SPE 108045

A Case Study in the Bakken Formation: Changes to Hydraulic Fracture Stimulation

Treatments Result in Improved Oil Production and Reduced Treatment Costs
Z.D. Phillips, Pinnacle Technologies; R.J. Halverson, SPE, S.R. Strauss, SPE, and J.M. Layman II, SPE, Hess Corp.; and
T.W. Green, Bass Enterprises

Copyright 2007, Society of Petroleum Engineers

This paper was prepared for presentation at the 2007 SPE Rocky Mountain Oil & Gas
Technology Symposium held in Denver, Colorado, U.S.A., 1618 April 2007.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, Texas 75083-3836 U.S.A., fax 01-972-952-9435.

Abstract
The Bakken Formation in the Williston Basin has experienced
multiple periods of activity and the current level of
exploration/exploitation in North Dakota is widespread. The
prevalent drilling practices include single and multi-lateral
horizontal wellbores requiring hydraulic fracture stimulation
to be productive.
Many operators are experimenting with an assembly of
completion and stimulation techniques in order to achieve
economic production. Knowledge and techniques used in
other basins with shale production have been applied with
positive results. This paper describes the methodologies
applied and subsequent production enhancements.
By changing stimulation techniques and methodologies along
with an evolving thought process, economic productivity has
been achieved enabling significant development of the
resource for the operator.
Introduction
The Williston Basin encompasses portions of Montana, North
and South Dakota, and Manitoba and Saskatchewan, Canada.
The oil-bearing Middle Bakken formation in the Williston
Basin has attained a large degree of visibility in recent years.
The first oil production occurred in the Bakken in 1953, with
the first horizontal Bakken well being completed in September
1987 by Meridian Oil Inc. 1 All wells discussed in this paper
are horizontal completions in several fields in western North
Dakota, particularly the Beaver Lodge (BLMU), Blue Buttes
(BBMU) and Tioga (TMU) areas. A map of these fields is
given in Figure 1.
In North Dakota, the Middle Bakken is an extremely lowpermeability formation, with typical matrix permeabilities in

the microdarcy to nanodarcy range. Reservoir quality varies

widely, and many operators drill long horizontal sections to
maximize reservoir contact. These laterals are typically
completed either openhole or with uncemented, pre-perforated
liners. Fracture stimulation is typically required to make
productive wells.
Early Bakken development with vertical wells posed some
interesting conclusions. The best initial fracturing results were
obtained with high concentrations of proppant2, however two
refracturing treatments incorporating less than 30,000 pounds
of ceramic proppant actually reduced total well production,
while three restimulation attempts with an average of 70,000
pounds of sand were effective; suggesting that increased fluid
and proppant volumes may be a greater key to success than an
increase in fracture conductivity when restimulating vertical
Bakken wells.2
However, most recent Bakken development has utilized
horizontal wellbores. When stimulating horizontal wells, it is
impossible to optimize the treatment design without an
understanding of the created fracture geometry and the
intersection with the wellbore. If the fractures are longitudinal
(Figure 2) with excellent wellbore/fracture intersection,
essentially any propping material will provide acceptable
conductivity.

However, if the fractures are transverse (Figure 3) the

produced oil and gas must travel hundreds of feet within the
proppant pack, and flow convergence around the
circumference of the well imposes very high pressure losses
near wellbore.
With transverse fractures, proppant
characteristics are key, and most transverse fractures will be
conductivity-limited.

SPE 108045

The Bakken formation is upper Devonian-lower Mississippian

in age and is comprised of three sub-intervals. The Upper
Bakken Shale is typically 15 thick, exhibits fissile fracturing,
and is characterized by high total organic carbon (~ 8%). The
Lower Bakken Shale is similar to the Upper Bakken Shale but
is commonly 30 thick and has slightly less organic material
present.

