geological and technical factors that prevent any significant migration of fracturing fluids upward from targeted production formations and into shallow drinking water wells thousands of feet above. In addition, well construction practices and zonal isolation techniques such as casing and cementing help ensure that fluids in the well bore will not come in contact with groundwater that may serve as a source of drinking water.

With the shale boom continuing to gather steam, hydraulic fracturing is expected to be the subject of ongoing scrutiny. While the prospects for federal legislative activity in the near future are slight, EPA will proceed with its study of hydraulic fracturing and may continue to explore its regulatory authority over fracturing operations. State regulators and legislators likely will continue to undertake efforts to impose increased regulatory requirements on hydraulic fracturing operations, regardless whether those additional requirements actually serve to mitigate any real risk.

These efforts could impact the extent to which the nation can economically access and produce the vast natural gas supplies that could significantly affect the nation’s energy security and lower greenhouse gas emissions generated by America’s energy use.

HOUSTON–A barrel of crude oil has the equivalent Btu heating capacity of 6 Mcf of natural gas, yet a barrel of oil was priced 25 times higher than an Mcf of gas on the New York Mercantile Exchange in late April. This widening valuation gap between oil and gas is leading many North American operators to rebalance their portfolios toward liquids reserves, and is driving new investment opportunities in unconventional plays with oil and natural gas liquids.

The strategic shift to liquids includes high NGL-yield gas shales such as parts of the Marcellus, Eagle Ford and Woodford, as well as oil shales such as the Bakken, Niobrara, Avalon/Bone Spring and the northern part of the Eagle Ford. In addition, companies are targeting multiple-zone “tight” liquids plays, including the Granite Wash in the Texas-Oklahoma panhandle region and the Wolfberry in the Permian Basin.

Although liquids plays use the same horizontal drilling and multistage hydraulic fracturing methods that opened gas shales from the Barnett to the Fayetteville and Haynesville, drilling and fracturing technologies are being both applied and evolved in a “fit for purpose” manner to target liquids reserves.

No two shales are alike; they vary vertically and aerially, especially along a well bore perpendicular to the frac direction. Operators must understand differences in shale fabric, in-situ stresses and geologic characteristics within each play, and how the formation will react to various drilling and stimulation designs that change in situ stresses.

Creating Complexity

The step-rate application of hydraulic pressure in shale formations during treatment creates complexity by opening and interconnecting natural fracture systems within the rock matrix. In shale plays, complexity is definitely a good thing, but it is not possible in some shales that do not have natural fractures or where stress differences between minimum and maximum horizontal directions are greater than about 5 percent.

In contrast to a simple biwing planar fracture that is perpendicular to the least principle stress, hydraulic fracturing in ultralow-permeability shales with low stress differences links the well bore to primary and secondary natural fracture mechanisms, resulting in complex networks that provide sharply increased access to the reserves.

Natural fracture systems dominate permeability in shales, with some formations more extensively fractured than others. Natural factures may be closed or even sealed by calcite, salt and clay, but they can be broken down at only 50-60 percent of the pressure that it would take to break rock that is not naturally fractured.

This lower-pressure pathway development is critical to increased complexity, because achieving economic production rates from shale depends on opening fractures–and keeping them propped open– to provide an adequate stimulated reservoir volume and create conductive pathways for hydrocarbons to flow to the well.

Opening more natural fractures increases the stimulated volume of rock surrounding the well bore and provides more reservoir access. A horizontal gas well with 10 frac stages is estimated to create as much as 100 million cubic feet of stimulated reservoir volume in “sweet spots” where there is significant natural fracturing.

Fracturing complexity also contributes to both the high initial rates and the characteristic steep production declines seen in both gas and liquids shale plays. Horizontal gas wells can decline 80-85 percent during the first year of production. Oil is more viscous and depletes slower, but decline rates still are considerably higher than in conventional reservoirs. It is not unusual for “good” wells in the Bakken and Eagle Ford oil window to be making only 30-40 percent of their initial daily rates after one year of production, with most of the declines occurring in the first few months.

