Distance Metric Based Multi-Attribute Seismic Facies Classification to Identify Sweet Spots within the Barnett shale : A Case study from the Fort Worth Basin, TX

URTeC Control ID 1619856

1

URTeC Control ID Number: 1619856

Distance Metric Based Multi-Attribute Seismic Facies Classification to
Identify Sweet Spots within the Barnett shale: A Case Study from the
Fort Worth Basin, TX
Atish Roy, University of Oklahoma
Vikram Jayaram, Oklahoma Geological Survey
Kurt Marfurt, University of Oklahoma
Copyright 2013, Unconventional Resources Technology Conference (URTeC)
This paper was prepared for presentation at the Unconventional Resources Technology Conference held in Denver, Colorado, USA, 12-14 August 2013.
The URTeC Technical Program Committee accepted this presentation on the basis of information contained in an abstract submitted by the author(s). The contents of this paper
have not been reviewed by URTeC and URTeC does not warrant the accuracy, reliability, or timeliness of any information herein. All information is the responsibility of, and, is
subject to corrections by the author(s). Any person or entity that relies on any information obtained from this paper does so at their own risk. The information herein does not
necessarily reflect any position of URTeC. Any reproduction, distribution, or storage of any part of this paper without the written consent of URTeC is prohibited.

Summary
Conventional reservoirs benefit from a long scientific history that correlates successful plays to seismic
measurements through depositional, tectonic, and digenetic models. Unconventional reservoirs are less well
understood, however benefit from significantly denser well control. Thus, allowing us to establish statistical rather
than model-based correlations between seismic data, geology, and successful completion strategies. One of the more
commonly encountered correlation techniques is based on computer assisted pattern recognition. The pattern
recognition techniques have found their niche in a plethora of applications ranging from flagging suspicious credit
card purchase patterns to rewarding repeating online buying patterns. Classification of a given seismic response as
having a “good” or “bad” pattern requires a “distance metric”. Distance metric “learning” uses past experiences
(well performance) as training data to develop a distance metric. Alternative distance metrics have demonstrated
significant value in the identification and classification of repeated or anomalous behaviors in public health,
security, and marketing. In this paper we examine the value of three of these alternative distance metrics of 3D
seismic attributes to the identification of sweet spots in a Barnett Shale play.
Introduction
Similarity of waveforms or attribute patterns along a horizon to those about good and bad wells is a well-established
practice in seismic data interpretation. In such studies, typically, the interpreter compares a vector of samples (e.g.
Johnson, 2000) or attributes (e.g. Michelena et al., 1998) extracted from productive or non-productive wells to every
trace along the horizon. Poupon et al. (1999) correlated wells to seismic waveforms, where the supervision was not
only the actual seismic about the well but also a suite of synthetic seismic traces generated through petrophysical
modeling and fluid substitution. The performance of these algorithms can depend delicately on the manner in which
distances are measured. Distance metrics are a vital component in many applications ranging from supervised
learning and clustering to product recommendations and document browsing. Since, designing such metrics by hand
is difficult, we look at the problem of learning a metric from exemplars. In particular, we consider relative and
qualitative exemplars of the form “P is closer to Q than P is to R”. In essence, a distance metric is a mathematical
operator that conveys how similar two (possibly vector-valued) members of a set compared with a single, scalar
value, based on a notion of similarity. The most common and most easily understood waveform similarity metric is
based on Euclidian distance.
In this paper, we discuss three different schemes of distance measures to classify 3D seismic facies: the minimum
Euclidean distance (MED), spectral angle mapper (SAM) and the Bhattacharya distance, to quantify similarity to a
given suite of wells. The use of such well control makes this a “supervised” learning task, with each seismic voxel

2


being clustered with the most similar well. Depending on our distance scalar, we may choose to leave some voxels
as “unclassified”.
We begin our analysis by providing brief formulations of the three distance metric methods utilized in the paper. We
then summarize the training step and the generation of seismic facies volumes. Lastly, we discuss how each distance
metric uniquely identifies and maps facies for sweet spot detection within the Barnett shale.
Minimum Euclidean Distance
Throughout this paper we have compared two types of vectors. One is the attribute data-vector Pn, which are the
normalized attribute values at each voxel location n. The other is the well data-vector Qw, which is the average of
the attribute data-vector around each well location w. In a Cartesian coordinate system, the Euclidean distance
between vectors Pn and Qw is described as the length of the distance connecting the two vectors. This distance metric
is valid and can be extended to an N-dimensional space, where n is the number of seismic attributes at any voxel.
The formulation of the distance between Pn and Qw for a single well location w is given by
√∑
where,

is the vector at

|

| ,……………………. (1)
is the vector at wth well.