Many recent treatments in the Middle Bakken have utilized

high-viscosity crosslinked fluids and ceramic proppants
pumped at moderate rates to create highly conductive
hydraulic fractures. These conductive fractures are justified if
transverse fracturing is observed; however, in most cases there
is insufficient knowledge of whether created fractures are
longitudinal or transverse. Significant height growth resulting
in decreased fracture width is observed in many Bakken
treatments, and the use of high-viscosity fluids was also used
to offset this decreased width and effectively place proppant.3
The use of this fluid/proppant system has been successful in
Montana, where the Bakken often exhibits higher
permeability.4 However, the North Dakota Bakken features
lower permeability and lower reservoir quality, and
crosslinked fractures have often resulted in disappointing posttreatment production rates.
Recent treatments have focused upon the use of very lowviscosity fluids to reduce height growth and thereby achieve a
longer hydraulic fracture with more fracture face/reservoir
contact. Maximum proppant concentrations were initially
reduced as it has been shown that in extremely lowpermeability environments, low proppant concentrations are
capable of achieving high values of dimensionless
conductivity. 5 These treatments also have eliminated ceramic
proppants in favor of 40/70-mesh white sand and are pumped
at high rates ranging from 80 to 120 bpm to effectively treat a
greater portion of the lateral with acceptable proppant
placement.
While there is still a great deal of treatment optimization
ahead, initial results from these treatments are promising.
Although a small sample size makes it difficult to draw
definite conclusions, early production data indicates a 30%
increase in production achieved with a corresponding
reduction in treatment costs. Refracture treatments with lowviscosity fluids have also shown favorable outcomes.
Geology
The Lodgepole formation overlies the Bakken formation and
is comprised of approximately 700 of tight, cryptocrystalline
limestone. The lower 30 of the formation typically exhibits
fracturing with an increase in reservoir quality.

The Middle Bakken member ranges from 40-75 in thickness.

Lithologies vary from argillaceous dolostones and siltstones to
clean, quart-rich arenites and oolitic limestones. The Upper
and Lower Bakken Shales were deposited in anoxic conditions
which preserved the organic content for hydrocarbon
generation. The shales are believed to have generated several
hundreds of billions of barrels of OOIP for the Williston
Basin.
The Three Forks formation lies beneath the Bakken formation.
The upper 15 is predominantly a silty, dolomitic sand referred
to as the Sanish Sand and the remaining section is
characterized by alternating apple-green shales and red-brown
dolomitic mudstones and wackestones.
The Bakken is typically found at a depth of 9,600 ft to 10,400
ft TVD in the study area. Typical fracture gradients have been
in the range of 0.80 to 0.86 psi/ft. A Bakken type log is given
in Figure 4.
Reservoir Characterization
The Middle Bakken member has been targeted for horizontal
drilling and development. Porosity in the Middle Bakken
ranges from 4-10% and is composed of primary, depositional
intergranular and solution-enhanced, intercrystalline pores.
Porosity value correlations from core-to-log data closely
match and lithologic variations do not seem to adversely affect
this relationship.
Matrix permeability measurements from core data indicate
nano- to microdarcy permeabilities (0.0001-.01 md) in the
productive Middle Bakken member. Intraformational
fracturing has been observed and is believed to enhance
production and act as a permeability-assist component for the
Middle Bakken member. The effectiveness of this fracture
network is thought to impact productivity and decline
characteristics of the well.6
The Middle Bakken is an oil-wet reservoir and oil saturations
range from 75-90%. The sweet crude has an API of 44 degrees
with no H2S present. Water production is minimal and the gas
is extremely rich with an average BTU content of 1500.
Treatment Design
At the start of the project, several design modifications were
implemented which were designed to reduce completion costs
while maintaining current productivity. These modifications
included changes in treating fluid type and volume, proppant
type and volume, and treatment rate and proppant staging.
The primary change was an elimination of crosslinked fluids
in favor of lower-viscosity fluids. The initial fluid package

SPE 108045

consisted of friction-reduced water for those portions of the

treatment with low sand concentrations, and a 20 lb/Mgal
linear gel for the higher-concentration stages. There were
several reasons for this change. First, a reduction in the
treating fluid viscosity was intended to generate more fracture
length and create a larger contact area between the fracture
face and the reservoir. Second, it was anticipated that a
reduction in gel loading would reduce formation damage and
the corresponding loss in conductivity.
The overall fluid volume was drastically increased as well.
Typical crosslinked fracture treatments consisted of 3,0006,000 bbl of crosslinked fluid. The authors felt that this
volume, coupled with the high fluid viscosity, was not
sufficient to create the fracture/reservoir contact area required
for optimal production in this naturally fissured formation. A
typical low-viscosity treatment consists of 10,000-18,000 bbl
of fluid, depending on wellbore design and the ability to place
water tanks on location.
Several modifications were also made to the proppant
schedule.
Among these, the biggest change was the
elimination of 20/40 ceramic proppants in favor of 40/70 mesh
white sand. Without clear evidence of the fracture geometry,
it was decided to reduce proppant costs and place the less
expensive sand. Although the authors are aware that some
Bakken treatments have broken into offset wells over 2000
feet away in the transverse direction4, there is no conclusive
information on fracture geometry in this part of the field. In
the absence of data regarding fracture geometry, 40/70 mesh
white sand was used, which allowed a reduction in proppant
cost and thus overall job expense.
The maximum proppant concentration was also decreased.
Crosslinked jobs were designed for 300,000-700,000 lb of
lightweight ceramic proppant at a maximum proppant
concentration up to 6 ppg in order to create high-conductivity
fractures, with some 10 ppg stages for diversion. The newer
low-viscosity treatment designs use 300,000 to 500,000 lb of
sand at a maximum concentration of 2-5 ppg. As in previous
designs, the higher concentrations are used primarily for
diversion.

