PETROPHYSICS, VOL. 56, NO.

5 (OCTOBER 2015); PAGE 493–510; 21 FIGURES; 3 TABLES

Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA1
Mehrnoosh Saneifar2, Mark Skalinski4, Paul Theologou5, Jeroen Kenter6, Clifford Cuffey7, and Rafael Salazar-Tio8

ABSTRACT

McElroy ¿eld, located in the Permian Basin, is a typical were populated in the core and the well-log domains
example of a complex carbonate reservoir. Discovered in applying a supervised model trained using the k-Nearest
1926, McElroy ¿eld has been under waterÀood since the Neighbors algorithm (KNN).
early 1960s. However, maximizing oil recovery is still Computed tomography (CT) scan imaging techniques
a major challenge in this ¿eld. We have applied a rock- correlated to log-derived estimates of porosity were
typing workÀow based on conventional well logs and used to predict vuggy porosity in the well-log domain.
core data to incorporate both depositional and diagenetic Assessment of vuggy porosity using CT-scan image
attributes for characterizing the heterogeneity within the analysis showed that the separation of sonic porosity
McElroy ¿eld. The resulting rock-type distribution may and density-neutron porosity is not a reliable tool for
be used to ultimately enhance future development and oil estimating vuggy porosity in gypsum-bearing reservoirs.
production in the McElroy ¿eld. All of the generated geological and petrophysical data
The applied rock-typing workÀow consists of several were integrated to de¿ne the petrophysical rock types that
sequential steps. Firstly, the depositional rock types were control the reservoir’s dynamic characteristics. Validation
described and consolidated in the core domain for the of the petrophysical rock types by dynamic injection
purpose of propagation into the well-log domain. Next, pro¿les con¿rmed earlier assertions that Àuid Àow in this
reservoir typing was conducted to identify controls on ¿eld is dominantly controlled by diagenetic modi¿cations.
reservoir properties. This analysis indicated that diagenetic Finally, we studied the distribution of the identi¿ed
overprint has the dominant inÀuence on Àuid Àow in the petrophysical rock types to establish trends for ¿eldwide
McElroy ¿eld. In a subsequent step, pore types were spatial distribution of petrophysical rock types. The spatial
classi¿ed by clustering attributes of Gaussian function ¿ts trends of petrophysical rock types in the ¿eld serve to
to the pore-throat-radius distributions derived from mercury identify the potential for future development opportunities
injection capillary pressure (MICP) measurements. in the McElroy ¿eld.
The identi¿ed depositional rock types and pore types

INTRODUCTION attributes at core and well-log scales are crucial for appropriate
spatial distribution of rock classes in 3D static reservoir
Hydrocarbon recovery from vuggy and heterogeneous models (Skalinski and Kenter, 2014 and 2015). Common
carbonates can be challenging as pore systems and Àow rock classi¿cation methods in carbonate reservoirs are based
properties result from the interaction between depositional on either geological properties (i.e., depositional attributes)
and diagenetic processes. Successful development and interpreted at core scale, or petrophysical properties (i.e.,
reliable characterization of carbonate reservoirs can be capillary pressure, porosity, and permeability) determined at
enhanced by rock classi¿cation based on both depositional core or well-log scales.
and diagenetic properties and their impact on Àuid Àow. In The ¿rst category, geological rock classi¿cation,
addition, interrelations between geological and petrophysical includes techniques that are predominantly focused on

Manuscript was received by the Editor on September 4, 2015; revised manuscript received October 9, 2015.
1
Originally presented at the SPWLA 56th Annual Logging Symposium, Long Beach, California, USA, July 18-22, 2015, Paper M.
2
BHP Billiton, 1360 Post Oak Blvd. #150, Houston, TX 77056, USA; [email protected]
4
Chevron ETC, 1500 Lousiana St., Houston, TX 77002, USA; [email protected]
5
Chevron Australia Pty Ltd, 250 SGT, Perth QV1, Australia; [email protected]
6
Independent Consultant, 710 Roy St., Houston, TX 77007, USA; [email protected]
7
Chevron MCBU, 15 Smith Rd., Midland, TX 79705, USA; [email protected]
8
Chevron ETC, 1500 Louisiana St., Houston, TX 77002, USA; [email protected]