voxel and

Also, to be noted that the distances such as the Euclidean distance, have some well-known properties such as




d(P, Q)  0 for all p and q and d(P, Q) = 0 only if P=Q (Positive definiteness)
d(P, Q) = d(Q, P) for all p and q (Symmetry)
d(P, R)  d(P, Q) + d(Q, R) for all points P, Q, and R (Triangle Inequality),

where, d(P, Q) is the distance (dissimilarity) between points (data objects), P and Q are vectors. A distance that
satisfies these properties is a metric, and such a space is called a metric space.
The Euclidean distance is a special case of Mahalanobis distance measure, which is formulated as
d Mahalanobi (P, Q)  (P  Q) 1 (P  Q)T . When the covariance matrix, Σ, is identity matrix, the Mahalanobis distance
s
is the same as the Euclidean distance.
Spectral Angle Mapper

The mathematical formulation of SAM attempts to obtain the angles formed between the reference well data-vector
Qw, and the attribute data-vector Pn, treating them as vectors in a space with dimensionality equal to the number of
attributes (Kruse et al., 1992; Boardman, 1992). SAM presents the following formulation:
∑
√∑

√∑

……………………. (2)

where, α is the angle formed between the well data vector and the attribute data vector. The expression is simply the
dot product of vectors P and Q. The SAM value is expressed in degrees where α is the angle subtended by the two
vectors P and Q, higher the angle means better the separation between the two vectors. The angle α, determined by
, presents a variation anywhere between 0 degree and 90 degree. Equation 2 can also be expressed as
∑
√∑

√∑

where, the best estimate acquires values close to 1.

……………………. (3)

3


The Bhattacharya Distance
In this metric, we use a probability distribution functions to measure similarities or dissimilarities between two
PDFs. Initially we define a 2D latent space grid with uniformly defined control points K and we project the posterior
probabilities of the vectors onto the grid points k=1,2,….,K. Let Rwk be the posterior probability distribution
corresponding to the wth well data-vector Qw, at kth grid location of the latent space (Figure 1a). Let Rnk be the
posterior probability projected at kth grid point, of any nth voxel seismic attribute data-vector Pn (Figure 1b). Thus
each vector Qw and Pn forms a PDF in the defined 2D latent space. Then by Bhattacharya (1943) measure, we can
find the similarities (Figure 1c) between the two PDFs by
∑

√

……………………. (4)

where, k are the grid points of the 2D latent space. Thus when two distributions are identical (Rnk = Rwk), we have a
∑
coefficient of
. In contrast when there is no overlap between the PDFs
. Thus, this
coefficient ranges from
.
For generating the PDF similarity seismic volumes we follow the workflow proposed by Roy (2013) who uses
Generative Topographic Mapping (GTM) (Bishop, 1998) to generate a probability density functions (PDFs) of the
well vectors and the attribute data-vectors. The GTM is a non-linear dimensional reduction technique that provides a
probabilistic representation of the data-vectors in a corresponding lower dimensional latent space.

Figure 1: A schematic representation of the supervised GTM analysis workflow: (a) The posterior probability of the well data vector (PDF)
projected onto the latent space. (b) The PDF of a data vector corresponding to voxel n in the seismic attribute volume projected onto the latent
space. (c) The joint PDF of the average well data vector and the attribute data vector. The coefficient dn from the Bhattacharya measure is the
measure of the similarity (overlap) between the two PDFs.

A set of non-linear continuous and differentiable basis functions is used to map points in the L-dimensional latent
space into a lower dimensional non-Euclidean manifold S embedded within the D-dimensional data-space. Data
vectors are modeled by a suite of Gaussian PDFs centered on the mapped grid points onto the non-Euclidean
manifold S. This manifold defines the space in which the data vector lies. Each mapped grid point on this 2D
manifold holds a finite probability that they represent a data-vector. For the purpose of visualization these
probabilities are projected as posterior probabilities back onto the 2D grid space, using Bayes theorem. Initially each
target well data-vector forms a PDF in the 2D latent space (Figure 1a). These PDFs are compared with the PDFs of
the data-vectors (Figure 1b). In this manner, we compute the overlap value of d n for all the data-vectors in the survey


4

(Figure 1c) resulting in a supervised facies “similarity” volume, quantitatively comparing each voxel to good and
poor wells (Roy, 2013). Thus, we calculate the overlap of the posterior probability distribution of each of these well
data-vectors and the remaining data-vectors by calculating the coefficient from the Bhattacharya (1943) measure
(equation 1). Separate “similarity” volumes for each shale petrotype are generated.