lateral. Friction-reduced water was used for the majority of

the job, and a 20 lb/Mgal linear gel was used for sand
concentrations greater than 1.5 ppg. These designs were fairly
conservative and were pumped at lower rates, generally
around 85 bpm. An example of this treatment design is given
in Figure 5.
Based on the pressure response during treatment for the jobs
pumped according to this general schedule, very little
diversion was seen. The desire for more diversion led to
further job modifications. The first of these was the
elimination of the linear gel stages. The friction-reduced
water exhibited no problems with proppant transport at these
rates and concentrations, and the elimination of linear gel
resulted in a reduction in overall job costs as well as a
reduction of polymer placed in the formation.
Additional changes included higher sand concentrations and
an elimination of sweeps between the majority of the sand
stages. Sand concentrations are stepped up gradually with no
sweeps in between each step. Each of these step-ramps
culminates in a brief stage of significantly higher
concentration, followed by a flush. This high-concentration
sand slug is intended to achieve diversion into other portions
of the lateral.
An increase in treating rate was also
implemented in these treatments. An example of the modified
design is shown in Figure 6.
Two refracture treatments have also pumped to date utilizing
the modified treatment design. These treatments were
performed in wells initially stimulated with a crosslinked
gel/ceramic proppant combination. These treatments were
smaller in total volume due to the shorter lateral length. For
the two refracture treatments the average job size was
approximately 7,200 bbl of fluid placed in a 3,400 lateral
section. Both laterals were completed with pre-perforated,
uncemented liners, and the possibility of using perforation
balls or bioballs to assist in achieving diversion was explored;
however, this was ultimately dismissed due to operational
concerns.

The third important modification was an increase in treatment

rate. Crosslinked gel jobs were pumped at an average rate of
59 bpm and a maximum rate of approximately 75 bpm. New
designs are pumped at an average rate of 103 bpm and achieve
maximum rates of up to 120 bpm. This higher rate is needed
not only for proppant transport, but to stimulate as much of the
lateral as possible. Some recent wellbores contain over 9,000
of openhole section, and with lower treating rates it is likely
that significant portions of this would go untreated.

Results
Nine wells examined for the purposes of this study were
stimulated using a crosslinked fluid with ceramic proppant.
One of these wellbores was a dual-lateral design, while the
remaining wellbores were single laterals. All of these fracture
treatments took place between April 2005 and June 2006.
Five wells were examined using the low-viscosity fracturing
design; one of these wells used lightweight ceramic proppant
due to availability, while the remaining wells used sand.
Three of these wellbores were dual-lateral designs. These
treatments took place between July and October 2006.

Two basic job designs have been utilized during this program.
The first used low-concentration sand slugs with sweep stages
between each sand-laden stage. These designs were used
early in the project to determine how the formation would
react to the higher treating rates and low-viscosity fluid. It
was also hoped that the sand slugs would help to achieve
diversion and to allow treatment of larger portions of the

Table 1 summarizes job costs and average post-frac

production rates for these wells. In cases of wellbores with
dual laterals, both the total job cost and production rates were
split to give a total cost and production per lateral. Wells
designated XL were treated with crosslinked fluids, while
wells designated LV were treated with low-viscosity fluid
systems. Production comparisons were based on 30-day