October 2015 PETROPHYSICS 493

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

depositional properties and do not include diagenetic to the dynamic petrophysical properties, they lack the critical
overprint corresponding to Àuid Àow in carbonate rocks links to geology and the spatial trends for populating rock
(Dunham, 1962; Embry and Klovan, 1971). These techniques classes in a reservoir model. In addition, these techniques
assume that reservoir properties can be adequately are not adequate for determining secondary pore systems
represented by depositional facies, while most carbonate (i.e., vugs and fractures) due to measurement-scale issues.
rocks are strongly altered by diagenetic mechanisms. Other Whole cores and advanced logging tools such as borehole
geological techniques are pore-typing methods based on image logs can be used for large-scale characterization of
pore- or grain-size measurements and core interpretations pore systems in the formation (Skalinski and Kenter, 2014).
(Choquette and Pray, 1970; Lucia, 1995 and 2007; Ahr, To address the aforementioned limitations of the
2008). Although these techniques incorporate textural conventional rock classi¿cation techniques in carbonate
information for classi¿cation, the rock types de¿ned based reservoirs, a recent study introduced a rock-classi¿cation
on these techniques are often not relevant to Àuid Àow as the workÀow based on an integrated application of core and
result of diagenetic modi¿cation. well logs (Skalinski and Kenter, 2014). The workÀow
The second category, petrophysical rock classi¿cation, establishes ties between geological and petrophysical
can be subdivided into (a) petrophysical partitioning properties in carbonate reservoirs, as well as spatial trends
techniques based on core measurements of porosity and for reservoir modeling. This technique was successfully
permeability (Leverett, 1941; Amaefule et al., 1993) and applied in two large carbonate ¿elds, Tengiz Field and
log clusters (Serra and Abbott, 1980), and (b) rock-typing Wafra Field. In this paper, we apply the petrophysical rock
techniques based on pore-throat-radius distribution, derived classi¿cation workÀow in the McElroy ¿eld, to further
from saturation-dependent mercury injection capillary validate its reliability in heterogeneous carbonate reservoirs
pressure (MICP) measurements (Pittman, 1992; Marzouk et with diverse data inventory.
al., 1995; Skalinski et al., 2006; Clerke, 2009).
Core-based petrophysical partitioning techniques METHOD
including Leverett J-function (Leverett, 1941) and FZI
(Amaefule et al., 1993) are methods that generate rock Petrophysical rock-type (PRT) workÀow consists
types by binning the core measurements of porosity and of eight consecutive steps that are adaptable to different
permeability. These techniques assume that pore-throat- reservoirs and data availability. Skalinski and Kenter (2014)
radius distributions can be represented by a normalizing provided a general description of the workÀow and the
parameter, square root of the ratio of permeability to porosity. procedures required for the application of the workÀow
Well logs can provide information about petrophysical in different data scenarios. The following section brieÀy
properties of the formation at all depths along the wellbore. explains the seven steps of the PRT workÀow conducted in
Serra and Abbott (1980) de¿ned electrofacies based on the McElroy ¿eld. The eighth step of this workÀow, which
distinct log responses corresponding to speci¿c rock classes is focused on using geostatistical tools (e.g., variograms and
in the formations. multiple point statistics) for distributing the identi¿ed PRTs
Several MICP-based pore-typing techniques were in 3D static reservoir models, is not the focus of this paper.
developed in recent decades. The Winland R35 technique
(Pittman, 1992) uses the pore-throat radius corresponding to Step 1—Data Scenario
35% of mercury (nonwetting phase) saturation, derived from The ¿rst step in the PRT workÀow is to investigate core
the MICP measurements, as an indicator of the effective Àow and log data availability and density in the ¿eld under study
properties. Marzouk et al. (1998) de¿ned three pore types (Skalinski and Kenter, 2014).
(micropores, mesopores, and macropores) in carbonate
formations based on the pore-throat radius measured by Step 2—Depositional Rock Typing
MICP or air-water centrifuge. Clerke (2009) illustrated the As previously mentioned, one of the main goals of the
use of Thomeer’s hyperbolas (Thomeer, 1960) for ¿tting PRT workÀow is integrating geology with petrophysics.
MICP measurements to quantitatively assess the pore- Therefore, a critical step in this workÀow is de¿ning
size distribution in complex carbonate formations. MICP- depositional rock types (DRTs) based on geological
based rock-classi¿cation techniques used in isolation lack attributes in the core domain. They are then predicted in the
an adequate prediction in the well-log domain, as required log domain using inputs including well logs and well-log-
for spatial distribution of the core-based rock classes. based estimates of petrophysical properties and mineralogy.
Although core-based and well-log-based petrophysical rock DRT prediction in the log domain usually requires iterative
classi¿cation techniques bene¿t from an indirect relationship lumping of the DRTs based on their similarity in terms of

494 PETROPHYSICS October 2015

 Saneifar et al.

depositional region, petrophysical properties, well-log magnetic resonance and borehole image logs), where
response, and statistical representativity in the core domain. available.
Multivariate statistical tools, supervised techniques, and
neural networks can be applied for DRT prediction in the
log domain (Skalinski and Kenter, 2014). In this paper,
we apply KNN algorithm (Cover and Hart, 1967), to build
a supervised model for prediction of the DRTs. KNN is a
classi¿cation method by which the data are classi¿ed based
on a majority vote of its neighbors, with the data point being
assigned to the most common class among its neighbors.

Step 3—Reservoir Typing

Reservoir typing is determination of the relative impact
of depositional, diagenetic, and structural (e.g. fracturing)
processes on the present-day Àow characteristics of the
reservoir. Several techniques can be incorporated for de¿ning
Fig. 1—Reservoir-typing procedure for RT assignment in the formation
reservoir types (RTs), depending on the data scenario, as based on relative contribution to the Àuid Àow (Skalinski and Kenter,
illustrated in Fig. 1. Convergence/divergence of porosity- 2014).
permeability (‫׋‬-K) transforms of the identi¿ed DRTs can
indicate whether petrophysical characteristics are strongly Step 5—PRT De¿nition
inÀuenced by the original DRTs. If ‫׋‬-K transforms in DRTs PRTs are ¿nally determined by combining the predicted
are nearly overlapping and exhibiting minimal divergence DRTs, PTs, barriers and other diagenetic modi¿ers affecting
in trend, diagenetic modi¿cation may be inferred as a the log response. The combination order of these elements
signi¿cant contributor to the Àuid Àow characteristics of the is de¿ned based on reservoir typing conducted in Step 4.
reservoir. Furthermore, conformity between predicted DRTs The identi¿ed PRTs have the following characteristics:
and logs can be tested to evaluate the impact of DRTs on (1) represent category of rocks that are characterized by
the log responses. Additional reservoir-typing techniques are speci¿c ranges of petrophysical properties (e.g. porosity and
described by Skalinski and Kenter (2014). Reservoir typing permeability); (2) exhibit distinct petrophysical relationships
can yield three different RT categories: RT1 Depositional, in relevant for reservoir modeling; (3) are identi¿ed by wireline
which Àuid Àow is controlled by DRTs; RT2 Diagenetic, in or logging-while-drilling (LWD) logging surveys; and (4)
which Àow characteristics are dominated by diagenesis and are linked to geological attributes, such as, primary texture
truncate DRT trends; RT3 Fracture, in which fractures are or diagenetic modi¿cations (Skalinski and Kenter, 2014).
the overriding Àow paths, and RT4 Hybrid, in which Àuid
Àow is variably controlled by any combination of RT1, RT2, Step 6—PRT De¿nition in a Multiwell Setting and
and RT3. The RT classi¿cation is directly included in Step 5 Quality Control
of the PRT workÀow, PRT de¿nition (Skalinski and Kenter, PRTs are populated in all wells using the supervised
2014). techniques implemented in the PRT workÀow. After an
assessment of possible outliers and overlaps, trends and
Step 4—Pore Typing interrelations for ¿eldwide spatial distribution of PRTs are
Carbonate formations are commonly characterized by extracted.
complex pore networks that involve different pore systems
with various shapes and interconnectivity. Petrophysical Step 7: Dynamic Validation of PRTs
properties and Àow characteristics are strongly controlled Available Àow indicators, such as, PLTs, DSTs, wireline
by distribution of pore systems in the formation. Thus, formation test data, and injection pro¿les are used for
classi¿cation and prediction of pore types are essential validating the identi¿ed PRTs and their correlation to Àuid
steps in carbonate rock typing. MICP measurements can be Àow.
used for accurate pore typing by assessing the pore-throat-
radius modality and distribution (Skalinski and Kenter, MCELROY FIELD APPLICATION
2014). Furthermore, MICP-derived PTs can be combined
with large-scale pore observations (i.e., vugs and fractures) The McElroy ¿eld, which is one of the major oil ¿elds in
from whole-core analysis and specialty logs (e.g., nuclear the Permian Basin, is located in Crane and Upton Counties

