Beyond gamma ray: Determining a geosteering program

for a new unconventional reservoir

Jason Pitcher 1 and Frederic Gallice 2

The Leading Edge 2015.34:514-522.

Downloaded from library.seg.org by Piyush Pasari on 05/31/15. For personal use only.

Abstract
Geosteering programs used in unconventional reservoirs
typically are restricted to using basic natural gamma-ray tools.
Although that approach might be eective, it does not mean
that all unconventional reservoirs are amenable to the same
approach. A reservoir in the South Central Oklahoma Oil
Province is examined to determine what kind of geosteering
program can add value, with tool-selection criteria being developed in consideration of the reservoir character. Analysis
of the usefulness of gamma-ray, resistivity, neutron, density
and sonic tools can result in notes on operational eectiveness and impact in terms of planning a geosteering program.
Pilot-log petrophysical models can be created and examined
for dierent tool congurations and styles to assist in program design. Examples illustrate the variability in tool design
and impact on planning. The results demonstrate that an integrated approach to program design and tool selection that
involves geologic, petrophysical, and geomechanical considerations can determine a methodology that meets operational
needs. Additional considerations pertain to selection of tools
available and their operational value in horizontal unconventional well development.

Introduction

Alternate geosteering methods can be considered by modeling the theoretical output from the logging-while-drilling
(LWD) tools (Calleja et al., 2010). For nonazimuthal tools
with a depth of investigation that is limited to the wellbore
wall or close to it, such as gamma ray, density, and porosity,
a simple total-stratigraphic-thickness (TST) pseudolog or
stretch log can be projected along the well path in accordance
with the relevant structural position correlated to the oset
logs (Jackson et al., 1998). With the advent of azimuthal circuitry, density (Ballay et al., 2001), gamma ray (Pitcher et al.,
2009b) and sonic (Market and Bilby, 2011) real-time loggingwhile-drilling wellbore images are also potential options that
can be modeled.
Beyond the immediate vicinity of the wellbore, LWD options are more limited. Omnidirectional propagation-wave resistivity tools are readily available from many vendors and can
be modeled using nonvendor-specic software (Zhou, 2008).
Azimuthal electromagnetic propagation-resistivity sensors have
been available since 2005 (Li et al., 2005), with fewer providers
typically oering integrated modeling and geosteering services.
These tools also can be modeled using commercially available
software, although the full capability of these tools often is restricted to the provider. In the unconventional-resource play
market, these tools are seen as cost prohibitive and dicult to
use eectively.

In the South Central Oklahoma Oil Province (SCOOP),

there are several promising intervals within the Springer
Shale. The current work is to determine an ecient approach Target interval geology
to wellbore placement that meets operational and productivThe reservoir target is within the Mississippian Chesterian
ity needs. The constraints are the result of the geology of the series of the Anadarko Basin (Figure 2). At the study location,
reservoir, which consists of thinly layered mudstones. Because this interval is thick, as much as 3000 ft, mostly composed
of the depositional environment, there is little gamma-ray con- of shale at the bottom and grading into sandstone at the top,
trast between the target interval and the bounding formations representing a typical transgressive-regressive (T-R) cycle in
(Figure 1).
a marine setting. The expanded section is the consequence of
As can be seen from Figure 1,
there is also considerable lateral heterogeneity in the system. The problems from a drilling perspective
are complicated by geomechanical
instability in the bounding formations, with time-determinant wellbore collapse being a major concern
in case a horizontal wellbore exits the
target formation. Given these formation characteristics, geosteering
systems that commonly are used in
unconventional plays that use pattern
matching against oset logs (Pitcher
and Jackson, 2012) are considered inappropriate as a geosteering solution
Figure 1. Oset vertical logs: gamma ray, resistivity, neutron/density.
in this target interval.
1
2

514

Cameron.
Vitruvian Exploration II, LLC.

THE LEADING EDGE

May 2015

http://dx.doi.org/10.1190/tle34050514.1.

Special Section: Well geosteering

The Leading Edge 2015.34:514-522.