averages. While the use of 30-day averages does not give a

complete picture of ultimate recovery, in the absence of longterm production data, 30-day averages were used to facilitate
an equitable comparison between treatment groups. Figure 6
gives a comparison of 30-day production data for both groups.
Production data shows a 30% increase in initial average
production for the wells stimulated with the low-viscosity
fluids.
Average production per lateral increased from
approximately 128 BOPD to 170 BOPD with the new designs.
The best-performing low-viscosity well (LV2) exhibited
slightly better performance per lateral than the bestperforming crosslinked well (XL1) at a slight reduction in job
cost per lateral. It must be noted that the LV2 used
lightweight ceramic proppant instead of sand.
However, the small sample size coupled with the large scatter
in production data means that the production increase is within
the statistical variation of the data set. It is therefore difficult
to draw definite conclusions regarding the production increase
resulting from low-viscosity treatments.
Total average costs per lateral were approximately $410,000
for crosslinked treatments and $241,000 for the new treatment
designs. When the initial production is compared to the
treatment costs, the resulting treatment costs are
approximately $3,200 per initial BOPD for crosslinked fracs
and $1,400 per initial BOPD for low-viscosity treatments.
While this data too cannot be used to draw definite
conclusions, it is an encouraging sign that low-viscosity
treatments can achieve results comparable to crosslinked
treatments at a lower cost.
Refracture treatments have also exhibited good responses to
low-viscosity treatments. Table 2 compares production rates
following each stimulation treatment in the two wells
examined. The Post-Frac 1 average refers to the average
following the initial crosslinked fracture treatment. Pre-Frac
2 gives the 30-day average immediately prior to the
refracture treatment, while Post-Frac 2 is the 30-day average
following the low-viscosity refracture stimulation.
The XL1 refracture treatment resulted in a doubling of the prefrac production rate. The XL6 refracture treatment resulted in
a nearly five-fold production increase, and surpassed the initial
production rate of the well following the initial crosslinked
fracture treatment. The average total cost of these refracture
treatments was $165,000, compared with an average cost of
$408,500 for the initial crosslinked treatments.
Future Work
The optimization of treatment design in the Bakken is
ongoing. With the advent of longer openhole sections, a
means of consistently achieving stimulation along the entire
wellbore will likely become even more crucial.
An important factor in the quest to achieve diversion is the
necessity of understanding the fracture characteristics that are
currently being obtained.
Methods to determine these

SPE 108045

characteristics will likely include microseismic or tiltmeter

fracture mapping, the use of radioactive tracers, and the
application of production and temperature logging to identify
fluid entry points and fracture geometry.
Additional means of achieving diversion will be examined.
These could include mechanical means such as the use of
coiled tubing and other pinpoint stimulation methods.
Treatment designs could also be modified further in terms of
rate and maximum proppant concentration.
The use of white sand will continue to be examined. In the
absence of information regarding fracture geometry, it is
difficult to justify the use of high-conductivity proppants.
However, since the production from the only new well that
received ceramic proppant (LV2) is the best in the group,
further field tests using low-viscosity fluids in conjunction
with ceramic proppants may be warranted. Additional
evidence showing the creation of transverse hydraulic
fractures would also be a cause for further examining the use
of ceramic proppants.
As stimulation experience grows in the Bakken formation and
wellbore design becomes more standardized, an optimized
stimulation plan will be developed. The application of a
standardized fracture treatment allows for the effect of a single
variable to be measured.
Examples of future job
modifications include the elimination of sand to judge the
effectiveness of a waterfrac treatment and the use of
significantly greater fluid volumes to determine the correlation
between fluid pumped and production.
Conclusions
Based on initial production data, low-viscosity fracturing
treatments in the Middle Bakken have increased production
while significantly lowering stimulation costs. Based on 30day averages, a production increase of approximately 30% has
been observed in the wells that were treated with low-viscosity
fluids, while stimulation costs have decreased by 40%. Future
treatments will focus on increasing hydraulic fracture
coverage along the wellbore, optimizing treatment volumes
and pump rates, understanding fracture geometry, and
optimizing proppant selection.
Acknowledgements
The authors would like to thank Hess Corporation and
Pinnacle Technologies for permission to publish this paper.
Unit Conversions
API
141.5/ (131.5 + API)
bbl
*
1.589 873
E - 01
ft
*
3.048
E - 01
F
(F 32)/1.8
gal
*
3.785 412
E - 03
lbm
*
4.535 924
E - 01
md
*
9.869 233
E - 04
psi
*
6.894 757
E +00

= g/cm3
= m3
=m
= C
= m3
= kg
= m2
= kPa

SPE 108045

References
1.

2.

3.
4.

5.

6.