October 2015 PETROPHYSICS 495

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

in West Texas, on the Southeastern portion of the Central

Basin Platform (Fig. 2). The reservoir is a stratigraphic-
structural ramp facing the Midland Basin, with oil production
coming from Permian-age heterogeneous dolomites of
Grayburg formation, overlying San Andres formation
evaporitic dolomites (Avasthi et al., 1991; Nolen-Hoeksema
et al., 1994; Dehghani et al., 1999). Although dolomite is
the dominant mineral in this formation, quartz, gypsum,
anhydrite, and clay are also present locally, where they are
commonly associated with speci¿c DRTs. The McElroy
¿eld, discovered in 1926, has been under continuous
waterÀooding since the early 1960s. However, ultimate oil
recovery has been a major challenge in this ¿eld, due to
vertical and interwell heterogeneity (Nolen-Hoeksema et al.,
1994; Nguyen and Somerville, 2014).

Fig. 3—Map of the McElroy ¿eld showing the four production areas with
their respective reservoir performance.

has been recognized as the key to improve production in the

McElroy ¿eld (Nolen-Hoeksema et al., 1994; Nguyen and
Somerville, 2014). In this section, application of the PRT
workÀow for characterizing heterogeneity within McElroy
¿eld is described step-by-step.

Data Scenario
The PRT workÀow was conducted using the following
available data resources in the McElroy ¿eld:
• 15 wells with core measurements (i.e, porosity,
permeability, and grain density), core descriptions,
and geological facies de¿nitions. Porosity
measurements were conducted at both high and low
Fig. 2—Map showing the location of the McElroy ¿eld (Dehghani et al., temperatures, to account for the presence of gypsum
1999). in the formation.
• 62 wells with a total of 850 MICP sample
Petrophysical properties and Àow characteristics are measurements. This database was narrowed
highly variable across the ¿eld. Thus, the ¿eld is divided down to 500 samples after an extensive quality
into four production regions, (from west to east): Low control based on the accuracy of porosity
Permeability, High Quality, Low Pressure, and East Flank measurements and the availability of permeability
(Fig. 3). measurements in the core domain. In addition,
The highest oil production is from the central area measurements were checked for blank
(i.e., High Quality area) and the lowest oil recovery is from corrections, closure (conformance) corrections,
Àanking areas (i.e., East Flank region). Fluid Àow in certain merging of low- and high-pressure chamber data,
parts of McElroy ¿eld, especially the Low Pressure area, is removal/interpolation of bad data points, and
strongly inÀuenced by large porosity features including vuggy smoothing or rejection of noisy data (Theologou et al.
porosity, which enhance fracture porosity in the reservoir. (2015).
These features are largely driven by diagenetic overprint of • X-ray diffraction (XRD) data of a limited depth
the original pore systems. Reliable reservoir characterization interval in one well.

496 PETROPHYSICS October 2015

 Saneifar et al.

• Image analysis of whole-core CT-scans in a well of the DRT prediction was veri¿ed using a supervised and
characterized by vuggy porosity. a blind dataset on 80 and 20% of the input data in the core
• 42 wells with quad-combo log suite required for the domain, respectively. Tables 2 and 3 list the statistics of the
depth-by-depth assessment of porosity, permeability, supervised and the blind test, respectively. The results show
and mineralogy. Wells were selected from all four a relatively strong conformity between well logs and lumped
aforementioned production regions in the McElroy DRTs. DRTG was separately determined based on its distinct
¿eld. mineralogy from estimated mineralogy.

A “¿t-for-purpose” ¿eldwide multimineral model

was built to generate predictors of DRTs and PTs, namely
volumetric concentrations of minerals and porosity.
Mineral types were identi¿ed based on core descriptions
and geological information. Volumetric assumptions in the
model included dolomite, calcite, gypsum, anhydrite, quartz,
orthoclase feldspar, illite, and “effective Àuid”. Inputs to the
model included: gamma ray, bulk density, neutron porosity,
photoelectric factor, electrical resistivity, and compressional-
wave slowness. As the ¿eld was under an extensive
waterÀood, high uncertainties were assigned to the electrical
resistivity measurements. A constraint relating volumetric
concentrations of orthoclase feldspar and quartz was also
used to make the model more robust. The outcome of the
multimineral analysis was the depth-by-depth estimates of
porosity and mineralogy in all wells with quad-combo log
suite. Core measurements of porosity, core descriptions,
geological information, and XRD data were used to validate
the results of multimineral analysis. Furthermore, gypsum
Fig. 4—An illustration of DRT lumping in the McElroy ¿eld. On the left,
dehydrates and converts to water at temperatures above 12 original core-based DRTs are shown as lumped into the ¿nal DRTs
140°F (Tilly et al., 1982). Recognizing this phenomenon, (lumped), on the right.
volumetric concentration of gypsum was also validated
using differences in the high- and the low-temperature core
measurements of porosity. Permeability estimates in the log Reservoir Typing
domain were obtained by applying a supervised model based RT analysis was conducted using ‫׋‬-K transforms in the
on the KNN algorithm, with an input of core porosity and identi¿ed DRTs. Figure 5 shows the crossplot of porosity
permeability measurements. The model was trained using and permeability measurements, colored by the lumped
80% of the input dataset and tested on the remaining 20%. DRTs in the core domain. The negligible divergence in
Estimates of permeability were validated using permeability the ‫׋‬-K trends of the lumped DRTs indicates the minimal
core measurements. inÀuence of geological attributes on the Àuid Àow. Based
on this observation, it can be inferred that diagenetic
Depositional Rock Typing modi¿cations have the dominant inÀuence on the Àuid Àow
Twelve DRTs were de¿ned based on depositional in the McElroy ¿eld (Fig. 5). However, the relatively strong
environments, mineralogy (Table 1). DRTs were lumped for conformity between DRTs and logs, observed from the log-
the best predictably in the log domain, while maintaining domain prediction of DRTs in the previous step, indicates a
their distinct geological characteristics after lumping. The potential correlation between the geological attributes and
optimum DRT lumping resulted in seven lumped DRTs (Fig. the formation properties (e.g., mineralogy) represented by
4). A supervised model based on the KNN algorithm was log measurements. This is expected as DRTF and DRTG are
used for prediction of the lumped DRTs. Well logs and well- de¿ned mainly based on their mineralogy.
log-based estimates of porosity and mineralogy that resulted The diagenetic inÀuence on Àuid Àow was further
in the best prediction of the DRTs in the core domain were investigated in the pore-typing step for a reliable de¿nition
used for DRT prediction in the log domain. The accuracy of PRTs.