Decline rates are a function of flush production from voids within the rock connected by natural fracturing (penetrating the voids yields the high initial rates), and how much rock volume has been effectively stimulated by hydraulic fracturing (accounting for much lower, but more stable rates after the voids have been drained). A related secondary problem is that quickly draining hydrocarbons trapped in the voids reduces reservoir pressure. Flattening decline rates and slowing pressure loss requires propping open as much of the natural fracture system as possible to better drain the stimulated reservoir volume after flush production.

Liquids-Rich Plays

The development of fracture system complexity was first documented in the Barnett Shale. The discovery well–the C.W. Slay No. 1–was drilled in 1981, but it would be 17 years before the Barnett would become a commercial success. Mitchell Energy Corp. had used the same stimulation techniques developed in the Devonian Shale in the 1970s, and had first foam fractured and then gel fractured the well.

However, the C.W. Slay did not produce economic rates until Mitchell Energy performed a third frac job using a large slick-water treatment pumped in steps to high injection rates. The reason was simple: A large-volume, high-rate slick-water frac was the missing element in getting the natural fractures to open.

Gas shale drilling and completion has evolved into a highly efficient “manufacturing” process in a few shales after sufficient experience has been gained with a type of completion. Field development commonly consists of drilling horizontal laterals and pumping multistage slick-water fracs starting with a fine (100-mesh) proppant, usually sand. However, because liquid is more viscous than gas, it is more difficult to flow through long, narrow fractures in low-permeability rock. Therefore, in liquids plays the goal is to create shorter and wider fractures that do not develop as much complexity, but have higher conductivities.

To control fracture length and enhance conductivity, more stages are pumped with more precise placement. Instead of 10 stages, a newer design for a liquids shale well might use 28 or more stages positioned to achieve short, highly-conductive fracs stacked along the entire length of the lateral. Higher-viscosity linear and cross-linked gels that will not leak down the natural fracture system, and higher-strength proppants produce wider fractures that will remain open at higher compressive strengths.

To selectively position stages at the most favorable locations along the lateral, frac points are selected using hydrocarbon shows, formation brittleness and mineralogy data acquired during drilling. Essentially, the operator looks at the mud log to find oil shows and the sonic log to pick the most brittle areas. Hydrocarbon shows indicate areas with higher permeability through the matrix or where the well has intersected a natural fracture or cluster. Brittleness is important because the more ductile the shale, the more difficult it is to establish fracture flow capacity.

Ideally, frac stages are placed at points where the mud log shows oil flowing during drilling and where the shale is most brittle, because that is potentially where voids exist for flush production and where hydraulic fracturing can best establish communication with the reservoir.

Treatment Designs

There are distinct differences between dry gas and liquids shales that impact field development strategies, and different types of fracturing treatment designs are evolving to exploit liquids plays. There is no “one size fits all” solution, and each play has its own drilling and completion styles.

In the Barnett liquids window, operators initially attempted to apply the same frac designs that had worked so well in the gas window. While wells produced commercial quantities of oil immediately after pumping frac jobs, they rapidly tailed off to subeconomic rates as the fractures closed. Recognizing the difficulties in achieving fracture height growth and carrying sufficient proppant into open fractures, operators have adopted a “W” pattern, where wells are landed alternatively high and low in the zone, and frac stages are spaced more densely to increase coverage throughout the zone.

In contrast, the Middle Bakken is a limey silt stone rather than a shale, and has a higher permeability matrix. The key in this play appears to be maximizing well bore contact with the formation and any existing natural fractures to access more reservoir area. The obvious way to develop a larger fracture-to-formation contact area is increasing lateral length and the number of frac stages. Laterals have been extended to more than 9,000 feet with reportedly up to nearly 40 frac stages, and the industry is seeing better performing wells with lower decline rates.

In both the Bakken and Granite Wash, where reservoir contact also is the main consideration, another level of technology may soon be added to field development: multilateral drilling. If the objective is to maximize contact area and fracture growth can be controlled with mechanical or purpose-built intelligent junctions, why not drill two laterals instead of one, perhaps in opposite directions? Bakken operators already are drilling short multilaterals from one vertical bore hole in a “bird’s foot” or “five finger” pattern. Fit-for-purpose fracturing systems eventually will be adapted to meet the needs of Bakken and Granite Wash multilaterals.