Barnett Shale from the Fort Worth Basin
The Fort Worth Basin is a foreland basin, located in north-central Texas and is associated with the late Paleozoic
Ouachita orogeny. This basin is bounded by the Muenster Arch to the northeast, the Ouachita Thrust Front to the
east, the Bend Arch to the west, the Red River Arch to the north, and the Llano Uplift to the south (Figure 2a). In
this study area the Barnett sits on an angular unconformity above the Cambrian to upper-Ordovician-age carbonates
of the Ellenberger Group and Viola Formation and overlying Pennsylvanian-age Marble Falls Limestone (Figure
2b). In between, the Forestburg Limestone divides the Barnett formation into Upper and Lower Barnett zone (Figure
2c). The Barnett Shale is not homogeneous, but rather can be subdivided into siliceous shale, argillaceous shale,
calcareous shale, and limestone layers, with minor amounts of dolomite (Singh, 2008).
The Fort Worth Barnett shale gas play is traditionally more of an engineering driven play. It requires hydraulic
fracturing for gas production. Our 3D seismic survey (Figure 3a) consists of a 200 square mile survey in the North
East Fort Worth Basin. The data are sampled at 110ft by 110ft by 2ms. However, the above survey was acquired
after numerous vertical and horizontal wells have been drilled and hydraulically fractured. For effective well
placement within the survey in future drilling, care should be taken to identify the brittle zones by mapping the
geomechanical rock type of the Barnett shale. The dataset between the Marble Faults (overlaying the Barnett
formation) and the Viola limestone (below the Barnett formation) are considered for attribute analysis.
Input Seismic Attributes for Seismic Facies classification
The inputs to our algorithms are different seismic inversion volumes (P-impedance, lambda-rho, mu-rho), which
help in understanding the highly heterogeneous nature of the Barnett shale. For the above attribute generations the
seismic data between the Marble faults horizon and the Viola limestone is considered. The impedance volumes
better reflect a heterogeneous shale reservoir based rock type. The Lamé parameters of seismic inversion such as
lambda-rho (λρ) and mu-rho (μρ) correlate to "fracability" and different elastic properties of rocks. Simple cross-plot
between two such elastic properties from the wells sometimes helps in segregating different rock types. However, it
is very difficult to separate between classes when we cross-plot any of these two seismic volumes.
We pre-condition our input data so that all the three input attributes are associated with each of the x, y, z location of
the survey (we call them as attribute data-vectors). Also the dataset is normalized to remove any bias to one of the
input attributes.
Study of the different Petrotypes in Barnett shale
The petrotype analysis for Barnett Shale is done for three different zones with their boundaries defined by the
sequences in the gamma-ray log from Figure 3. The Barnett shale is not homogeneous and based on the
parasequences, several lithofacies have been defined within the Barnett shale formation (Singh et al, 2008).
Analyzing various well logs from the survey, we have considered only three different lithofacies or shale petrotypes
: shale petrotype A represents the Upper Barnett Shale, shale petrotype B represents the brittle upper section of
Lower Barnett shale, and shale petrotype C represents the ductile, TOC rich lower section of Lower Barnett
shale. Zone A is defined within the upper Barnett shale shown in Figure 3c. An average attribute data-vector is
extracted around well A (Figure 3a and b) from this zone. We call this as shale petrotype A. Petrotype A (well data
vector) shows a higher lambda-rho value compared to the mu-rho and the P-impedance (Figure 3b).
Zone B is defined within the upper section of the Lower Barnett Shale. An average attribute data-vector (Shale
petrotype B) is extracted around well B (Figure 3a and b) from this zone. This zone mostly has siliceous noncalcareous shale lithofacies (Singh, 2008). Petrotype B shows a low lambda-rho and P-impedance (Figure 3b).