Downloaded from library.seg.org by Piyush Pasari on 05/31/15. For personal use only.

tectonic activity resulting from the far-eld loading of a thrust

sheet south of the study area (Feinstein, 1981).
This created an accommodation space which lled with clays
and silt relatively quickly compared with the Woodford target
below. This is exemplied by the high water retention in the
shale, reecting the time of dewatering during deposition and
the abundance of soft sediment features from high-resolution
wireline image logs which reect fast deposition on the slope.
During the early stage of this deposition while shale was still
being deposited, several cycles of rich organic-matter sections

also were deposited and preserved. These biologic events were

recorded by a high percentage of biogenic silica associated with
phosphorus and organic-rich material.
One of these cycles is the reservoir target, bounded by
weak and reactive shale. This unconventional reservoir was
bypassed on the way down to the Woodford reservoir, mostly
because of the lack of the characteristic high gamma-ray signature typical of a source-rock reservoir play. The increase in
total organic carbon (TOC) is reected by an increase in uranium concentration, but because of the increase in silica at the
expense of the background clay, both thorium and potassium
concentrations decrease. This results in a total gamma-ray log
response that is relatively constant and atypical for sourcerock reservoirs.
The lack of a carbonate section above and below is not common
for source-rock reservoirs currently being worked in the United
States and makes the landing and geosteering of the well a delicate operation. In addition, the horizontal and vertical variability
at the reservoir level seen in Figures 1 and 3 results in a need for
careful geosteering-tool selection while drilling these wells.

Offset well analysis

A critical part of the planning process for geosteering analysis is the determination of useful data from the oset wireline
logs (Figure 1) and the impact the data will have on the selection of horizontal-well geosteering tools. As mentioned above,
the conventional gamma-ray response does not delineate the
target formation with a characteristic denitive response. From
these basic logs, resistivity and porosity give the best indicators
of the target-interval boundaries. Additional data from spectral
gamma measurements (Figure 4) show a clear, unambiguous response in the relationship between uranium and thorium, delineating the top and base of the section of interest with a crossover
eect on the logs.
Figure 5 shows the relative response of an acoustic tool in
the section of interest. VP /VS is a potential indicator of being in
the target zone and has the advantage of allowing the determination of relative competence.
Examination of the resistivity logs shows that the 4-ft
resolution-array induction response (AIF90 in Figure 1), alFigure 2. Stratigraphy of target zone. After Cipriani (1963).
though providing a good average resistivity value, misses a great deal of
detail in the target interval. The other
log responses and core photos indicate
much more layering in the formation.
The highest-resolution resistivity measurement available from the resistivity-suite run (1-ft-resolution AIO90)
is plotted. With the close agreement
between curves, the highest-resolution curve can be used as the input for
resistivity modeling.
A major challenge in using propagation-resistivity tools for geosteering
is that resistivity tools do not respond
in the same manner in horizontal wells
Figure 3. CT scan of a 3-ft cored interval showing a thinly layered mudstone.
as they do in vertical wells when the

Special Section: Well geosteering

May 2015

THE LEADING EDGE

515

formations are heterogeneous (Bittar and Rodney, 1994) because

of anisotropy and polarization eects. For resistivity-model input, the most accurate representation of the layers within the
target zone is essential for understanding the impact a resistivity
tool can have on the geosteering solution.
For forward modeling of a multilayered system, the technique used here is to generate a square log based on input parameters from the most detailed resistivity log available (Figure 6).
This output is the basis for the resistivity-tool response modeling
in the horizontal well.

Figure 4. Spectral gamma-ray response.

Maxwell Dynamics, Inc., was used. The technique is to use a

simple at geologic model based on a single oset well, with a
planned horizontal well that traverses the geology from above
the top of the target interval to below the base (Figure 7). For
ease of analysis, the planned well then can be plotted to traverse
back up through the section (Pitcher et al., 2009a). This is important if asymmetric tool response is anticipated. The well trajectory is kept constant, with a 2 angle of incidence to the bed
boundaries. This is done to simplify the model and gain an understanding of tool response in a simple geometric relationship.
With a geologic and well-trajectory model developed, modeled tool responses can be plotted along a theoretical well projected as a curtain plot. The tools with limited depth of investigation
gamma ray, density, and porosity are simply projected to the
well at the appropriate true-vertical-depth (TVD) location. Because gamma-ray and density imaging tools are common, the images for these tools are simple to model,
as shown in Figure 7.
For the resistivity model, the
modied square log is used as the
input. In the scenario presented, the
individual layers are considered isotropic. The model was run for a selection of common tool congurations
and geometries. Presented here is the
resistivity model for a compensated
electromagnetic-wave resistivity tool
consisting of two frequencies (2 MHz

Figure 5. Sonic response in target formation.

