t Technical paper

Interpretation of porosity and fluid constituents from well logs

using an interactive neutron-density matrix scale
Olabode Ijasan1, Carlos Torres-Verdn1, and William E. Preeg2
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Abstract
Neutron and density logs are important borehole measurements for estimating reservoir capacity and infer-
ring saturating fluids. The neutron log, measuring the hydrogen index, is commonly expressed in apparent
water-filled porosity units assuming a constant matrix lithology whereby it is not always representative of actual
pore fluid. By contrast, a lithology-independent porosity calculation from nuclear magnetic resonance (NMR)
and/or core measurements provides reliable evaluations of reservoir capacity. In practice, not all wells include
core or NMR measurements. We discovered an interpretation workflow wherein formation porosity and hydro-
carbon constituents can be estimated from density and neutron logs using an interactive, variable matrix scale
specifically suited for the precalculated matrix density. First, we estimated matrix components from combina-
tions of nuclear logs (photoelectric factor, spontaneous gamma ray, neutron, and density) using Schlumbergers
nuclear parameter calculator (SNUPAR) as a matrix compositional solver while assuming freshwater-filled for-
mations. The combined effects of grain density, volumetric concentration of shale, matrix hydrogen, and neu-
tron lithology units define an interactive matrix scale for correction of neutron porosity. Under updated matrix
conditions, the resulting neutron-density crossover can only be attributed to pore volume and saturating fluid
effects. Second, porosity, connate-water saturation, and hydrocarbon density are calculated from the discrep-
ancy between corrected neutron and density logs using SNUPAR and Archies water saturation equation,
thereby eliminating the assumption of freshwater saturation. With matrix effects eliminated from the neu-
tron-density overlay, gas- or light-oil-saturated formations exhibiting the characteristic gas neutron-density
crossover become representative of saturating hydrocarbons. This behavior gives a clear qualitative distinction
between hydrocarbon-saturated and nonviable depth zones.

Introduction 2N 2D
2t ; (1)
Porosity calculated from neutron and density mea- 2
surements is still the most commonly used estimate
of pore volume in rock formations penetrated by wells. where D and N are the density- and neutron-apparent
In complex lithologies, inadequate characterization of porosities, respectively. Mao (2001) studies the correla-
the matrix could yield inaccurate porosity and satura- tion characteristics of D and N for identification of
tion estimates. The petrophysical effects of lithology, oil- and gas-saturated zones. Spears (2006) applies lith-
saturating fluid, and borehole conditions on nuclear logs ofacies-based porosity corrections derived from neu-
are exhaustively discussed by Ellis and Singer (2007). tron-density crossplots for t calculations in geologic
Using departure curves from log interpretation charts and reservoir models. Fertl and Timko (1971) extend
(Schlumberger, 2009), corrections are applied such that Gaymard and Poupons (1968) formulations for calcula-
interpreted properties are representative of the forma- tion of hydrocarbon density hc and detection of oil- and
tion only. Extensive studies and publications on neutron gas-bearing intervals in shaly sands.
and density logs, being ubiquitous for porosity and hy- The neutron-density overlay technique relies on the
drocarbon estimation, can be found in the literature. difference between apparent porosities, on a prede-
Historically, total porosity t in gas-bearing for- fined matrix scale, for inferring hydrocarbon saturation
mations is approximated with the following formula (S hc ), t , and hc . Several petrophysical factors ad-
(Gaymard and Poupon, 1968): versely affect the reliability of the overlay technique.

1
University of Texas at Austin, Austin, Texas, USA. E-mail: [email protected]; [email protected].
2
Private consultant, Austin, Texas, USA. E-mail: [email protected].
Manuscript received by the Editor 5 June 2013; revised manuscript received 12 July 2013; published online 10 October 2013. This paper appears
in Interpretation, Vol. 1, No. 2 (November 2013); p. T143T155, 8 FIGS., 6 TABLES.
http://dx.doi.org/10.1190/INT-2013-0072.1. 2013 Society of Exploration Geophysicists and American Association of Petroleum Geologists. All rights reserved.