5

Zone C is defined within the lower section of the Lower Barnett Shale. The shale petrotype C is extracted around
well C (Figure 3a and b) from this zone. This lowermost Barnett Shale is the zone of hot shale (Pollastro et al.,
2007) with very high gamma ray. Petrotype C shows a low lambda-rho and P-impedance.
We use the three distance measures (explained before) to classify the attribute data vectors based on these shale
petrotypes A, B and C (target vectors) in the Barnett shale formation.

Figure 2: (a) The map of Texas highlighting major basins and uplifts. This Ft. Worth basin is bounded by the Muenster Arch to the northeast,
the Ouachita Thrust Front to the east, the Bend Arch to the west, the Red River Arch to the north, and the Llano Uplift to the south. (b)
Stratigraphic section including the Gamma ray and the resistivity log showing the major units. The Barnett sits on an angular unconformity above
the Cambrian to upper-Ordovician-age carbonates of the Ellenberger Group and Viola Formation and overlying Pennsylvanian-age Marble Falls
Limestone with the Forestburg Limestone in dividing the Barnett formation. (c) Cross-section of the stratigraphy of Ft. Worth basin from
Montgomery et al. (2005).


6

Figure 3: (a) The Location of the three horizontal wells Well A, Well B and Well C overlaid over the upper Barnett time horizon of the 3D
survey. (b) The location of the wells and their horizontal extension within the Barnett zone and the sub-volumes for extracting an average welldata vector. (c) The zones for extracting the average well data-vector are highlighted on a gamma ray log from the survey. Shale petrotype-A is
extracted from a zone within the upper Barnett shale, shale petrotype-B facies is extracted from the upper region of the lower Barnett zone, and
petrotype-C is extracted from the “hot” gamma ray region of the lower Barnett shale.

Discussion
Each of the three zones A, B and C within the Barnett shale formation are analyzed separately and separate
algorithm outputs separate classification volumes, which will be discussed subsequently. Each zone has five
separate volumes; one from minimum Euclidean distance measure (MED), one from spectral angle mapper (SAM)
and three PDF similarity volumes from GTM classification (Figures 4, 5, and 6). The results from each of the
classification volumes are analyzed and a comparative study is made.
Figure 4 shows the comparative study of all the five volumes from the Upper Barnett shale (zone A). The average
well data-vector is extracted around Well A, which acts as the target vector. The MED classification (Figure 4b) and
the SAM classification (Figure 4c) give similar results around the well (cyan color facies). However, the shale
petrotype-A spatially varies over the survey depending on the cutoff factor between MED and SAM. The PDF
similarity measure from Bhattacharya measure has three separate volumes (Figure 4d, 4e and 4f) corresponding to
the three predefined shale petrotypes. Analysis shows that the Upper Barnett zone has mostly rocks belonging to
petrotype-A (Figure 4d). However some rocks in this zone A have properties similar to petrotype C (Figure 4f).
Further we infer that it is extremely unlikely that the petrotype-B is present in this zone (Figure 4e).


7

Figure 4: Supervised classification analysis for Zone A, which corresponds to the Upper Barnett zone. (a) The shale facies petrotype-A is
extracted from this upper Barnett shale (blue arrow) around the Well A (red arrow). The shale petrotype A is colored cyan, shale petrotype B is
colored green and the shale petrotype C is colored red. Strata slice within zone A from the results of (b) Minimum Euclidean Distance (MED)
measure and (c) Spectral Angle mapper (SAM). The regions similar to shale petrotype-A (from the MED classification and the SAM classifier)
are colored cyan. The gray areas in b and c are unclassified and do not belong to any of the pre-defined shale petrotype. Figure c, d, and e show
the results of supervised GTM analysis. The “red” color in the colorbar highlights regions with highest probability and the “dark grey”
corresponds to the least probability, respectively of occurrence of the petrotype-A shale facies. (c) PDF Similarity volume for the petrotype -A
facies. The stratal slice shows that petrotype-A is abundant in this zone. (d) PDF Similarity volume showing the probability of occurrence of
shale petrotype-B (corresponding to zone B) within zone A. Mostly dark gray color highlights that the petrotype B has the least probability of
occurrence in zone A. (e) The probability of occurrence of the petrotype-C (corresponding to the hot-gamma zone C) are color-coded on the
strata-slice. The analysis shows that there are localized occurrences of the petrotype-C in zone A.


