Barnett shale rising star in Fort Worth basin

EMERGING U.S. GAS RESOURCES-5

Vello A. Kuuskraa, George Koperna
Advanced Resources International Inc.
Arlington, Va.
James W. Schmoker
John C. Quinn
U.S. Geological Survey
Denver

The Mississippian-age Barnett shale of the Fort Worth basin, North Texas, has emerged as a new and active natural gas play.

Natural gas production from the Barnett shale at Newark East field in Denton and Wise counties, Tex., has reached 80 MMcfd from more than 300 wells. However, very little publicly available information exists on resource potential and actual well performance.

The U.S. Geological Survey 1995 National Assessment of U.S. Oil and Gas Resources1 categorized the "Mississippian Barnett shale play" (play number 4503) as an unconventional gas play but did not quantitatively assess this resource. This article, which expands upon a recent USGS open-file resource assessment report,² provides an updated look at the Barnett shale and sets forth a new quantitative assessment for the play.

The Barnett shale is the reservoir for a continuous (unconventional) gas accumulation-in effect a single very large gas field-underlying hundreds and perhaps thousands of square miles (Fig. 1 [109,089 bytes]). USGS Open-File Report 96-254, cited above,² estimated 3.4 tcf (mean value) of undeveloped recoverable gas resource for the Barnett shale.

Using the updated analysis of well performance and drainage area presented in this article, the Barnett shale might hold 10 tcf of technically recoverable natural gas (depending on eventual size of the productive trend, distribution of well performance, and intensity at which the play is developed. A 10 tcf gas volume is equivalent on a BTU basis to a giant 1.67 billion bbl oil field.

The initial and updated resource estimates for technically recoverable gas resources for the Barnett shale follow the general methodology for assessing continuous accumulations used in the USGS 1995 National Assessment.³

Geologic setting

Continuous-type deposit

Mississippian-age organic-rich shales are the reservoir for the Barnett shale unconventional gas accumulation in the Fort Worth basin.

This large, geologically continuous gas accumulation occupies a structurally low position straddling the basin axis. The accumulation is characterized by the presence of gas downdip from water-saturated rocks, no obvious lithologic or structural barriers that separate the updip water and downdip gas, very low (micro-darcy level) matrix permeability, the importance of natural fractures for production, and the absence of truly dry holes.

The Barnett shale was deposited on the southwestern flank of the Southern Oklahoma aulacogen (Fig. 1). Depositional patterns of the Barnett shale reflect the generally NW-SE trend of the aulacogen axis, whereas burial patterns reflect the subsequent depositional and structural influence of the present-day Fort Worth basin.

The Barnett shale lies unconformably on sedimentary rocks of Ordovician age (Ellenburger Group, Simpson Group, Viola limestone) and is conformably overlain by the Pennsylvanian Marble Falls formation.⁴

The majority of published studies of the Barnett shale focuses on surface outcrops on the flanks of the Llano uplift in central Texas (Fig. 1). These exposures are about 150 miles south-southwest of the area of current gas production in Denton and Wise counties, in a direction roughly perpendicular to depositional strike.

Therefore, these outcrop studies are of limited relevance to the petroleum geology of the Barnett shale in the much deeper subsurface. Henry's comment in 1982 that the subsurface Barnett shale "is so modestly represented in the literature" still holds true today.

The Barnett shale includes various local, informal members that differ somewhat in their names and physical and geochemical properties. The shale interval can be divided into an upper and a lower unit, separated by the Forestburg limestone. However, in this study, the Barnett shale is assessed as a single unit.

Regionally, the thickness of the Barnett shale approaches 1,000 ft near the southwest fault boundary of the Southern Oklahoma aulacogen and thins to the southwest as it crops out along the flanks of the Llano uplift.⁴ The Barnett shale is approximately 500 ft thick near the center of the present producing trend. Most wells are completed in the lower part of the Barnett shale with perforations typically spanning 200-300 ft.

Adsorbed gas

The Barnett shale is a black, organic-rich shale with a high resistivity value, gamma-ray intensity of 150-400 API units, and mean bulk density of about 2.50 g/cc.