The Eagle Ford oil window is different in that while it has a little higher permeability than other shales, it lacks widespread natural fracturing. As a result, fracture systems are more planar and less complex. Many Eagle Ford wells use “hybrid” fracs, where treatments combine slick water with linear or cross-linked gels to minimize leak off and increase fracture width (local stresses act to keep fractures narrow). The Eagle Ford oil play also requires stronger and larger propping agents, including resin-coated sands, ceramics and even intermediate-strength proppant systems that use nanoparticles.

Interestingly, the oil recovery factors in both the Bakken and the Eagle Ford shales is about 1.5 percent of the initial oil in place. This is comparable to the initial estimates of 2 percent gas recovery in the Barnett that was driven to 20-30 percent by technology advances driven by applications learning. Similar technology advances will drive recovery in the Bakken and the Eagle Ford.

On the other hand, the Wolfberry consists of the conventional Wolfcamp, Spraberry and Dean oil intervals, but like shales, it is challenged by low relative permeability issues. The recovery factor in these reservoirs historically has been less than 10 percent, and they are not good waterflood candidates. Production and recovery rates have improved significantly by linking the pay zones through horizontal/deviated drilling and performing multistage completions custom designed for each zone. There are as many as eight productive zones down to the Fusselman at +10,000 feet. Some zones require only acid washing, while others require high-rate gelled fracs and high-strength proppants.

Keys To Success

Most problems encountered in liquid shale plays are related to relative permeability issues caused by oil viscosity and the capillary trapping effects that limit the amount of liquid that can flow through low-permeability rocks, especially considering that many oil windows are at shallower depths with lower pressures. If I was looking for an oil shale play, I would search for a deeper reservoir that was borderline thermally mature with a vitrinite reflectivity between 0.9 (in the oil window for most kerogens) and 1.3 (in the wet gas window).

Relative permeability issues can be mitigated with the right drilling and completion designs. The keys to success include better well placement, improved fracturing effectiveness and increased reserves recovery. In many formations, 3-D seismic is essential to identifying geological hazards as well as assisting in placing the well bore to maximize reservoir access (laterals are generally positioned transverse to the primary orientation of natural fractures). Integrating 3-D seismic with mud log, logging-while-drilling and mineralogy data optimizes placing laterals as well as hydraulic fracturing stages.

It also is important to understand how to initiate, drive and stabilize the fracture system itself–whether opening primary or secondary natural fractures, or generating a planar fracture. The industry is acquiring that understanding through experience, but good petrophysics programs can provide short cuts and eliminate “trial and error” field experimentation.

Managing stress regimes within the formation is extremely important. Several technologies are being incorporated into interpreting 3-D seismic and well logging data to delineate high stress areas with the greatest potential for natural fractures, modeling stress profile changes through the formation, and determining directional differences between the horizontal minimum and maximum stresses, which ultimately control where hydraulic fractures will go. Stress regime management is key in sequential and simultaneous fracturing techniques, which use real-time stress changes created by a frac treatment in an offset well to reroute fractures in adjacent wells to unstimulated parts of the formation.

Slick water remains the dominant fracturing fluid, with several million gallons of water pumped per well in the Haynesville, Barnett, Marcellus, Bakken and Eagle Ford. Slick water’s low viscosity reduces polymer damage and increases penetration into microfractures intersecting the well bore. As treatment pressure is stepped up, the water invades the natural fractures and acts like a wedge to open the fractures and create even more area in which to penetrate.

Because shales are undersaturated with water, they retain much of the frac water that enters natural fractures. This appears to be a positive in some shales because the water acts as a propping agent (it may not be the water itself that helps keep fractures propped, but minerals in the formation that swell or change shape in reaction to water). However, freshwater is not always the best choice. Higher-viscosity gelled systems are better suited to liquids plays to achieve the wider fracs and controlled lengths to accommodate oil and condensate flows.

There also is considerable environmental push back to using freshwater in hydraulic fracturing. Although the industry routinely recycles water flowed back after fracturing, only 30-50 percent of the total water pumped during treatment typically is produced. The majority is absorbed by the shale. One solution is to use saltwater from high-salinity zones that are unsuitable for human or agriculture use. In the Horn River gas shale, Encana and Apache are using a high flow capacity saltwater zone (the Debolt) located above the Horn River as a secondary fluid source. In fact, we are now fracturing Horn River wells with saltwater from the Debolt. Operators are looking for similar zones in other areas, especially in more arid regions.