October 2015 PETROPHYSICS 497

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

Table 1—McElroy Field Core-Based DRT Assignment and Description of Depositional Regions and Mineralogy

Table 2—Statistics of DRT Prediction in the Supervised Test

Table 3—Statistics of DRT Prediction in the Blind Test

Fig. 5—Crossplot of core permeability and porosity measurements,

colored by the lumped DRTs in the McElroy ¿eld. Minimal divergence
in the ‫כ‬-K transforms indicates the strong inÀuence of diagenetic
modi¿cations on the Àuid Àow in the McElroy ¿eld.

498 PETROPHYSICS October 2015

 Saneifar et al.

Pore Typing ranges of porosity and permeability (Fig. 8) and a unique

Pore-throat-radius distributions were obtained from pore-throat-radius distribution (Fig. 9). PT1 is characterized
the available MICP measurements and were quantitatively by the highest ranges of porosity and permeability, as
characterized by ¿tting a multimodal Gaussian function into well as a monomodal pore-throat-radius distribution. PT2
each sample measurement. In general, Gaussian function exhibits intermediate ranges of porosity and permeability,
provides a better ¿t to the MICP data compared to Thomeer’s while having a predominantly bimodal pore-throat-radius
hyperbolas (Thomeer, 1960). Furthermore, for multimodal distribution. PT3 corresponds to the rock segments with
pore systems, Thomeer’s method produces arti¿cial the lowest porosity and permeability. Furthermore, PT1
discontinuities in the capillary pressure and pore-throat-size and PT3 were categorized as macropores and micropores,
distributions at the merge pressure of the hyperbolas, which as mean values of the pore-throat radii in PT1 and PT3 are
are a direct result of the superposition of one hyperbola on centered at approximately 1 and 0.1 ȝm, respectively (Fig.
another, and are not present in the measured data. (Xu and 9). This observation also explains the higher permeability
Torres-Verdín, 2013; Theologou et al., 2015). and hydraulic conductivity in PT1 compared to PT3. Next,
Gaussian function of total pore volume for a formation MICP-based pore types were predicted in the log domain
with more than one pore-throat-radius mode is given by using a KNN supervised model, trained by an input of core
porosity and permeability measurements and an output of
(1) the MICP-based pore types. The model was then applied to
populate the identi¿ed pore types in the core domain using
where p is the number of modes present and x(i) is given via inputs of core porosity and permeability measurements.
Another KNN model, trained by core porosity and
permeability measurements and identi¿ed core-based pore
. (2) types, was used to predict pore types in the log domain.
Input log predictors used in this model were evaluated and
selected based on a statistical test for achieving the optimum
VP’, Pm, and S are bulk volume, modal pressure, and pore predictability in the core domain.
geometric factor of each mode, respectively (Theologou
et al., 2015). A maximum number of four modes were
used in this work, to maintain viable petrophysical bounds
and consistency with previous observations in carbonate
formations. Furthermore, ¿t attributes and the number of
modes were optimized to obtain the best ¿t to the data.
To minimize the discrepancy in prediction of pore types
(PTs) at different measurement scales (i.e., MICP core chip,
core plug, and log), PTs are sequentially extrapolated from
the smallest to the largest measurement scales. PTs in the
MICP domain are determined by clustering the Gaussian
function ¿t parameters using Multi-Resolution Graph-Based
Clustering (MRGC) technique (Ye and Rabiller, 2001).
Supervised models (e.g., based on the KNN algorithm) and
inputs of core measurements and well logs are then used to
populate the PTs in the core and log domains, respectively.
Figures 6a and 6b show examples of the Gaussian function
¿t into a bimodal pore-throat-radius distribution and its
corresponding pressure measurements (plotted against water
saturation), respectively.
Clustering on the Gaussian function ¿t attributes
resulted in three pore type categories in the MICP domain:
PT1, macropores, PT2, bimodal pores, and PT3, micropores.
Figure 7 illustrates distinct distribution of the Gaussian
parameters: modal pressure (Pm), modal bulk volume (VP’),
and modal geometrical factor (S) in the identi¿ed pore types.
Furthermore, Figure 8 shows the identi¿ed pore types on a
‫׋‬-K crossplot and pore-throat-radius distributions in all pore Fig. 6—Application of the multimodal Gaussian function ¿t for
characterizing pore-throat-radius distribution, obtained from MICP
types are illustrated in Fig. 9. Each pore type exhibits distinct measurements. (a) plot of MICP-derived pore-throat-radius distribution;
(b) plot of MICP measurements.

October 2015 PETROPHYSICS 499

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

Fig. 7—Distribution of Gaussian parameters, modal pressure (Pm), modal bulk volume (VP’), and modal geometrical factor (S) in identi¿ed pore types
(color scale).

Fig. 8—Crossplot of permeability and porosity measurements, colored by the identi¿ed pore types based on the MICP data in the McElroy ¿eld. Pore
types exhibit distinct ranges of porosity and permeability. PT1 has the highest porosity and permeability, while PT2 and PT3 have intermediate and
low porosity and permeability, respectively.

500 PETROPHYSICS October 2015

 Saneifar et al.

Fig. 9—Pore-throat-radius distribution for three identi¿ed pore types based on the MICP data in the McElroy ¿eld. PT1 corresponds to macropores
with mean pore-throat radius of approximately 1 ȝm. PT2 exhibits bimodality of pore-throat radius. PT3 represents micropores with mean pore-throat
radius of approximately 0.1 ȝm.

Fig. 10—Crossplots of permeability and porosity for each pore type in the core and log domains.