This description is similar to that for black shales in the U.S. of Devonian-Mississippian age that produce natural gas, such as the Antrim shale of the Michigan basin, New Albany shale of the Illinois basin, Woodford shale of the Anadarko basin, and "Devonian" shales of the Appalachian basin.

Based on analogies with these other black shales, the high resistivities of the Barnett shale can be attributed to the generation and retention of hydrocarbons.⁵ The high gamma-ray intensity and low bulk density are indicative of relatively high organic-matter content. A mean bulk density of 2.50 g/cc for the Barnett shales suggests a mean total organic carbon (TOC) content of about 4.5 wt %,⁶ a value comparable to that for the Woodford shale.

TOC measurements on core samples from the lower Barnett interval in the 2 T.P. Sims well, collected by the Gas Research Institute and Mitchell Energy Corp., showed an average TOC of 4.5 wt %, consistent with the bulk-density based estimates.

The organic-rich shales are a storage site for adsorbed gas. The organic matter in the shale contains approximately 60 scf of natural gas/ton of reservoir rock,⁷ about 25% of the natural gas stored in the pore space. This gas is released (desorbed) as formation pressure is lowered and contributes to the long life of Barnett shale gas wells.

Play foundations

Historical perspective

Until recently, the continuous (unconventional) Barnett shale accumulation has been overlooked as a prospective gas play.

Henry⁴ emphasized the potential for conventional fields in carbonate-dominated Barnett-age strata deposited on the shallower parts of the southwest shelf but did not discuss the potential for gas production from more basinward shale facies. The current gas producing area was not included in the 1989 Atlas of Major Texas Gas Reservoirs.⁸

Active development

Today, gas production from the Barnett shale is firmly established. At the end of 1997, over 300 Barnett shale wells concentrated in Denton and Wise counties (Fig. 1) were producing 80 MMcfd (authors' update of Reeves and others, 1996).⁹

Significantly, drilling in the Barnett shale has accelerated since expiration at the end of 1992 of the Sec. 29 Nonconventional Fuels Tax Credit, indicating that this gas play has become economic on its own terms.

The active development drilling in the area appears to be influenced by proximity to existing infrastructure as well as by an improved understanding of the geologic controls governing the productive areas.

Examination of Barnett shale well drilling and gas production since the mid-1980s showed that the Sec. 29 tax credit helped stimulate the early drilling and testing of technology that "opened the door" for this gas play. However, the great bulk of today's gas production from the Barnett shale has been developed without the support of tax credits.

GRI sponsored studies

In the early 1990s, GRI supported a series of engineering studies that contributed significantly to reservoir characterization of the Barnett shale gas play.

In a series of GRI reports summarized by Lancaster and others,¹⁰ the Barnett shale was characterized as a layered reservoir in which well deliverability is greatest in thinner, higher permeability naturally fractured zones, and much of the gas-in-place is held in thicker, extremely low permeability intervals.

Subsequent evaluation of this formation, as reported here, indicates that the gas flow units are considerably more complex than this initial characterization, and the distribution of permeability depends on the local interplay of structure, faults, and stress.

A particularly significant finding of the GRI-sponsored studies was that natural fractures in the Barnett shale have a mean strike of N. 114° E., whereas induced fractures have a mean strike of about N. 60° E. Because of this indicated shift in stress-field orientation from past to present, hydraulic fractures induced from a wellbore may intersect, rather than parallel, the natural fracture system.

An important research question for future Barnett development is whether the present stress-field direction, measured in the current producing area, results from the influence of local fault systems or represents a regional stress field with affinities to the extensional tectonics of the Gulf Coast.

Understanding drainage

The initial analysis of the Barnett shale, using the two-layer model mentioned above, forecast long hydraulic fracture lengths (1,200 ft) and a drainage area of 320 acres/well.10 These results helped establish the 320 acre well-spacing "mind-set" currently followed by industry. However, based on analogs from other naturally fractured reservoirs, the Barnett natural fracture system is probably tectonically controlled and dispersed throughout the shale interval, providing little foundation for the two-layer model and the resulting large 320 acre drainage assumption.