Artificial Lift

Artificial lift also must be considered in designing well completions for liquids-rich shales. In some cases, artificial lift can be necessary from the first day of production. Obviously, completions in the Barnett and Eagle Ford–both of which produce 38-40 degree API gravity crude oil–need to accommodate artificial lift equipment. But even a multicomponent wet gas mixture can have liquids that condense in the well and require pumps and tubulars. For instance, in the Granite Wash, where liquids are produced as condensate, artificial lift may not be required initially, but it becomes necessary as reservoir pressure declines.

Almost 95 percent of all oil wells in conventional reservoirs are on some form of artificial lift, and that percentage could ultimately be even higher in oil shales. Once flush oil from reservoir voids is produced, flow rates decline very rapidly, and the operator is soon left with a large, interconnected fracture system with reduced reservoir pressure. Artificial lift increases the ability for the liquids to flow to the surface and assists in draining oil through fractures connected to parts of the reservoir outside the near-well-bore area.

Beam pumps, gas lift, electric submersible pumps, plunger lift, jet pumps and progressing cavity pumps all could find applications in producing liquids in horizontal shale wells. Technologically speaking, however, the industry remains challenged for a good system capable of lifting both liquids and gases in horizontal wells. While beam pumping is the most efficient lift method, it also is most susceptible to gas locking in highly deviated wells. The action of the rods working inside the tubing also can create friction problems. There is clearly an opening for artificial lift companies to develop better systems for these long-lateral horizontal wells.

One of the questions impacting lift system selection is whether to drill the lateral section “toe up” or “toe down.” Drilling toe down creates a sump that collects the liquids, but getting an artificial lift system to effectively move liquids from the lowest point of the well at the end of the lateral through the curve and up the vertical section is becoming increasingly more difficult as lateral lengths extend beyond 4,000-5,000 feet.

Consequently, many operators are designing toe-up laterals, with the “belly” at the heel of the lateral near the kickoff point and the horizontal section angled upward 92-95 degrees. Although not as restrictive as the toe-down design, it still limits the type of lift system that can be used.

Some types of pumps are disqualified by the high deviation angles as the well turns horizontal and angles upward toward total depth. Gas lift systems can be run fully horizontal and flow gas, but they begin to lose efficiency at about 20 degrees. To get enough turbulence to efficiently lift liquids, so much gas has to be pumped down hole that the friction begins to create back pressure (the way to increase production is to lower bottom-hole pressure, not increase it!).

Enhanced Recovery

Beyond the need for artificial lift, liquids-rich shales are good candidates for refracs as well as enhanced oil recovery. Shales respond very well to refracs, achieving a much higher success rate than conventional reservoirs. On the EOR front, testing of both mechanical and chemical techniques is occurring in some shales.

A good example is microemulsions that incorporate carbon dioxide. This is a relatively inexpensive chemical technology that could prove very beneficial, especially in formations that do not yield to pure mechanical solutions. The microemulsions clean oil from the rock and mobilize it. They can be applied either as flood treatments or as part of frac treatments. Initial laboratory test results on cores look very promising.

Longer-term, I fully expect to see full-scale CO2 EOR projects in both oil and gas shale plays. The benefits of carbon dioxide flooding are well understood in oil reservoirs, but initial work in CO2 sequestration suggests that carbon dioxide molecules also can displace gas molecules absorbed on the rock. There can be substantial amounts of absorbed gas in place, and it is typically richer gas (ethanes, propanes and butanes). Absorbed gas is not typically produced with free gas because reservoir pressure simply cannot be reduced enough.

The hydrocarbon content of the gas produced in shale plays supports this concept. Initially, shale wells produce gas very high in methane. Over time, methane concentration drops by a small percentage while ethane, propane and butane content increases. This is the absorbed gas coming off the rock as reservoir pressure declines. By displacing absorbed hydrocarbons, CO2 injection represents a potential EOR system to increase recovery rates, especially in deeper and higher-pressure reservoirs such as the Haynesville, Eagle Ford and Bakken.