8

Figure 5: Supervised classification analysis for Zone B, which corresponds to the upper region of Lower Barnett zone. (a) The shale petrotype-B
is extracted from this upper region of Lower Barnett zone (blue arrow) around the Well B (red arrow). The shale facies A is colored cyan, shale
petrotype B is colored green and the petrotype C is cored red. Strata slice within zone B from the results of (b) Minimum Euclidean Distance
(MED) measure and (c) Spectral Angle mapper (SAM). The regions similar to shale petrotype-B (from the SAM classification and the MED
classifier) are colored green. Regions similar to Type-C shale facies are colored red. Figure c, d, and e show the results of supervised GTM
analysis. The “red” color in the colorbar highlights regions with highest probability and the “dark grey” corresponds to the least probability of
occurrence respectively of the Type-B shale facies. (c) PDF Similarity volume for the Type-A facies. Mostly dark gray color highlights that the
shale petrotype A has the least probability of occurrence in zone B. (d) PDF Similarity volume showing the probability of occurrence of shale
petrotype-B (corresponding to zone B) within zone B. Zone B shows the highest probability of spatial occurrence of Type B shale facies. (e) The
probability of occurrence of the petrotype-C (corresponding to the hot-gamma zone C) are color-coded on the strata-slice. The analysis shows
that there are localized occurrences of the petrotype-C in zone B.

Figure 5 above shows the comparative study of all the five volumes from the upper region of Lower Barnett shale
(zone B). The average well data-vector is extracted around Well B, which acts as the target vector. The MED
classification (Figure 5b) and the SAM classification (Figure 5c) give similar results around the well (green color
facies). However, the shale petrotype-B spatially varies over the survey for SAM classifier compared to MED
classifier, which is varying less spatially. We see the presence of petrotype C (red color facies) in this zone from the
SAM classifier. The PDF similarity measure from Bhattacharya measure has three separate volumes (Figure 5d, 5e
and 5f) corresponding to the three predefined shale petrotypes. Analysis shows that the upper region of Lower
Barnett zone has mostly rocks belonging to petrotype-B (Figure 5e) and less of petrotype-C (Figure 5f). We infer
that it is extremely unlikely that the petrotype-A is present in this zone (Figure 5d). Generally this region of the
Lower Barnett shale has siliceous non-calcareous shale lithofacies (Singh, 2008) and is brittle in nature (Perez,
2013). Shale petrotype B corresponds to brittle shale. Mapping these brittle zones is helpful because these are the
regions where a horizontal well can be placed for effective fracturing.


9

Figure 6: Supervised classification analysis for Zone C, which corresponds to the lower region of Lower Barnett zone. (a) The shale petrotype-C
is extracted from this lower region of Lower Barnett zone (blue arrow) around the Well C (red arrow). The shale facies A is colored cyan, shale
petrotype B is colored green and the petrotype C is cored red. Strata slice within zone C from the results of (b) Minimum Euclidean Distance
(MED) measure and (c) Spectral Angle mapper (SAM). The regions similar to shale petrotype-B (from the SAM classification and the MED
classifier) are colored green. Regions similar to Type-C shale facies are colored red. Figure c, d, and e show the results of supervised GTM
analysis. The “red” color in the colorbar highlights regions with highest probability and the “dark grey” corresponds to the least probability of
occurrence respectively of the Type-B shale facies. (c) PDF Similarity volume for the Type-A facies. Mostly dark gray color highlights that the
shale petrotype A has the least probability of occurrence in zone B. (d) PDF Similarity volume showing the probability of occurrence of shale
petrotype-B (corresponding to zone B) within zone B. Zone B shows the highest probability of spatial occurrence of Type B shale facies. (e) The
probability of occurrence of the petrotype-C (corresponding to the hot-gamma zone C) are color-coded on the strata-slice. The analysis shows
that there are localized occurrences of the petrotype-C in zone B.

Figure 6 above shows the comparative study of all the five volumes from the lower region of Lower Barnett shale
(zone C). The average well data-vector is extracted around Well C, which acts as the target vector. The MED
classification (Figure 6b) and the SAM classification (Figure 6c) give similar results around the well (red color
facies). However, the shale petrotype-C spatially varies over the survey for SAM classifier compared to MED
classifier, which is varying less spatially. We see the presence of petrotype B (green color facies) in this zone from
the SAM classifier. The three separate volumes from PDF similarity measure (Figure 6d, 6e and 6f) correspond to
the three predefined shale petrotypes. Analysis show that within the survey this zone C has mostly rocks belonging
to petrotype-B (Figure 6e) and petrotype-C (Figure 6f). We infer that it is extremely unlikely that the petrotype-A is
present in this zone (Figure 6d). From the study it is evident that the northeast corner of the survey around the well
C, it is mainly shale petrotype C. Petrotype B is present mostly in the center and south of the survey. The SAM
results are more consistent with the PDF similarity outputs. Shale petrotype C corresponds to the high gamma values
in the well logs and corresponds to regions of most ductile shale, which may be due to high TOC concentration in
this zone (Singh, 2008 and Perez, 2013).