Recent evaluations of gas production data and hydraulic fracture performance, given in this article, show fracture lengths on the order of 200 ft and limited drainage areas of 10-30 acres/well.

Well performance

Basic data

Total depths of most Barnett gas wells in Denton and Wise counties are 7,200-8,200 ft. As such, the Barnett shale continuous gas accumulation is deeper than Appalachian Devonian shales (2,000-5,000 ft) and Michigan Antrim shale (about 1,500 ft).⁹

The mean liquids-to-gas ratio for the Barnett shale is about 1.5 bbl NGL/MMcf of gas. This ratio may depend on the thermal maturity of the shale and thus may vary systematically within the accumulation.

Water production is erratic and generally low. Many recent wells have little or no reported water production, perhaps indicating an improvement in completion practices or more favorable well locations.

Unlike Antrim shale and coalbed-gas operators, Barnett shale producers do not need to practice rigorous dewatering to improve gas desorption and production.

Type-wells

To better understand the reservoir properties of the Barnett shale, seven wells representative of the evolution in completion practices and technology were selected for in-depth study using production history type-curve matching. The reservoir properties used as input for the history matching were derived from work sponsored by GRI and are shown on Table 1 [24,420 bytes]. The six "type" wells and one "special" well are listed on Table 2 [67,333 bytes].

Wells 1 and 2 reflect the interval selections, completion practices, and performance of the initial 74 wells, drilled in the 1980s. These wells were completed in about 180 ft of shale interval with small fracs (Table 2).

These initial wells have EURs of 0.4-0.5 bcf/well;
The wells were drilled into reservoir settings with 2 to 3 micro-darcies of matrix permeability; and,
In general, the wells drain a little over 10 acres.

As an example representative of the initial 74 wells drilled into the Barnett shale gas play in 1985-90, Fig. 2 [50,743 bytes] provides 8 years' production history for Well 2 (Royce Chism 1-T) that has produced 0.38 bcf of gas to date. Fig. 3 [57,001 bytes] provides the type-curve history match for this well, showing an EUR of about 0.5 bcf. (For more information on the type-curve history matching, see Cox and others.¹¹)

Fig. 3 also shows the estimated permeability, drain- age area and completion efficiency for this well.

Special Well 3 (4 Stella Young, drilled in 1986) was a GRI and Mitchell Energy study well that helped define the productive potential of the Barnett shale (Fig. 4 [46,890 bytes]). This well, with over 10 years of production history, provides an excellent benchmark for evaluating the impact of improved completion practices on well performance (Table 2).

Well 3 has produced 0.78 bcf (through yearend 1996) toward an EUR of 1.16 Bcf, and
The type-curve match (Fig. 5 [54,844 bytes]) indicates that the well penetrated a reservoir with about 3 micro-darcies of permeability, drains an area of 30 acres, and has a hydraulic fracture half-length of 190 ft.

Type Wells 4 through 7 reflect the improved completion practices for the 217 Barnett shale wells drilled in 1991-96. These wells were completed in 230-290 ft of shale interval with moderate (160-190 ft) high-conductivity hydraulic fractures (Table 2).

These four wells have EURs of 0.8-0.9 bcf/well;
The wells were drilled into areas of the Barnett shale with permeability of about 2 micro-darcies; and,
Drainage areas are 11-21 acres/well.

As an example of the analysis of Wells 4 through 7, Fig. 6 [43,652 bytes] provides the initial three years of production history for Well 7 (A.L. Peterson 3-T) that had produced 0.33 bcf of gas as of mid-1997. Fig. 7 [55,546 bytes] provides the type-curve history match showing an EUR of 0.83 bcf plus the other matched reservoir parameters for this well.

Resource assessment: method, input data

Methodology

A continuous accumulation or play can be regarded as a collection of contiguous hydrocarbon-charged cells for which EUR is represented by a probability distribution. The key steps in the resource assessment procedure (described in detail by Schmoker³) are establishing the areal extent of the play, the well spacing (which equates to cell size), the success ratio of wells, and the EUR of successful wells (cells). The statistical combination of these parameters yields an estimate of undeveloped technically recoverable resources.