Breit, V., Stright Jr., D., Dozzo, J: Reservoir

Characterization of the Bakken Shale From Modeling
of Horizontal Well Production Interference Data,
paper SPE 24320 presented at the 1992 SPE Rocky
Mountain Regional Meeting, Casper, WY., May 1821.
Cramer, D.:
Reservoir Characteristics and
Stimulation Techniques in the Bakken Formation and
Adjacent Beds, Billings Nose Area, Williston Basin,
paper SPE 15166 presented at the 1986 SPE Rocky
Mountain Regional Meeting, Billings, MT., May 1921.
Cramer, D.: Treating-Pressure Analysis in the
Bakken Formation, JPT (Jan. 1992) 20-27.
Wiley, C., Barree, B., Eberhard, M., and Lantz, T.:
Improved Horizontal Well Stimulations in the
Bakken Formation, Williston Basin, Montana, paper
SPE 90697 presented at the 2004 SPE Annual
Technical Conference and Exhibition, Houston, Sept.
26-29.
Mayerhofer, M., Richardson, M., Walker, R.,
Meehan, D., Oehler, M., and Browning, R.:
Proppants? We Dont Need No Proppants, paper
SPE 38611 presented at the 1997 Annual Technical
Conference and Exhibiton, San Antonio, Oct. 5-8.
Reisz, M.: Reservoir Evaluation of Horizontal
Bakken Well Performance on the Southwestern Flank
of the Williston Basin, paper SPE 22389 presented
at the 1992 SPE International Meeting on Petroleum
Engineering, Beijing, Mar. 24-27.

SPE 108045

Figure 1 Bakken Development Areas

SPE 108045

Figure 3 Bakken Geologic Type Log

SPE 108045

Middle Bakken Fracture Treatment

Early Low-Viscosity Design
Slurry Flow Rate (bpm)
Surf Press [Csg] (psi)

150.0
7500

Proppant Conc (ppg)

5.000

120.0
6000

4.000

90.0
4500

3.000

60.0
3000

2.000

30.0
1500

1.000

0.0
0

0.0

40.0

80.0

120.0

160.0

200.0

0.000

Time (min)

Figure 4 Early Low-Viscosity Hydraulic Fracture Treatment Design

Middle Bakken Fracture Treatment

Modified Low-Viscosity Design
Blender 1 Slurry Rate (bpm)
Proppant Conc (ppg)

100.0
10.00

Blender 2 Slurry Rate (bpm)

Surf Press [Csg] (psi)

100.0
7500

80.0
8.00

80.0
6000

60.0
6.00

60.0
4500

40.0
4.00

40.0
3000

20.0
2.00

20.0
1500

0.0
0.00

0.0

36.0

72.0

108.0

144.0

Time (min)

Figure 5 Modified Low-Viscosity Hydraulic Fracture Treatment Design

180.0

0.0
0

SPE 108045

Figure 6 30-Day Production Data Comparison

Table 1 Treatment Results

Well
Name

Number of
Laterals

XL1
XL2
XL3
XL4
XL5
XL6
XL7
XL8
XL9

1
1
1
2
1
1
1
1
1

Fluid Pumped
per Lateral
(bbl)
3,700
3,692
5,234
3,102
5,585
6,323
5,714
5,315
2,726

LV1
LV2
LV3
LV4
LV5

1
2
2
2
1

12,994
7,076
8,040
8,517
10,648

Proppant Pumped
per Lateral (lb)

Lateral
Length (ft)

Proppant

Treatment Cost
per Lateral

301,252
306,563
518,460
344,253
516,962
364,980
525,586
514,275
240,318

3109
3158
5034
3648
5639
3768
5392
5266
4640

20/40 LWC
20/40 LWC
20/40 LWC
20/40 LWC
20/40 LWC
20/40 LWC
20/40 LWC
20/40 LWC
20/40 LWC

$328,000
$347,000
$474,000
$293,500
$496,000
$489,000
$518,000
$505,000
$355,000

Initial 30-Day Average-Post-Frac Production per

Lateral (BOPD)
343
117
36
194.5
48
152
75
80
38

299,803
215,750
264,806
234,866
259,287

5266
4997.5
4779.5
3903
6100

40/70 white
30/50 LWC
40/70 white
40/70 white
40/70 white

$292,000
$320,500
$150,000
$197,000
$301,000

39
370
184
79
54

Wells designated as XL were treated with a crosslinked fluid.

Wells designated as LV were treated with a low-viscosity fluid.
LWC: Lightweight Ceramic

Table 2 Refracture Treatment Results

Well Name
XL1
XL6

30-Day Average (Post-Frac 1)

343
152

30-Day Average (Pre-Frac 2)

74
35

30-Day Average (Post-Frac 2)

147
173