October 2015 PETROPHYSICS 501

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

Distribution of the predicted pore types in core and log DRTs in Track 8 show a good match with the core DRTs in
domains in permeability-porosity crossplots is illustrated in Track 7. Tracks 10 and 11 in Fig. 11 also indicate a relatively
Fig. 10. Results show that pore types exhibit similar ranges of good agreement between the MICP-based pore types (red
porosity and permeability in both core plug and log domains. dots) and the pore types determined in the core and log
Figure 11 shows (a) stratigraphic intervals, including domains, respectively.
MCLR, top of Grayburg, A1 and E, upper Grayburg, D5, To investigate the impact of pore types on Àuid Àow,
middle Grayburg, and M, lower Grayburg; (b) results ‫׋‬-K transforms of the PTs determined in the log domain
of multimineral analysis, including estimates of mineral were compared. Figure 12 illustrates PTs in a crossplot
concentrations and estimates of porosity and permeability, of well-log-based estimates of permeability and porosity.
compared against corresponding core measurements (red The distinct ‫׋‬-K trend in each PT con¿rms that digenetic
dots); (c) identi¿ed DRTs (lumped) in the core and log attributes are the dominant factor inÀuencing Àuid Àow in
domains; and (d) identi¿ed pore types in the MICP, core, and the McElroy ¿eld.
log domains, in a key well in the McElroy ¿eld. Predicted

Fig. 11—Results of multimineral, DRT, and PT analysis in a key well in the McElroy ¿eld. From left to right: Track 1, depth; Track 2, stratigraphic tops;
Track 3, gamma ray and caliper logs; Track 4, estimates of volumetric concentrations of gypsum, anhydrite, dolomite, calcite, orthoclase feldspar,
quartz, and illite; Tracks 5 and 6, estimates of total porosity and permeability, compared to core measurements (red dots); Tracks 7 and 8, DRTs
determined in the core and log domains, respectively; Tracks 9 to 11, PTs identi¿ed in MICP, core, and log domains, respectively (red dots correspond
to the MICP-based-PTs).

502 PETROPHYSICS October 2015

 Saneifar et al.

Image analysis was also conducted on CT-scans of approximately 12%. In addition, a trend between the vuggy
approximately 12 m of core from a well with a noticeable porosity and porosity measurements in interval D5 can be
vuggy porosity to estimate the concentration of vugs in the deduced. This correlation was then used for estimating
formation. Figure 13 demonstrates the obtained CT-scans vuggy porosity from the well-log-based estimates of total
and vugs (in red) observed on the cores. The scanned section porosity in the uncored sections of D5 and M intervals in all
was from the stratigraphic intervals D5 and M. wells. Estimates of vuggy porosity were overlaid on PTs in a
crossplot of permeability and porosity in the log domain, as
shown in Fig. 15.

Fig. 12—Crossplot of well-log-based estimates of permeability and

porosity, colored by PTs. Divergence of ‫כ‬-K trends in PTs indicates the
strong inÀuence of diagenetic attributes on Àuid Àow in the McElroy ¿eld. Fig. 14—Crossplot of vuggy porosity estimated by image analysis and
core measurements of porosity in the stratigraphic intervals D5 and M in
a key well in the McElroy ¿eld.

Fig. 13—CT-scans of ‫׽‬12 m of core from intervals D5 and M of a key

well with vuggy porosity in the McElroy ¿eld. Vugs are shown in red.

Figure 14 shows the estimates of vuggy porosity, based

on image analysis, plotted against core measurements of
porosity. Results show negligible vuggy porosity in interval Fig. 15—Estimates of vuggy porosity in stratigraphic intervals D5 and
M, while vuggy porosity exists and increases in interval M, overlaid on PTs in a crossplot of permeability and porosity in the log
D5 for core intervals with a total porosity higher than domain. PT1, macropores, and PT2, bimodal pores, dominate vuggy
porosity.

October 2015 PETROPHYSICS 503

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

The results indicate that rock segments in PT1, i.e., the McElroy ¿eld as the result of diagenetic modi¿cation.
macropores, and PT2, i.e., bimodal pores, have been PRTs were ¿nally de¿ned as:
signi¿cantly affected by dissolution processes leading to • PRT1—Depositionally controlled evaporitic layers,
vuggy porosity. with the maximum concentration of anhydrite in the
Previous studies of secondary porosity in the McElroy formation and K <1 md
¿eld applied separation of sonic and density-neutron • PRT2—Depositionally controlled terrigenous beds,
porosity for the assessment of vuggy porosity (Dehghani containing laminations of quartz, clay, and feldspar,
et al., 1999). It should be noted that neutron-porosity with K <1 md
response overestimates porosity in the presence of gypsum. • PRT3—Diagenetic micropores
This increase in neutron porosity causes an increase in the • PRT4—Diagenetic bimodal pores
separation of sonic and density-neutron porosity that might • PRT5—Diagenetic macropores (Fig. 18).
be misinterpreted as the secondary porosity in the formation.
Figure 16 shows the estimates of vuggy porosity compared Two of the PRTs retain depositional attributes (PRT1
against separation of sonic and density-neutron porosity. and PRT2) because of their distinct mineralogy while the
Results show that separation of sonic and density-neutron remaining PRTs have little or no correlation with the DRTs.
porosity does not always indicate the presence of vugs in this Among the identi¿ed PRTs, PRT1 and PRT2 are barriers and
reservoir. will have a signi¿cant impact on Àuid Àow. PRT4 and PRT5
are the most productive rock types in the McElroy ¿eld,
because of high permeability and presence of diagenetic
features (i.e., vuggy porosity) in these PRTs.

Fig. 16—Comparison of vuggy porosity in the core domain against

separation of sonic and density-neutron porosity. The results show that
separation of sonic and density-neutron porosity is not a reliable tool for
estimating vuggy porosity in gypsum-bearing reservoirs.

PRT De¿nition
PRTs were de¿ned based on the outcome of DRT,
RT, and PT analyses. An additional prerequisite for PRT
de¿nition was to investigate distribution of PTs in DRTs and
vice versa (Figs. 17a and 17b). The results indicate that no Fig. 17—(a) Distribution of PTs in DRTs, and (b) distribution of DRTs
in PTs in the McElroy ¿eld. No clear trend can be observed between
clear trend exists between the occurrence PTs and DRTs in occurrence of DRTs and PTs.

504 PETROPHYSICS October 2015

 Saneifar et al.

be drawn from the cross-section plot, including: (a) high

concentrations of PRT1, PRT2, and PRT3 in the stratigraphic
intervals MCLR and A1 in all regions; (b) higher occurrence
of PRT5 in the intervals E and D5 in the central area (i.e.,
High Quality and Low Pressure areas), compared to the
same stratigraphic zone in the Low Permeability and the
East Flank areas; (c) dominance of PRT3 and PRT4 in
the intervals E and D5 in Low Permeability area; and (d)
higher concentration of PRT5 in the interval E in the East
Flank area, compared to the Low Permeability area. These
observations can be applied, as spatial rules, to geostatistical
tools used for distribution of PRTs in 3D static reservoir
models. In addition, the high concentration of PRT5 in the
midsection of the East Flank region, which is estimated to
Fig. 18—Final PRT de¿nition in the McElroy ¿eld.
have the highest estimated oil in place in the ¿eld, unveils
PRT De¿nition in a Multiwell Setting and Quality the potential for future development in this area.
Control
PRTs were populated in all wells following the Dynamic Validation of PRTs
appropriate quality control and cutoffs applied for prediction PRTs in the McElroy ¿eld were validated using available
of PRT1 and PRT2, based on multimineral analysis results. injection pro¿les. Figure 21 shows an example of the side-
Figure 19 illustrates the identi¿ed PRTs in a crossplot of by-side comparison between the identi¿ed PRTs and the
permeability and porosity in the log domain. injection pro¿les. The High Quality PRT (i.e., PRT5) in all
wells with injection pro¿les corresponded to an increase in
Àuid injectivity. This conformance con¿rmed the validity of
the identi¿ed PRTs and also indicated the impact of the PRTs
on the formation Àow characteristics.