This assessment methodology has the benefit of simplicity in that the in-place hydrocarbon volume does not need to be established or a recovery factor calculated. The approach relies on well production histories to characterize the recovery expected from undrilled portions (cells) of the accumulation. Reservoir characteristics such as porosity, permeability, and water saturation and the efficiency of the well completion are embedded in the well production histories.

A limiting aspect of the assessment methodology is that, unless adjustments are made, existing technology and development practices are projected to continue into the future. Gas production from Barnett shale wells has improved significantly in recent years, with the gas recoveries from recent wells about 1.7 times those of pre-1990 wells (Table 3 [29,945 bytes]).

Given that the Barnett shale is, to a large extent, a technology driven gas play, considerable thought needs to be given to the question: How might resource assessment methodologies better capture and reflect continued technology progress?

One analysis involved using data that track the performance of all Barnett shale wells for which data are available according to the year they were drilled using the best 6 months of gas production as the measure. The data show a steady improvement in well performance through 1992 with a small drop in performance for more recent wells. A return to improving well productivity may occur if less damaging well completion and stimulation practices evolve and advanced natural fracture detection technology is applied to more efficiently find other "sweet spots."

Conversely, if the best well locations in the play have been mostly drilled, a return to improving well productivity might not occur. Such issues illustrate the assessment question posed in the preceding paragraph.

The recently drilled 1 Callejo step-out well, located near the northwest corner of Dallas County, 12 miles east of Newark East field, provided early indications of strong performance from the Barnett shale. This well, drilled by AFE Oil and Gas Consultants, was reported to have flowed at 2 MMcfd against a 1,569 psi flowing tubing pressure and had an estimated absolute open flow of 15 MMcfd.¹²

A review of the first two months of production data for this well showed an average daily rate of around 300 Mcfd, somewhat lower than the typical performance of recent Barnett shale wells in Wise and Denton counties.

Maintaining current information on well performance, success ratio, and areal extent of a gas play is essential for providing updated resource assessments and capturing the impacts of technology progress. The resource assessment of the Barnett shale provides one example of how this can be accomplished.

Extent of the play

The areal extent and boundaries of the Barnett shale continuous gas accumulation provide the first set of input data for the resource assessment.

The present producing trend is in Denton and Wise counties. The boundaries of the productive accumulation are not yet known outside this trend. To reflect the uncertainty of the play's areal extent, three definitions of areal extent are discussed in this article (Fig. 1).

1. Minimum area.
The boundaries of the minimum area enclose 285 sq miles and surround the existing trend of producing Barnett wells. The rationale is that existing wells define a favorable area-a sweet spot-having geologic characteristics that might not to occur elsewhere. The boundaries of the minimum area straddle the axis of the Fort Worth basin in Denton and Wise counties. The updip (southwest and northeast) limits correspond approximately to the -6,600 ft structural contour on the Ellenburger Group. This structural contour is assumed to map a constant (but unknown) level of thermal maturity within the Barnett shale. No measured or published vitrinite reflectance (Ro) values for the Barnett shale were available for this study.

2. Intermediate area.
The boundaries of the intermediate area enclose 2,479 sq miles and include the present-day productive area plus northwest/southeast extensions of this area. The rationale is that the geologic characteristics of the present producing area might continue along the basin trough, extending the existing productive trend.

The extension area generally parallels trends of depositional environment, basin structure, erosional subcrop of underlying Ordovician age rocks, and (presumably) thermal maturity. Few wells within the extension area have fully penetrated and adequately tested the Barnett shale.

Maximum depth to the base of the Barnett shale in the intermediate area occurs at the intersection of the basin axis with the Ouachita structural front, where the top of the Ellenburger Group is approximately 9,800 ft below sea level. By analogy to the current producing trend, thermal-maturity levels in the Barnett shale within the extension area are assumed to exceed the threshold for thermogenic gas generation.

The boundaries of the intermediate area straddle the axis of the Fort Worth basin and track the -6,600 ft Ellenburger structural contour or are coincident with boundaries of the maximum area (Fig. 1).

To the northeast, these coincident boundaries are dictated by erosional removal of the Barnett shale on the Muenster arch, which parallels the southwest fault boundary of the Southern Oklahoma aulacogen and separates the Fort Worth basin from the Marietta basin farther to the northeast.