Interpretation / November 2013 T143

 For example, gas detection is challenging in shaly where m and f are assumed matrix and fluid densities,
sands or shale gas reservoirs due to opposite effects respectively, e.g., limestone matrix of 2.71 gcm3 and
of shale-hydroxyls and gas density in overlay character- freshwater of 1 gcm3 . The porosity departure D
istics. Similarly, in oil-saturated or invaded gas zones, due to m and f assumptions (Ellis et al., 2007) is quali-
the decreased difference between neutron and density- tatively and quantitatively intuitive such that
apparent porosities masks the characteristics of light-
hydrocarbon crossovers (Mao, 2001). Consequently, b m 1
D 2
m f ; (4)
f m f m m
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

application of the overlay technique requires implemen-

tation of a suitable matrix correction.
In this paper, we estimate S hc , t , and hc using where m and f are differences in matrix and fluid
an interactive interpretation workflow based on the densities, respectively, between assumed and true
neutron-density overlay technique, with explicit con- properties.
sideration of neutron matrix scale and shale content.
The interpretation workflow improves reliability of the Neutron-apparent porosity
overlay technique in the presence of arbitrary lithology The neutron log is an apparent porosity measure-
and fluid effects. These effects and their influence on ment, given that it refers to an equivalent hydrogen in-
neutron and density-apparent porosities, along with dex (HI) response in water-filled lithology units, usually
conventional well-log interpretation methods, are de- limestone. Limestone unit implies an equivalent re-
scribed with a synthetic example of known lithology sponse of water-filled limestone formation where the
and fluid constituents. Additionally, the interactive pore volume equals that of the neutron log. As shown
matrix scale method is applied to field examples of by Gaymard and Poupon (1968), the environmentally
varying geology and lithology, namely, carbonate, sil- corrected neutron-porosity log N across invaded for-
iciclastic, and shale reservoirs, where porosity and mations can be expressed as
fluid-saturation estimates are compared to laboratory
core measurements. The calculated hc enables differ- 1 X M

entiation between gas- and oil-saturated intervals when N HI V t N ; (5)

HImf i1 i i
hc < 0.25 gcm3 and hc 0.25 gcm3 , respectively,
for simplified interpretation. Hence, the interactive where HI is the hydrogen index, subscript mf identifies
analysis method is implemented for qualitative identi- mud filtrate, and N is porosity departure due to neu-
fication of fluid zones, fluid contacts, and reservoir tron-apparent porosity measurement. Equation 5 im-
compartments. plies that the neutron response is a superposition of
the volumetric contributions of component hydrogen
concentrations. The apparentness in N is determined
Interpretation of apparent porosity
by the neutron-porosity unit (pu), usually water-filled
The porosity value associated with neutron logs is
limestone. For example, quartz, calcite, and dolomite
inherently apparent for given matrix and fluid units.
blocks yield N of 2, 0, and 0.5 limestone pu, respec-
On the other hand, compensated bulk density measure-
tively. This matrix effect is qualitatively intuitive be-
ments bear no apparentness until density porosity is
cause quartz and dolomite have lower and higher
calculated with constant values of matrix and fluid
matrix densities, respectively, than limestone. On the
properties. This is a significant difference between den-
other hand, unlike D , the matrix effect is quantitatively
sity and neutron logs.
obscure and cannot be calculated directly from equa-
tions 3 or 4. This effect is exacerbated in complex mix-
Density-apparent porosity
tures of various lithologies. Similarly, a gas-saturated
The bulk density measurement b principally re-
limestone formation yields negative N because the
sponds to formation electron density such that
HI of gas is typically lower than that of water.
X
M Hence, a physically intuitive parameter representa-
b 1.0704e;i 0.188V i ; (2) tion of neutron-porosity responses is necessary. Using
i1 neutron characteristic lengths, specifically migration
length Lm a calibration of Lm -to-neutron porosity is
where e;i and V i are the electron density and volume used to quantify matrix and lithology effects (Ellis et al.,
fraction, respectively, of the ith fluid/matrix component 2007).
up to M components. In hydrocarbon-bearing forma-
tions, t can be directly calculated from density logs
if and only if matrix density m and fluid density f are Neutron parameter model
known precisely. Otherwise, density-apparent porosity The Schlumberger nuclear parameter calculator
D is obtained using (SNUPAR, McKeon et al., 1989) calculates nuclear
properties such as Lm , HI, photoelectric factor (PEF),
b m
D t D ; (3) capture cross section, , etc., for any given mixture of
f m rocks and fluids. In this paper, we implement Lm for