DISCUSSION

As indicated previously, the present paper is an

application of the PRT workÀow (Skalinski and Kenter,
2014) to the McElroy ¿eld in the Permian Basin. We consider
this an important validation/testing of the workÀow, which
was applied before to different carbonate reservoirs, such as,
Tengiz (low porosity limestone with bitumen) and 1st Eocene
(high-porosity dolomite with evaporates). Both reservoirs
had limited MICP data, which prevented capturing a full
impact of diagenesis on pore systems. The McElroy ¿eld
also contains evaporates and clastic lithofacies. Although
none of them were part of the main reservoir, it helped to
validate the workÀow in a mixed carbonate/clastic system.
We believe that McElroy ¿eld application showed that PRT
workÀow is Àexible enough to be applied to carbonate,
Fig. 19—Final PRTs illustrated on a crossplot of permeability and mixed carbonate-clastic and clastic reservoirs. Application
porosity in the log domain in the McElroy ¿eld.
to unconventional reservoirs would require modi¿cations to
include mechanical properties and kerogen content/porosity.
Furthermore, PRT distribution in all production
The presented results are showing an example of
regions and stratigraphic intervals was investigated for
PRTs strongly inÀuenced by diagenetic pore types, which
extracting spatial trends for future enhancement in reservoir
were not only derived from a robust MICP database (500
characterization and development in the McElroy ¿eld.
samples), but interpreted using a Gaussian modal inversion
Figure 20 indicates a typical west-east cross-section plot
and successfully predicted in log domain. The pore types
that was used for this purpose. Several interrelations can

October 2015 PETROPHYSICS 505

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

effectively subdivided carbonate formations into rock types rather than identifying certain DRTs with porosity and
linked to the productivity of the reservoir. The pore types permeability ranges. Similar geological DRTs were lumped
can also be the rationale for creating robust saturation height to optimize the prediction scores by logs, however, pore
functions for reservoir models. Application of conventional types only display a signi¿cant link with two of the DRTs
electrofacies clustering would not provide rock types with and are later used as PRTs (DRTG and DRTF). The linkage
clear petrophysical meaning. with pore types allows embedding depositional properties
Furthermore, the observed DRTs are commonly in the PRT de¿nition (PRT1 and PRT2, respectively). This
recognized in the Grayburg Formation and each has a relationship is critical for future interwell prediction of PRTs.
particular spatial position in the progression from (bottom No visible link is established between the DRTs and PTs
to top) open shelf, shallow and ¿nally shallow shelf and (Fig. 17). Homogenization by diagenesis is also con¿rmed
intertidal environments as a result of transgression and by the minimal divergence in ‫׋‬-K transforms (Fig. 5).
subsequent progradation (Harris et al., 1984; Tucker et al., Two of the PRTs retain depositional attributes (PRT1 and
1998). The PRT workÀow provides a systematic approach, PRT2) while the remaining PRTs (PRT3 to 5) have little

Fig. 20—A typical west-east cross-section plot illustrating the lateral (in production regions) and vertical distributions (in stratigraphic intervals) of
PRTs in the McElroy ¿eld.

506 PETROPHYSICS October 2015

 Saneifar et al.

Fig. 21—PRT validation using injection pro¿les in the McElroy ¿eld. PRT5 corresponds to an increase in Àuid injectivity, as shown in Tracks 7 and
8, respectively.

October 2015 PETROPHYSICS 507

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

to no correlation with the DRTs. Diagenetic modi¿cation, DRTs and PTs, respectively, indicated that diagenesis is
suggested as a key parameter, is the only process that can the main factor inÀuencing Àuid Àow in the ¿eld. Vuggy
explain these observations. Tucker et al. (1998) con¿rm such porosity was also estimated in two stratigraphic intervals D5
overprint by diagenesis obscuring lithological differences and M based on a CT-scan image analysis of whole cores.
and making prediction of reservoir quality based on DRTs A correlation between vuggy porosity and total porosity
dif¿cult or impossible. Figure 20 shows the presence was used for estimating vuggy porosity in the log domain.
of PRTs 1 and 2 near the top of the prograding tidal Àat, Results showed the highest estimated concentration of vugs
while most of the underlying section displays PRT3 to 5 in macro- and bimodal PTs in the formation. Furthermore,
trends that are crosscutting the larger depositional stacking estimated vuggy porosity was compared against separation
patterns, observed in previously published work (Harris of sonic and density-neutron porosity. This comparison
et al., 1984; Tucker et al., 1998). Diagenetic PRTs, which con¿rmed that the separation of sonic porosity and density-
are represented by pore types, allow de¿ning clear spatial neutron porosity is not a reliable tool for estimating vuggy
trends of reservoir quality. Identi¿ed zones with PT2 in the porosity in gypsum-bearing reservoirs. Finally, DRTs and
Low Permeability area suggest local diagenetic dissolution/ PTs were combined into ¿ve PRTs. Distribution of PRTs
enhancement, which can be used for reservoir management in vertical and lateral directions was investigated to extract
and development decisions. Evaporates, classi¿ed as PRT1, spatial rules for 3D static reservoir models and future
constitute a strong permeability barrier capping reservoir. development of McElroy ¿eld.
Pore types mappable at ¿eld scale would be a valuable
tool for reservoir management decisions and selection of ACKNOWLEDGMENTS
bypassed productive zones.
These observations need to be validated by future work, The authors would like to thank Robert Mallan for his
which would require, as a next logical step, prediction of the contributions in the pore-typing portion of this work. We
PRTs in a larger number of wells and subsequent extraction greatly appreciate Chevron MidContinent Business Unit for
(through kriging) of spatial trends in PRTs. Obtained patterns permission to publish this paper.
should be validated by diagenetic observations and analogs
from outcrop and literature. NOMENCLATURE
Most of the work used for this paper was completed
during a three-month internship of M. Saneifar at Chevron, Abbreviations
which restricted the scope of the work. Some of interesting CT = computed tomography
results, such as vuggy porosity quanti¿cation and their DRT = depositional rock type
prediction from porosity, need to be validated by more KNN = k-Nearest Neighbors Algorithm
robust data in order to formulate ¿eld wide predictions. LWD = logging while drilling
MICP = mercury injection capillary pressure
SUMMARY AND CONCLUSIONS MRGC = Multi-Resolution Graph-based Clustering
PRT = petrophysical rock type
Enhanced reservoir characterization in McElroy ¿eld is PT = pore type
critical for continuous development of this heterogeneous RT = reservoir type
¿eld. The PRT workÀow was successfully applied in Symbols
the McElroy ¿eld and its outcome was validated using K = permeability, md
the available dynamic data. As part of this workÀow, Pc = capillary pressure, psi
petrophysical properties and geological attributes were Pm = modal pressure of pore-throat radius, psi
integrated for reliable characterization of lateral and p = number of modes in pore types
vertical heterogeneity in the formation. Core-derived ‫ = ׋‬porosity, v/v
DRTs were ¿rst lumped for the best prediction in the log S = modal pore geometric factor
domain. Statistical analysis indicated a strong conformance VPT = total pore volume, cm3
between the predicted DRTs and log, due to the presence VP’ = modal bulk volume, cm3
of DRTs that were strongly dependent on lithology. Three
PTs (i.e., micropores, bimodal pores, and macropores) were
determined based on a quantitative analysis of an extensive
MICP database and were populated in the core and log
domains. Convergence and divergence of ‫׋‬-K trends in