To the northwest, these coincident boundaries reflect an erosional Barnett shale subcrop under the Pennsylvanian section.

To the east and southeast, the coincident boundaries follow the trace of the Ouachita structural front. (It is possible that the Barnett shale and its continuous gas accumulation extend farther east, under the Overthrust sheet.) The southeast-trending straight-line boundary segment at the southern extreme of the extension area has been drawn to exclude the carbonate-prone shelf platform to the southwest.

3. Maximum area.
The boundaries of the maximum area (Fig. 1) enclose 4,227 sq miles. The rationale is that levels of thermal maturity within the Barnett shale might be sufficient for gas generation to extend updip beyond the boundaries of the intermediate area. The maximum-area model also assumes that other favorable characteristics of the present producing area extend updip. About a dozen widely scattered wells have penetrated the Barnett shale within the added area.

The boundaries of the maximum area (Fig. 1) are somewhat arbitrarily drawn to track the -5,250 ft Ellenburger structural contour, except for the segments described in the preceding section. Most of the added area extends shelfward, to the west and southwest.

Well performance

Historical data on gas production rates and EUR provide the second set of input data for the resource assessment.

1. Annual production.
The peak-yearly-production (PYP) and peak-monthly-production (PMP) probability distributions representing 173 Barnett shale gas wells with data through fall 1994 are provided on Fig. 8. [35,247 bytes]

The median of the PYP distribution is 135 MMcf/12 month period. Eighty percent of productive Barnett gas wells have PYP greater than 76 MMcf/12 month period, and 20% have PYP greater than about 200 MMcf/12 month period.

The PYP probability distribution of Fig. 8 conceals a time and technology dependence in the production data (Fig. 9 [35,278 bytes]). PYP values for more recent Barnett gas wells are significantly higher than those for the older wells. The factor by which PYP values for the newer wells exceed those for the older wells ranges from about 1.5 for the higher productivity wells to more than 6 for the lower productivity wells.

Barnett shale gas wells are routinely stimulated by hydraulic fracturing. A portion of the improved performance of more recent wells is due to increases in treatment size and completion interval (Table 2), coupled with a better knowledge of the mechanical properties of underlying strata.

The northeastern half of the study area is underlain by the Viola limestone or the Simpson Group, whereas the southwestern half is underlain by the Ellenburger Group.⁴ The mechanical properties of strata influence the ability to contain the hydraulic fracture in the pay zone.⁹

Another portion of the improvement in well performance might stem from the use of seismic and other methods for locating new wells in the naturally fractured "sweet spot" portions of this vast gas accumulation.

2. Ultimate recoveries.
Gas in-place in the present producing area of the Barnett shale continuous accumulation was estimated at 25-35 bcf/sq mile.⁹ Reservoir properties computed in the GRI-supported 2 T.P. Sims research well (located in the present producing area) equate to a gas in-place volume of about 42 bcf/sq mile plus 10 bcf/sq mile of adsorbed gas.¹⁰ As shown by the two EUR probability distributions of Fig. 10 [43,309 bytes] (which are for spacing of 2 wells/sq mile), recoveries realized in actual practice are typically a small fraction of the gas in-place.

The EUR probability distribution labeled "A" (Fig. 10) is based on decline-curve analysis, using the current well spacing of 320 acres and an abandonment rate of 0.1 MMcf/month, for the 20 oldest wells in the trend. These 20 wells were drilled, from June 1982 to December 1985. The median EUR of Distribution A is 0.35 bcf. Eighty percent of the wells have EURs greater than 0.12 bcf, and 20% have EURs greater than 0.76 bcf.

Whereas the older wells of Distribution A provide longer production histories for analysis, the peak-yearly-production data of Fig. 9 clearly demonstrate increased production rates for more recent Barnett shale wells. Distribution A (Fig. 10) ignores recent technological progress and is too conservative for use in assessing future resources.