Figure 6. Oset log with square-log curve t.

516

Special Section: Well geosteering

The Leading Edge 2015.34:514-522.

Downloaded from library.seg.org by Piyush Pasari on 05/31/15. For personal use only.

Tool response modeling

To understand the value of any specic tool or suite of tools
used in the geosteering program, a model is built to evaluate the
system. To allow for analysis of multiple vendor tool types, a
commercial geosteering and tool-modeling software suite from

THE LEADING EDGE

May 2015

and 400 kHz) and dual spacing (36 inches and 22 inches). Four
curves are presented on the model the 36-inch 400-kHz
attenuation resistivity, 36-inch 400-kHz phase resistivity,
36-inch 2-MHz phase resistivity, and 22-inch 2-MHz phase
resistivity. These curves represent deep, medium, and shallow
depths of detection, used for visual interpretation in geosteering the horizontal well.

Figure 8. Modeled tool responses at base of target section.

geologic model, the real-time resistivity curves should match

the model curves.
This methodology begins to break down when the petrophysical model deviates from the original because of lateral
facies changes or unresolved characteristics in the input data.
When that breakdown occurs, visual interpretation based on
relative curve responses is often the most eective method of
reaching a steering decision in real time.
An example of this is where the three depths of detection
overlie one another, typically indicating that the tool response
is not being aected by boundary eects. A boundary is detected rst by the deepest detecting response, which will begin
to separate from the other curves, either high or low, depending on conditions. An example of this (in reverse, detecting
the boundary from the low-resistivity formation) can be seen
in Figure 7, reading the section from right to left from 3200
ft. The distance to the boundary can be determined using integrated radial pseudogeometric factor curves (Pitcher, 2004) or
inversion methods.
A dierent response might be caused by a second boundary
superimposing polarization on the primary response. This might
be the case here, at a depth of 2960 ft, where phase curves are
stacking and then the medium curve separates from the shallow
and deep curves. Note that in this case, the deep attenuation
resistivity curve rarely matches the deep phase curve, making
visual comparison problematic. In the modeled section (Figure
7), these superimposition eects are present throughout the section, in the zone of interest, based on the modeled geology, well
path, and petrophysical inputs.
The challenge in real-time geosteering will be to produce an eective
comparative model that matches realtime data in a timely manner to eect
steering decisions. From the standpoint of petrophysical evaluation, the
extent of the boundary eects might
also make a determination of Rt to be
challenging along the horizontal well
path. Software tools have to be used to
compare model responses against measured data, which requires additional
processing time and might not provide
a unique solution set.
The resistivity modeling performed in this work demonstrates
that simple omnidirectional resistivity tools can detect boundaries in this
reservoir eectively. The challenge is
that the resistivity boundaries do not
necessarily correspond to the zone
boundaries (labeled as top and base
in Figure 1). This is seen in pilot well
C, with multiple internal resistivity
boundaries in the main zone of interest. This is compounded by additional
layers that are not incorporated in the
model because they are unresolved in

518

Special Section: Well geosteering

The Leading Edge 2015.34:514-522.

Downloaded from library.seg.org by Piyush Pasari on 05/31/15. For personal use only.

Results
Figure 7 shows the completed model with tool responses.
The modeled tool response for the gamma-ray and density-image logs are shown. Throughout the section, resistivity responses
show typical responses expected from this type of tool. The deep
attenuation resistivity model has few polarization artifacts and
does not match the oset resistivity (Rh) in terms of magnitude.
The phase responses more closely match the oset resistivity in
magnitude but exhibit more polarization eects associated with
internal boundaries in the zone of interest. Figure 8 shows a
detailed part of the section, depicting the theoretical exit from
the base of the zone of interest.