508 PETROPHYSICS October 2015

 Saneifar et al.

REFERENCES dx.doi.org/10.2118/29883-MS.
Nguyen, H.D., and Somerville, J.M., 2014, Difference in Behavior
Ahr, W.M., 2008, Geology of Carbonate Reservoirs, Wiley- of the WaterÀood Mechanism in Two Typical Areas of a
InterScience, 144–176, ISBN: 978-0470164914 Fractured Carbonate Reservoir, McElroy Field, Paper SPE-
Amaefule, J.O., Altunbay, M., Tiab, D., Kersey, D.G., and Keelan, 169028 presented at the SPE Improved Oil Recovery
D.K., 1993, Enhanced Reservoir Description: Using Symposium, Tulsa, Oklahoma, USA, 12–16 April, DOI:
Core and Log Data to Identify Hydraulic (Flow) Units and http://dx.doi.org/10.2118/169028-MS
Predict Permeability in Uncored Intervals/Wells, Paper SPE- Nolen-Hoeksema, R.C., Avasthi, J.M., Pape, W.C., and El Rabaa,
26436 presented at the SPE Annual Technical Conference and A.W.M., 1994, WaterÀood improvement in the Permian
Exhibition, Houston, Texas, USA, 3–6 October, DOI: http:// Basin: impact of in-situ-stress evaluations, Paper SPE-24873,
dx.doi.org/10.2118/26436-MS. SPE Journal, 9(4), DOI: http://dx.doi.org/10.2118/24873-PA.
Avasthi, J.M., Nolen-Hoeksema, R.C., and Al Rabaa, A.W.M., Pittman, E.D., 1992, Relationship of Porosity and Permeability
1991, In-situ Stress Evaluation in the McElroy Field, West to Various Parameters Derived from Mercury Injection–
Texas, Paper SPE-20105, SPE Journal, 6(3), 301–309, DOI: Capillary Pressure Curves for Sandstone, AAPG Bulletin,
http://dx.doi.org/10.2118/20105-PA. 76(2), 191–198.
Xu, C., and Torres-Verdín, C., 2013, Core-Based Petrophysical Serra, O. and Abbott, H.T., 1980, The Contribution of Logging
Rock Classi¿cation by Quantifying Pore-System Data to Sedimentology and Stratigraphy, Paper SPE-9270, SPE
Orthogonality with a Bimodal Gaussian Density Function, Journal, 22(1), 117–131, DOI: http://dx.doi.org/10.2118/9270-
Paper SCA2013-2079 presented at the SCA International PA.
Symposium, Napa Valley, California, USA, 16–19 September. Skalinski, M., Gottlib-Zeh, S., and Moss, B., 2006, De¿ning and
Choquette, P.W., and Pray, L.C., 1970, Geologic nomenclature Predicting Rock Types In Carbonates—Preliminary Results
and Classi¿cation of Porosity in Sedimentary Carbonates, from an Integrated Approach Using Core and Log Data in
AAPG Bulletin, 54(2), 207–250. Tengiz Field, Petrophysics, 47(1), 37–52.
Clerke, E.A., 2009, Permeability, Relative Permeability, Skalinski, M., and Kenter, J., 2014, Carbonate Petrophysical
Microscopic Displacement Ef¿ciency, and Pore Geometry Rock Typing: Integrating Geological Attributes and
of M_1 Bimodal Pore Systems in Arab D Limestone, Paper Petrophysical Properties While Linking With Dynamic
SPE-105259, SPE Journal, 14(3), 524–531, DOI: http:// Behavior, in Agar, S.M., and Geiger, S., editors, Fundamental
dx.doi.org/10.2118/105259-PA. Controls on Fluid Flow in Carbonates: Current WorkÀows to
Cover, T.M., and Hart P.E., 1967, Nearest Neighbor Pattern Emerging Technologies, Geological Society of London,
Classi¿cation, IEEE Transactions on Information Theory, Special Publications, 406, 229–259, ISBN: 978-1-86239-659-
13(1), 21–27. DOI: 10.1109/TIT.1967.1053964. 3; DOI: http://dx.doi.org/10.1144/SP406.6.
Dehghani, K., Harris, P.M., Edwards, K.A., and Dees, W.T., 1999, Skalinski, M., and Kenter, J., 2015, Integrated WorkÀow
Modeling a Vuggy Carbonate Reservoir, McElroy Field, West or Method for Petrophysical Rock Typing in Carbonates,
Texas, AAPG Bulletin, 83(1), 19–42. U.S. Patent No. 9,097,821, Published August 4, 2015.
Dunham, R.L., 1962, Classi¿cation of Carbonate Rocks According Theologou, P.N., Skalinski, M., and Mallan, R.K., 2015, An MICP-
to Depositional Texture, in Hamm, W.E., editor, Classi¿cation Based Pore Typing WorkÀow—Core Scale To Log Scale,
of Carbonate Rocks, a Symposium, AAPG Memoir 1, 108– Paper L, Transactions, SPWLA 56th Annual Logging
121. Symposium, Long Beach, California, USA, 18–22 July.
Embry, A.F. and Klovan, J.E., 1971, A late Devonian Reef Tract Thomeer, J.H.M., 1960, Introduction of a Pore Geometrical Factor
on Northeastern Banks Island, Northwest Territories, Bulletin De¿ned by the Capillary Pressure Curve, Paper SPE-1324-G,
of Canadian Petroleum Geology, 19, 730–781. Journal of Petroleum Technology, 12(3), 73–77, DOI: http://
Harris, P.M., Dodman, C.A., and Bliefnick, D.M., 1984, Permian dx.doi.org/10.2118/1324-G.
(Guadalupian) Reservoir Facies, McElroy Field, West Texas, Tilly, H.P., Gallagher, B.J, and Taylor, T.D., 1982, Methods
in Harris, P.M., editor, Carbonate Sands—a Core Workshop, for Correcting Porosity Data in a Gypsum-Bearing Carbonate
SEPM Core Workshop, No. 5, 136–174. Reservoir, Paper SPE-9716, Journal of Petroleum Technology,
Leverett, M.C., 1941, Capillary Behavior in Porous Solids, Paper 34(10), 2449–2454, DOI: http://dx.doi.org/10.2118/9716-PA.
SPE-941152-G, Transactions, AIME, 142(1), 159–172. Tucker, K.E., Harris, P.M., and Nolen-Hoeksema, R.C., 1998,
Lucia, F.J., 1995, Rock-Fabric/Petrophysical Classi¿cation of Geologic Investigation of Cross-Well Seismic Response in a
Carbonate Pore Space for Reservoir Characterization, AAPG Carbonate Reservoir, McElroy Field, West Texas, AAPG
Bulletin, 79(9), 1275–1300. Bulletin, 82(8), 1463–1503.
Lucia, F.J. 2007, Rock-Fabric Classi¿cation, Chapter 2, in Ye, S., and Rabiller, P.J.Y.M., 2001, Multi-Resolution Graph Based
Carbonate Reservoir Characterization, 2nd Edition, Springer- Clustering, U.S. Patent No. 6,295,504, Published September
Verlag, 30–55, ISBN: 978-3-540-72740-8. 25, 2001.
Marzouk, I., Takezaki, H., and Miwa, M., 1995, Geologic controls
on Wettability of Carbonate Reservoirs, Abu Dhabi,
U.A.E., Paper SPE-29883 presented at the SPE Middle East
Oil Show, Manama, Bahrain, 11–14 March, DOI: http://