To correct for this problem, Distribution B was constructed by multiplying Distribution A by the ratio of PYP (newer wells) to PYP (older wells). This ratio was calculated from Fig. 9 for selected percentiles.2 This multiplier, which can be thought of as a technology-improvement factor, ranges from 6.2 at the 95th percentile (low EUR wells) to 1.5 at the 5th percentile (high EUR wells). Distribution B incorporates the impact of advancing technology on the EURs of future wells.

The median EUR of Distribution B is 0.6 bcf. Eighty percent of productive Barnett shale gas wells have EURs greater than 0.3 bcf, and 20% have EURs greater than 1.2 bcf (Fig. 10).

Although considerable opportunity exists for error, Distribution B of Fig. 10 seems reasonable and was used by Schmoker and others2 for their resource assessment of the Barnett shale continuous gas accumulation.

Resource assessments

1996 USGS assessment

The Barnett shale is an emerging natural gas resource that was not quantitatively assessed in the USGS 1995 National Assessment.

In 1996, the USGS prepared a special assessment2 of technically recoverable natural gas resources for the Barnett shale continuous gas accumulation. This special assessment is summarized in the following two subsections.

1. Input parameters.
The data used for the 1996 USGS assessment of the Barnett Shale gas play are given in Table 4. [29,945 bytes] Values of the EUR distribution are taken from Fig. 10 [43,309 bytes] (Distribution B). The minimum, median, and maximum numbers of untested cells are calculated from the respective areal extents by assuming 320 acres/well (corresponding to present-day well spacing) and subtracting the 210 tested cells (180 productive and 30 nonproductive) drilled as of September 1994.

The projected success ratio of 0.86 is based on drilling results in and near the current producing area of Denton and Wise counties. The liquids/gas ratio of 1.5 bbl NGL/MMcf is used to assess volumes of natural gas liquids.

The present-day well spacing of only 2 wells/sq mile, the relatively low recovery percentage, and the very limited reservoir drainage of 11-30 acres/well demonstrated here (Table 2) provide ample opportunities for infill development.⁹

If successful, infill development could substantially increase the technically recoverable resources of the Barnett shale. The significance of more intense development upon the resource assessment is further examined later in this article.

2. Assessment results
Using the above set of input data, the technically recoverable gas resources in the Barnett shale are estimated at 3,360 bcf mean estimate (Table 4, Fig. 11 [97,679 bytes]).

Based on the assumptions set forth on Table 4, a 19 in 20 chance exists for the occurrence of an undiscovered, technically recoverable gas volume of at least 1,781 bcf (F95) and a 1 in 20 chance exists for a gas volume exceeding 5,612 bcf (F5). The volume of natural gas liquids associated with these gas resource estimates ranges from 2.67 (F95) to 8.42 (F5) million bbl (Fig. 11).

1998 assessment

The Barnett shale provides an excellent example of how resource assessments need to be updated over time. To better reflect this idea, we have begun to use the term "forecasts" rather than assessments, with the understanding that resource forecasts, like those for long term weather, need to be progressively adjusted as more information becomes available.

To keep resource assessments (or forecasts) for the Barnett shale current, the key input parameters of area, well spacing, success rate, and EUR distribution should be periodically re-examined.

1. Input parameters.
The principal input parameter requiring re-evaluation for the "1998 updated assessment" is a growing recognition that the Barnett shale could be more intensely developed, particularly in the higher productivity areas. Based on the analysis of well drainage presented on Table 2, much closer well spacing (cell sizes) of 80-160 acres/ well (as compared to 320 acres/well) would provide significantly increased gas recovery from the more productive areas of the accumulation.

This updated resource forecast assumes that the best 30% of the intermediate area will be intensively drilled at 80 acres/well, the next best 30% of the area will be drilled at 160 acres/well, and the lowest productivity 40% of the area will not be further developed past today's standard 320 acre spacing.

The basic EUR distribution (B, Fig. 10) was not changed for the "1998 updated assessment" but deserves a revisit as additional wells are drilled in this gas play.

2. Assessment results.
Using the updated input data, the mean technically recoverable gas resources in the Barnett shale are estimated at 10,000 bcf (Table 5 [104,902]), with the bulk of the recoverable gas being from the best potential area.