HOUSTON–North American oil and gas companies and energy services providers are following with a keen and vested interest the news coming out of Washington regarding the “Frac Act,” a congressional proposal to amend the Safe Drinking Water Act to remove a 2005 Energy Policy Act exemption on hydraulic fracturing and effectively give the U.S. Environmental Protection Agency the authority to regulate fracturing operations.

A year or two ago, no one outside of the upstream oil and gas industry even knew what a hydraulic fracture was. Today, senators, congressmen and journalists are discussing fracturing as if they were career frac engineers.

A number of key questions are central to the debate over the Frac Act. Is there a cleaner fossil fuel available than natural gas? Is natural gas abundant in North America and many other parts of the world? Has the growth of unconventional gas in the past decade, primarily from shale, materially added to the recoverable gas (and sometimes oil) reserves in North America, eliminating the need for natural gas imports? Does this improve our energy security? Does incremental production create new sources of tax revenues and local jobs? Is there a more impacting service in low-permeability reservoirs than hydraulic fracturing? Does hydraulic fracturing imperil groundwater supply? Whoa, let’s park the pump trucks here for a moment.

On March 17, 1949, a team comprised of Stanolind Oil Company and Halliburton personnel converged on a well about 12 miles east of Duncan, Ok., to perform the first commercial application of hydraulic fracturing. Later the same day, Halliburton fractured another well near Holliday, Tx. The technique had been developed and patented by Stanolind and an exclusive license was issued to Halliburton to perform the stimulation process. In 1953, the license was extended to all qualified service companies.

Since that fateful day in 1949, hydraulic fracturing has done more to increase recoverable reserves than any other technique. In the more than 60 years following those first treatments, more than 2 million frac treatments have been pumped with no documented case of any treatment polluting an aquifer.

Recent proposals have been initiated by Congress to limit hydraulic fracturing in order to protect groundwater supplies. To limit or eliminate hydraulic fracturing would, of course, be extremely detrimental to domestic supplies of oil and natural gas. In most low-permeability reservoirs, the wells simply will not produce without the fracturing process.

As many as 90 percent of all gas wells completed in the United States have been fractured, and the process continues to be applied in new and innovative ways to boost production of American and worldwide energy in unconventional formations, such as tight gas sands, shale deposits and coalbeds. Conventional, higher-quality reservoirs also benefit from hydraulic fracturing in the form of accelerated recovery.

As a result, hydraulic fracturing is now responsible for 30 percent of domestic oil and natural gas reserves, and has aided in extracting more than 600 trillion cubic feet of natural gas and 7 billion barrels of oil, with much more to come. According to the National Petroleum Council, 60-80 percent of all wells drilled in the United States during the next decade will require fracturing to remain viable. The reason, simply put, is that the low-hanging fruit in the oil and gas industry is largely gone. Going forward, new oil and gas supplies increasingly will have to be recovered from lower-quality and unconventional reservoirs.

The concerns around groundwater contamination raised by Congress are primarily centered on one fundamental question: Are the created fractures contained within the target formation so that they do not contact underground sources of drinking water? In response to that key concern, this article presents the first look at actual field data based on direct measurements acquired while fracture mapping more than 15,000 frac jobs during the past decade.

Unanimous Conclusion

Studies conducted by governmental agencies and respected authorities have unanimously concluded that hydraulic fracturing is safe. The Environmental Protection Agency, the Ground Water Protection Council and the Interstate Oil & Gas Compact Commission all have found hydraulic fracturing nonthreatening to the environment or public health.

As displayed on the Energy in Depth Web site (http://www.energyindepth.com/), the GWPC’s survey of state energy regulatory agencies found no documented cases of contaminated drinking water linked to hydraulic fracturing. The GWPC also concluded that state regulations were sufficient to ensure the integrity of the water supply. A 2002 study conducted by the IOGCC confirmed the GWPC’s conclusion that no evidence could be found of contaminated drinking water related to hydraulic fracturing.