Discussion
The most critical areas in geosteering this section will be
approaching the top and base of the zone of interest. When
evaluating this type of resistivity tool and its capacity to detect
the boundary in a conventional reservoir, the primary method
of determining well placement is to compare the real-time
response against the modeled response. If the petrophysical
model is accurate and the well is positioned correctly in the

Figure 7. Geosteering model section.

THE LEADING EDGE

May 2015

The Leading Edge 2015.34:514-522.

Downloaded from library.seg.org by Piyush Pasari on 05/31/15. For personal use only.

the oset well. These are known to be present from the core
and will prevent the model from matching the measured responses until the input parameters to model are adjusted.
There are some assumptions in the modeling that must
be understood before building a geosteering plan based on
resistivity or any single tool. The input model does not have
high resolution and might not be an accurate representation
of the geology. This point is also important when considering
the other tools modeled. A thin (< 6 inches) but highly radioactive layer would have only a small response on the vertical
log, but with a tool penetrating the layer at a high angle, the
thin-bed response would have a pronounced response on a
gamma-ray log.
This issue of thin-layer properties can become signicant
when trying to interpret horizontal real-time logs for well
placement because the curves might represent formation properties dierent from those exemplied in vertical well logs.
Other parameter assumptions are that the individual layers are electrically isotropic. Even if this assumption is correct, there might be sharp, high-contrast thin layers in the
formation that are not resolved on oset logs but have a cumulative impact on the horizontal resistivity response, making the formation seem anisotropic even though individual
layers are isotropic.
As seen in Figures 7 and 8, images do aid interpretation
of relative geometry between the wellbore and the formation.
Relative dip picking from gamma-ray images is notoriously
unreliable because the depth of investigation for a gamma tool
is subject to many factors beyond the control of the interpreter,
but as an indicator of relative trajectory, the value is clear. Density images, with higher resolution, can be used for relative
dip picking, which aids geosteering, reducing the tortuosity
induced by pattern matching (Pitcher and Jackson, 2012).
As seen from pilot wells A, B, and C (Figure 1) and from
the projected horizontal well in Figure 7, there are two methods
to reliably determine whether the well is out of zone. These are
spectral gamma and porosity. As mentioned above, the resistivity responses do not correspond to the zone boundaries in the
area of interest. This is important to understand because of unanticipated faulting and the sensitivity of the bounding formations for collapse.
Because of the resistivity response of the bounding formations close to the boundary, resistivity on its own cannot be
used to reliably determine whether the well has been faulted
out, and gamma ray is similarly inconclusive. A key challenge would be to determine the sense and magnitude of the
fault. This can be done by steering to build a prole-to-pattern
match, and time-sensitive boundary formations make this operationally challenging.
An additional measurement for consideration is the use of
acoustic tools. Figure 5 shows the relative response of an acoustic tool in the section of interest. Using sonic as a geosteering
tool has been documented (Pitcher et al., 2011), but it is still
uncommon because of cost and operational understanding of
the application.

Conclusion

520

Special Section: Well geosteering

THE LEADING EDGE

May 2015

Some key questions need to be answered when planning to

geosteer a well in a new reservoir.

What tool or suite of tools will provide the maximum utility

in well placement?
Will the tool suite be able to determine whether the well is
in the target zone?
Will the tool suite enable the stratigraphic placement of the
well?

Modeling tool responses is an eective method of testing

assertions as to the value of any particular tool and allows for
consideration of a multitool approach. In this reservoir, the typical approach to a source-rock reservoir of using the lowest-cost
tool and a basic periodic look-back approach will not answer the
above questions and increases risk to the integrity of the well
and drilling process. Adding a resistivity tool does improve the
potential for success, but the limitations imposed by the physics
of the measurement have to be understood and accounted for.
To do this in real time while drilling is possible, but a straightforward approach of modeling, measuring, and adjusting will
struggle to compensate for the limitations of the model in a realtime drilling environment.
In this reservoir, use of an imaging tool to determine wellbore/stratigraphy relationships in combination with real-time
resistivity modeling and interpretation should suce in preventing the well from being drilled out of the zone of interest. In
case of a fault intersection, depending on the sense and degree
of throw, a real-time spectral gamma, neutron-porosity, or sonic
tool will be needed to determine whether the well has moved out
of zone, based on the oset-well log responses. The challenge
still would be to determine whether the well is above or below
the reservoir, which none of the presented tools could determine
from a single point sample in case the well has moved beyond
the range of detection for deep-reading sensors. There are methods to determine absolute position from a single point, but that
is beyond the scope of this study.