T144 Interpretation / November 2013

 property characterization of wireline neutron poro- shale-hydroxyl effect Nsh can be approximated from
sity responses, typically with an americium-beryllium equation 6 using the expression
(AmBe) neutron source. It follows that equation 5

can be rewritten as 1
Nsh V sh gl ; (8)

 Lmsh
1
N gl ; (6)
Lm where Lmsh is the SNUPAR-calculated migration length
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

in shale. For example, Lmsh is approximately 15.35 cm

where gl represents the Lm -to-porosity calibration func- for illite of density 2.78 gcm3 , whereby Nsh for pure
tion in limestone water-filled units and gl is obtained by illite (see Figure 1), i.e., V sh 1, corresponds to 0.156.
fitting a polynomial function to the inverse of SNUPAR- In unconventional reservoirs with an organic-rich ker-
calculated Lm and limestone pore volume. Figure 1 ogen matrix (Passey et al., 1990), the neutron porosity
shows gs , gl , and gd water-filled calibration functions response increases due to high hydrogen content of or-
for sandstone, limestone, and dolomite units, respec- ganic matter. The SNUPAR-calculated HI of kerogen
tively. Additionally, the SNUPAR-calculated compen- could be as high as 0.8, depending on the hydrogen-
sated neutron tool (CNT) thermal porosity response,
carbon ratio and kerogen density. Accordingly, equa-
shown in dashed blue, agrees well with gl . Unless oth-
tion 8 quantifies the matrix-hydrogen effect, where
erwise stated, neutron-porosity logs in this paper are
V sh 1 and Lmsh become V ker (volume fraction of ker-
expressed in limestone matrix units, where g1 l and gl ogen) and Lmker (SNUPAR-calculated migration length
are used to convert neutron porosity to Lm logs, and
of kerogen matrix), respectively.
vice versa, respectively.
It then follows that the total matrix effect on neutron-
Furthermore, we implement a SNUPAR-based com-
porosity logs is an addition of matrix and Nsh ,
positional solver (Heidari et al., 2012) for estimation of
mineral and fluid concentrations from nuclear logs. The i.e., interactive porosity departures due to apparent
solver uses nonlinear minimization of a constrained- limestone matrix scale (calculated from SNUPAR in
error, quadratic cost function between SNUPAR- freshwater-filled assumptions, equations 6 and 7) and
predicted properties and nuclear logs (b , N , PEF) shale-hydroxyl or matrix-hydrogen effect (equation 8).
for estimation of mineral and fluid volumetric fractions. The corrected or rescaled neutron-apparent porosity is
Additionally, volumetric concentration of shale V sh and given by
water saturation S w are calculated using linear scaling Ncorr N matrix Nsh : (9)
of the gamma ray (GR) log and Archies equation,
respectively.
Fluid and hydrocarbon saturation effects
Given equations 3, 5, 6, and 9, fluid and saturation
Lithology effects effects on rescaled neutron-apparent porosity can be
Matrix effect written as
Equation 4 describes the sensitivity of D such that
the matrix effect in water zones, i.e., when f 0 and
f cw (connate-water density), is given by
 