In addition, the EPA conducted an extensive survey of hydraulic fracturing practices and their effect on drinking water in 2004. Focusing on very shallow coalbed methane wells (since their extremely shallow depths would have the highest potential of harming the water supply), the EPA found that several factors (fluid recovery, the small amount of chemicals contained in frac fluids, their dilution in water, and their absorption by rock formations) minimize the potential risks associated with hydraulic fracturing. The EPA agreed with the GWPC and the IOGCC that hydraulic fracturing is safe. More specifically, the EPA concluded that hydraulic fracturing does not create pathways for fluids to travel between rock formations to affect the water supply.

On May 5, 1995, Carol M. Browner, then an EPA administrator and now energy adviser to President Obama, stated, “There is no evidence that the hydraulic fracturing at issue has resulted in any contamination or endangerment of underground sources of drinking water.”

At a state regulators conference in Washington last February, Steve Heare, director of EPA’s Drinking Water Protection division said, “I have no information that states are not doing a good job already (of protecting water supplies).” Despite claims by environmental organizations, Heare also reported that he had not seen any documented cases where hydraulic fracturing was contaminating water supplies.

ICF International recently completed another key study that specifically addressed hydraulic fracturing activities in the Marcellus Shale play in New York. The study confirmed that the EPA findings were valid and found that there would not be any risks to drinking water from hydraulic fracturing in the Marcellus Shale region.

Barnett, Marcellus Data

Extensive mapping of hydraulic fracture geometry has been performed in unconventional North American shale reservoirs since 2001. The microseismic and tiltmeter technologies used to monitor the treatments are well established, and are also widely used for nonoil field applications such as earthquake monitoring, volcano monitoring, civil engineering applications, carbon capture and waste disposal. Figures 1 and 2 are plots of data collected on thousands of hydraulic fracturing treatments in the Barnett Shale in the Fort Worth Basin in Texas and in the Marcellus Shale in the Appalachian Basin.

FIGURE 1

Barnett Shale Mapped Fracture Treatments (TVD)

FIGURE 2

Marcellus Shale Mapped Fracture Treatments (TVD)

More fracs have been mapped in the Barnett than any other reservoir. The graph illustrates the fracture top and bottom for all mapped treatments performed in the Barnett since 2001. The depths are in true vertical depth. Perforation depths are illustrated by the red-colored band for each stage, with the mapped fracture tops and bottoms illustrated by colored curves corresponding to the counties where they took place.

The deepest water wells in each of the counties where Barnett Shale fracs have been mapped, according to United States Geological Survey (http://nwis.waterdata.usgs.gov/nwis), are illustrated by the dark blue shaded bars at the top of Figure 1. As can be seen, the largest directly measured upward growth of all of these mapped fractures still places the fracture tops several thousands of feet below the deepest known aquifer level in each county.

The Marcellus data show a similarly large distance between the top of the tallest frac and the location of the deepest drinking water aquifers as reported in USGS data (dark blue shaded bars at the top of Figure 2). Because it is a newer play with fewer mapped frac stages at this point and encompasses several states, the data set is not as comprehensive as that from the Barnett. However, it is no less compelling in providing evidence of a very good physical separation between hydraulic fracture tops and water aquifers.

Almost 400 separate frac stages are shown, color coded by state. As can be seen, the fractures do grow upward quite a bit taller than in the Barnett, but the shallowest fracture tops are still ±4,500 feet, almost one mile below the surface and thousands of feet below the aquifers in those counties.

The results from our extensive fracture mapping database show that hydraulic fractures are better confined vertically (and are also longer and narrower) than conventional wisdom or models predict. Even in areas with the largest measured vertical fracture growth, such as the Marcellus, the tops of the hydraulic fractures are still thousands of feet below the deepest aquifers suitable for drinking water. The data from these two shale reservoirs clearly show the huge distances separating the fracs from the nearest aquifers at their closest points of approach, conclusively demonstrating that hydraulic fractures are not growing into groundwater supplies, and therefore, cannot contaminate them.

It is true that sound engineering practices make exceptional wells. But it also can be said that sound engineering practices make for a safe environment. Today’s combination of proven, measurement-based engineering practices, optimal fracture treatment designs and direct mapping measurements has led to the production success now enjoyed in low-permeability and unconventional gas plays in the United States and around the globe. Those same best practices also ensure that our drinking water supplies are protected for generations to come.