Acknowledgments
The authors extend their thanks to John Zhou of Maxwell
Dynamics for his time and eorts in introducing us to the software used to model the tool responses in this article. We also
thank Vitruvian Exploration II, LLC, for permission to publish
this article.
Corresponding author: Jason.Pitcher@c-a-m.com

References
Ballay, G., H. Al-Ali, S. Amos, and B. Dennis, 2001, In the drivers seat with LWD azimuthal density images: Presented at the
SPE/IADC Middle East Drilling Technology Conference, paper SPE 72282-MS, http://dx.doi.org/10.2118/72282-MS.
Bittar, M., and P. Rodney, 1994, The eect of rock anisotropy on
MWD electromagnetic wave resistivity sensor: 35th Annual
Logging Symposium, SPWLA, paper PP.

The Leading Edge 2015.34:514-522.

Downloaded from library.seg.org by Piyush Pasari on 05/31/15. For personal use only.

Calleja, B., J. Market, J. Pitcher, and C. Bilby, 2010, Multi-sensor

geosteering: Presented at the SPWLA Annual Symposium,
paper 82670.
Cipriani, D., 1963 General geologic sections of Oklahoma and
northern Arkansas, http://www.ogs.ou.edu/fossilfuels/pdf/
Geol_Sections.pdf, accessed 20 January 2015.
Feinstein, S., 1981 Subsidence and thermal history of southern
Oklahoma aulacogen: Implications for petroleum exploration:
AAPG Bulletin, 65, no. 12, 25212533.
Jackson, C. E., H. M. J. Illfelder, and G. Pineda, 1998. Field
implementation of geological steering techniques optimizes drilling in highly-deviated and horizontal wells: Presented at the SPE International Conference on Horizontal Well Technology, paper SPE 50381-MS, http://dx.doi.
org/10.2118/50381-MS.
Li, Q., D. Omeragic, L. Chou, L. Yang, and K. Duong, 2005, New
directional electromagnetic tool for proactive geosteering and
accurate formation evaluation while drilling: Presented at the
46th Annual Logging Symposium, SPWLA, paper SPWLA2005-UU.
Market, J., and C. Bilby, 2011, Introducing the rst LWD
crossed-dipole sonic imaging service: Presented at the 52nd

Annual Logging Symposium, SPWLA, paper SPWLA2011-DDD.

Pitcher, J., 2004, Interpreting resistivity responses while geosteering horizontal wells: Oral presentation at the Topical Conference on Formation Evaluation in Horizontal and High Angle
Wells, SPWLA.
Pitcher, J., and T. Jackson, 2012, Geosteering in unconventional
shales: Current practice and developing methodologies. Presented at the SPE/EAGE European Unconventional Resources
Conference and Exhibition, paper SPE 152580-MS, http://
dx.doi.org/10.2118/152580-MS.
Pitcher, J. L., J. Market, and D. Hinz, 2011, Geosteering with sonic
in conventional and unconventional reservoirs: Presented at the
Annual Technical Conference and Exhibition, SPE, paper SPE
146732, http://dx.doi.org/10.2118/146732-MS.
Pitcher, J., D. B. Schaefer, and P. Botterell, 2009b, A new azimuthal gamma at bit imaging tool for geosteering thin reservoirs:
Presented at the SPE/IADC Drilling Conference, paper SPE/
IADC 118328P, http://dx.doi.org/10.2118/118328-MS.
Zhou, Q., 2008, Log interpretation in high-deviation wells through
user-friendly tool-response processing: Presented at the Annual
Symposium, SPWLA, paper AAAA.

522

Special Section: Well geosteering

THE LEADING EDGE

May 2015