Assessments

Three examples, summarized below and on Table 5, show the impact that regularly updating the key assessment input parameters can have on the resource outlook for the play:

The first assessment-the "initial performance" model for the Barnett shale (Table 5, Column 1)-is prepared (in 1998) using data available as of 1990, after the first group of 74 wells was drilled. With relatively low typical EURs of 0.35 bcf/well (Distribution A, Fig. 10) and well spacings of 320 acres, the Barnett shale would be forecast to hold only 1,442 bcf of technically recoverable gas.
The second assessment-the 1996 USGS special assessment (Table 5, Column 2, and Table 4)-uses a higher EUR representative of more recent wells of 0.837 bcf/well (the mean of Distribution B) together with the current practice of spacing wells on 320 acres. This leads to a recoverable resource estimate of 3,360 bcf.
The third assessment-reflecting the "intensive development" model (Table 5, Column 3)-is prepared on the basis that industry recognizes the very limited drainage area being achieved by Barnett shale wells, finds the play's higher productivity areas, and intensively develops them. Thus, this forecast optimistically assumes that close well spacing and high EUR are correlated. This model yields a forecast of 10,000 bcf of technically recoverable gas for the Barnett shale. Note that this model forecasts that most of the potential resource resides in the intensely developed "sweet spots" (Table 5).

Future assessments might revisit the well productivity (EUR) and success rate inputs for the Barnett shale gas resource, in addition to play area and well spacing (cell size).

Summary

Gas production from the Barnett shale continuous gas accumulation of the Fort Worth basin has been firmly established by more than 300 wells in Newark East field. However, play boundaries are not yet delineated by drilling. The eventual productive area might be as small as 285 sq miles or as large as 4,200 sq miles.

The 1996 USGS special assessment of undiscovered technically recoverable gas resources for the Barnett shale is about 3.4 tcf (mean estimate). An initial forecast for this gas play, circa 1990, reflecting low performance by the initial wells, would have assigned only about 1.4 tcf to this play. Intensive development of the higher quality areas in which wells are projected to be drilled at 80 and 160 acre spacings leads to a "1998 updated forecast" of 10.0 tcf for the Barnett shale gas play.

The economic risk associated with developing the Barnett shale continuous gas accumulation is not so much that of finding gas but rather that of finding natural fractures and high quality pay to support sufficient rates of gas production. Drilling in the Barnett shale has accelerated in recent years, suggesting that at least a portion of the play is economic at today's prices and technology. Continued improvements in natural fracture detection, reservoir characterization, and well completion technologies could appreciably extend and improve the economics of this giant unconventional gas play.

Acknowledgments

The authors express their appreciation to the Gas Research Institute, particularly Tom H. Fate and Charles F. Brandenburg, for supporting this study of emerging gas resources. This article benefited from maps generated by Mitchell E. Henry, from technical support by Robert A. Crovelli, Vito F. Nuccio, and Timothy C. Hester, and from manuscript reviews by Ronald C. Johnson, Mitchell Henry, Katherine L. Varnes, and Thaddeus S. Dyman.