October 2015 PETROPHYSICS 509

 Integrated Petrophysical Rock Classi¿cation in the McElroy Field, West Texas, USA

ABOUT THE AUTHORS subsurface team leader for Kharyaga Field in Russia testing
new technologies. During his tenure with ConocoPhillips
Mehrnoosh Saneifar is a petrophysicist at BHP he held the post of Technical Team Leader Carbonates/UR
Billiton Petroleum. She recently received her PhD degree in in the Reservoirs Team at ConocoPhillips and worked on
Petroleum Engineering from the Harold Vance Department subsalt Angola Exploration and the Bakken/Three Forks
of Petroleum Engineering at Texas A&M University. She Field in North Dakota.
received her MSc and BSc degrees in Petroleum Engineering
from Texas A&M University. Mehrnoosh worked as a Clifford Cuffey is currently Senior Staff Geologist
graduate research assistant in the Texas A&M Multiscale for the McElroy Technical Team (Chevron Midcontinent
Formation Evaluation Joint Industry Research Program Business Unit). He received his MS degree from the
under Dr. Zoya Heidari’s supervision from spring 2012 to University of Oklahoma (1992). He has worked for Chevron
2015. During summer 2014, she worked at Chevron Energy since 1997, including ¿ve years in the Gulf of Mexico,
Technology Company as a petrophysicist intern. She also followed by 13 years in the Permian Basin.
worked at RasGas in Qatar during summers 2006 and 2007,
and Shell in Dubai during summer 2008. Mehrnoosh serves Rafael Salazar-Tio has worked for Chevron since 2006.
Petrophysics Journal and Journal of Natural Gas Science & He has been working on NMR inversion methods, pore-
Engineering as a technical reviewer. scale simulations, and multiscale petrophysics. He received
a BS in Physics from the San Marcos National University,
Mark Skalinski is currently a Senior Research Peru in 1995 and a PhD in Physics from The University
Consultant in Petrophysics in Chevron ETC. He has MSc of Balearic Island, Spain in 2000. Previously he was a
(1971) and PhD (1979) degree in Geophysics from Mining Research Associate at the Eindhoven Technical University,
University in Cracow. His previous assignments include The Netherlands and in the Washington University in St.
Tengizchevroil in Atyrau, Kazakhstan, Chevron Canada Louis, Missouri, both in Physical chemistry.
Resources in Calgary, CABGOC in Angola, Husky Oil in
Calgary, ONAREP in Morocco and University of Mining &
Metallurgy in Cracow Poland. Mark’s interests include rock-
typing methodology, petrophysical multimineral modeling,
application of statistical methods for facies and permeability
prediction and integrated petrophysical ¿eld studies. Mark is
a member of SPE, SPWLA and AAPG.

Paul Theologou is currently the Petrophysics Team

Lead for Chevron Australia Pty Ltd. in Perth. Paul received
his PhD.in applied geology from the University of South
Australia (1997). He worked as a petrophysicist at Santos
Ltd., A.C.S. Laboratories, and Mincom Ltd. before starting
a petrophysical consulting company (the Saros Group)
in 2000. In 2008, Paul joined Chevron ETC as a research
petrophysicist at in Houston, TX, where he was involved
with the development of petrophysical interpretation work
Àows for carbonates and mineral modelling.

Jeroen Kenter currently is a consultant in carbonate

reservoir characterization in Houston. He has an MSc from
Amsterdam University and a PhD from Free University
in The Netherlands, both in Geology. Jeroen has worked
with Chevron ETC on subsurface assets such as Tengiz,
Karachaganak, Wafra, McElroy at the interface of R&D and
technical deliveries for the business units. Subsequently,
with Statoil ASA in Bergen, Norway, he worked on
reservoir characterization and rock typing workÀows and as

510 PETROPHYSICS October 2015