10

Conclusions
In the paper, we have studied the three different supervised distance metric classifiers when applied on the multiattribute seismic data from the Barnett shale in the Fort Worth Basin. Our results demonstrate the unique divergence
properties of the three distance metrics discussed in the paper when seen on facies classification outputs. In all the
three zones of our case study, it is evident that the SAM classifier results are more consistent with the PDF similarity
results. The mapped brittle zones and the mapped high TOC zones are consistent within the results of the SAM
classifier and the PDF similarity. In this study we have tried to demonstrate that the classification based on distance
metrics have significant value in the identification and classification heterogeneous Barnett shale.
Acknowledgements
We thank Devon Energy for use their seismic and well data over the Barnett Shale in this work. Financial support
for this effort was provided by the industry sponsors of the Attribute-Assisted Seismic Processing and Interpretation
(AASPI) consortium. All the 3D seismic displays were made using licenses to Petrel, provide to OU for research
and education courtesy of Schlumberger.
References
Bishop, C. M., M. Svensen, and C. K. I. Williams, 1998, The generative topographic mapping: Neural Computation,
10, No. 1, 215-234.
Duda, R.O., Hart, P.E, Stork, D.E., Pattern Classification, 2nd Edition, Wiley, 2000, ISBN 978-0-471-05669-0.
Boardman, 1992, SIPS User’s Guide Spectral Image Processing System, Version 1.2, Center for the Study of Earth
from Space, Boulder, CO. 88 pp.
Michelena, J. R., E. S. M. Gonzalez, and M. Capello, 1998, Similarity analysis: A new tool to summarize seismic
attribute information: The Leading Edge, 17, 1361-1364
Montgomery, S. L., D. M. Jarvie, A. K. Bowker, and M. R. Pollastro, 2005, Mississippian Barnett Shale, Fort Worth
Basin, north central Texas: Gas-shale play with multi-trillion cubic foot potential, AAPG Bulletin;
February 2005; 89; 155-175.
Perez-Altamar, R., 2013, Correlation of surface seismic measurements to completion quality: Application to the
Barnett Shale, PhD dissertation, the University of Oklahoma.
Pollastro, R. M., D. M. Jarvie., R. J. Hill, and C. W. Adams, 2007, Geologic framework of the Mississippian Barnett
Shale, Barnett-Paleozoic total petroleum system, Bend Arch-Fort Worth Basin, Texas: AAPG Bulletin, 91,
405-436.
Poupon, M., Gil, J., D. Vannaxay, and B. Cortiula, Tracking Tertiary delta sands (Urdaneta West, Lake Maracaibo,
Venezuela): An integrated seismic facies classification workflow, The Leading Edge, 23, 909-912.
Singh, P., 2008, Lithofacies and Sequence Stratigraphic Framework of the Barnett Shale, Northeast Texas: Ph.D.
dissertation, University of Oklahoma.
Roy, A., 2013, Latent space classification of seismic facies: PhD dissertation, the University of Oklahoma.
Roy, A., Jayaram, V., Marfurt, K.J., 2013, Active learning algorithms for 3D seismic facies classification, extended
abstract SEG Annual Meeting, Sept. 2013, Houston, TX, USA.

Distance Metric Based Multi-Attribute Seismic Facies Classification to Identify Sweet Spots within the Barnett shale : A Case study from the Fort Worth Basin, TX

More Related Content

What's hot (18)

Similar to Distance Metric Based Multi-Attribute Seismic Facies Classification to Identify Sweet Spots within the Barnett shale : A Case study from the Fort Worth Basin, TX (20)

More from Pioneer Natural Resources (16)

Recently uploaded (20)

Distance Metric Based Multi-Attribute Seismic Facies Classification to Identify Sweet Spots within the Barnett shale : A Case study from the Fort Worth Basin, TX