References

U.S. Geological Survey National Oil and Gas Resource Assessment Team, 1995 National Assessment of U.S. oil and gas resources, USGS Circular 1118, 1995, 20 p.
Schmoker, J.W., Quinn C.J., Crovelli, R.A., Nuccio, V.F., and Hester, T.C., Production characteristics and resource assessment of the Barnett shale continuous (unconventional) gas accumulation, Fort Worth Basin, Texas, USGS Open-File Report 96-254, 1996, 20 p.
Schmoker, J.W., Method for assessing continuous-type (unconventional) hydrocarbon accumulations, in Gautier, D.L., Dolton, G.L., Takahashi, K.I., and Varnes, K.L. (eds.), 1995 National Assessment of U.S. oil and gas resources-Results, methodology, and supporting data, USGS Digital Data Series DDS-30 [CD-ROM], 1995.
Henry, J.D., 1982, Stratigraphy of the Barnett shale (Mississippian) and associated reefs in the northern Fort Worth basin, in Martin, C.A. (ed.), Petroleum geology of the Fort Worth basin and Bend arch area, Dallas Geological Society, Dallas, 1982, pp. 157-177.
Schmoker, J.W., and Hester, T.C., Formation resistivity as an indicator of oil generation-Bakken formation of North Dakota and Woodford shale of Oklahoma, The Log Analyst, Vol. 31, No. 1, 1990, pp. 1-9.
Hester, T.C., Schmoker, J.W., and Sahl, H.L., Log-derived regional source-rock characteristics of the Woodford shale, Anadarko basin, Okla., USGS Bull. 1866-D, 1990, pp. D1-D38.
Luffel, D.L., Lorenzen, J., Curtis, J.B., and Low, P.F., Formation evaluation technology for production enhancement, Report to the Gas Research Institute, GRI Contract No. 5086-213-1390, 1991.
Kosters, E.C., Bebout, D.G., Seni, S.J., Garrett, C.M., Jr., Brown, L.F., Jr., Hamlin, H.S., Dutton, S.P., Ruppel, S.C., Finley, R.J., and Tyler, Noel, Atlas of major Texas gas reservoirs, Gas Research Institute, Chicago, and Bureau of Economic Geology, Austin, 1989, 161 p.
Reeves, S.R., Kuuskraa, V.A., and Hill, D.G., New basins invigorate U.S. gas shales play, OGJ, Jan. 22, 1996, pp. 53-58.
Lancaster, D.E., McKetta, Steve, and Lowry, P.H., Research findings help characterize Fort Worth basin's Barnett shale, OGJ, Mar. 8, 1993, pp. 59-64.
Cox, D.O., Kuuskraa, V.A., and Hansen, J.T., Advanced type curve analysis for low permeability gas reservoirs, SPE paper 35595 presented at the Gas Technology Conference, Calgary, Alta., 1996 (not printed in proceeedings).
More drilling follows Fort Worth basin find, OGJ, May 5, 1997, p. 134.

The series

Part 1-Kuuskraa, Vello A., Outlook bright for U.S. natural gas resources, OGJ, Apr. 13, 1998, p. 92.
Part 2-Dyman, Thaddeus S., Schmoker, James W., and Root, David H., USGS assesses deep undiscovered gas resource, OGJ, Apr. 20, 1998, p. 99.
Part 3-Reeves, S.R., Kuuskraa, J.A., and Kuuskraa, V.A., Deep gas poses opportunities, challenges to U.S. operators, OGJ, May 4, 1998, p. 133.
Part 4-Collett, Timothy S., and Kuuskraa, Vello A., Hydrates contain vast store of world gas resources, OGJ, May 11, 1998, p. 90.
End Part 5 of 6

The Authors

Vello A. Kuuskraa is president of Advanced Resources International Inc. He was a 1985-86 Society of Petroleum Engineers Distinguished Lecturer, served on the Secretary of Energy's "Assessment of the U.S. Natural Gas Resource Base, and was a member of the National Academy of Sciences' Committee on the National Energy Modeling System. He received an MBA degree (highest distinction) from the Wharton School, University of Pennsylvania, and a BS degree in mathematics/economics from North Carolina State University.

George J. Koperna Jr. is a staff reservoir engineer for ARI. His primary areas of work are property evaluation, reservoir simulation, and production forecasting for conventional and unconventional gas resources. He is a graduate of West Virginia University with BS and MS degrees in both petroleum and natural gas engineering.

James W. Schmoker is a geophysicist with the Central Region Energy Team of the U.S. Geological Survey in Denver, where he has been employed since 1974. His areas of research include petroleum resource assessment, studies of reservoir quality (in particular the evolution of porosity with burial), and the petroleum potential of organic-rich black shales. He holds BS and MS degrees in physics from the University of Minnesota and a PhD in geophysics from Virginia Polytechnic Institute and State University.

John C. Quinn was involved in a wide range of production activities with industry in the Rocky Mountain region until 1994, when he joined the USGS as a petroleum reservoir engineer. Since then he has specialized in developing production models for natural gas fields and reservoirs and engineering aspects of petroleum resource assessment. He received BS and ME degrees in petroleum engineering from Colorado School of Mines in 1986 and 1992, respectively.