b m 1
D m : (7)
f m 2 f m

In the neutron-density overlay technique, water-satu-

rated zones are expected to overlap only if the matrix
scale for density and neutron corresponds to the pre-
cise formation lithology. Otherwise, the neutron-density
matrix effect, matrix N D (equations 3 and
5), depends on m and the neutron response, N , of
the matrix. Unlike D in equation 3, N of the matrix
is not quantitatively intuitive, and it is only obtained
from equation 6 by converting SNUPAR-calculated
Lm to neutron porosity. Qualitatively, with a limestone
matrix scale in water zones, matrix < 0 across sand-
stone and matrix > 0 across dolomite.
Figure 1. SNUPAR-calculated water-filled neutron porosity
Shale-hydroxyl or matrix-hydrogen effect calibration functions gs , gl , and gd for sandstone, limestone,
Typically, shales consist of clay minerals with high and dolomite units, respectively. The figure also shows neu-
hydroxyl (OH ) content such that N > D . The tron porosity responses across relevant formations.

Interpretation / November 2013 T145

 Ncorr t 1 S hc ; (10) saturating fluids. This paper introduces a new inter-
pretation method, or interactive analysis workflow,
where is the difference in neutron response between that combines petrophysical effects due to apparent
hydrocarbon- and water-saturated formations. Several matrix scale and hydrocarbon saturation. Using a syn-
forms of equation 10 are given in Gaymard and Poupon thetic example of a layered earth model, where well
(1968), Mao (2001), and Quintero and Bassiouni (1998). logs are simulated with the University of Texas at Aus-
Gaymard and Poupon (1968) characterize across in- tin petrophysical and well-log simulator (UTAPWeLS,
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

vaded formations as the relative difference in HI be- Voss et al., 2009), we describe the estimation of
tween residual hydrocarbon and mud-filtrate; i.e., t , hc , and S hc using the interactive interpretation
workflow.
HIhc HImf HImf ; (11)
Interpretation workflow
where the subscript hc identifies hydrocarbon. For gas-
The first part of the interpretation involves rock/
saturated formations at reservoir conditions, one has
matrix compositional interpretation from b , PEF,
HIg 9g 0.15 0.20.9 g 2
; (12) and GR logs using the SNUPAR-based solver under
the assumption of freshwater-filled saturation. We as-
where the subscript g describes gas. Equation 12 is re- sume freshwater-filled formations for two reasons:
plicated in SNUPAR for hc g < 0.25 gcm3 , while a (1) The environmentally corrected N is typically refer-
SNUPAR-derived functional relationship is obtained for enced on freshwater-filled units and (2) to independ-
oil (Cn H2n2 ) when hc g > 0.25 gcm3 . Estimation ently characterize matrix effects for estimation of m
of S hc , t , and hc thus requires solving equations 2, given that formation fluids have negligible or no effect
10, 11, 12, and inclusion of a water saturation model, on PEF and GR logs.
e.g., Archies equation, Using the estimated m from the matrix solver and
equation 3, we calculate density-apparent porosity
aRw under the freshwater-filled assumption, Dwf . Accord-
Rt ; (13)
m
t 1 S hc n ingly, neutron-apparent porosity under the fresh-
water-filled assumption, Nwf , is obtained by converting
where Rt is the resistivity log, Rw is connate-water or the predicted Lm from the matrix solver to neutron
mud-filtrate resistivity, a is Archies factor, m is the porosity. It follows from equations 8 and 9 that
porosity exponent, and n is the saturation exponent. matrix Nsh Nwf Dwf , i.e., the interactive
It follows that 0 corresponds to a water or deeply neutron-density lithology effect in limestone porosity
invaded zone. Consequently, the magnitudes of and scale, where V sh is calculated assuming linear scaling
hc dictate the hydrocarbon type, i.e., oil or gas. of the GR log. We then calculate the corrected neu-
tron-apparent porosity Ncorr from equation 9 for
rescaling with D . At this point, the overlay character-
Interactive analysis of matrix and fluid effects istics of Nwf and Dwf are solely due to porosity effects,
Well-log interpretation involves conceptual rock mod- and the overlay of Ncorr and D is due to hydrocarbon
els when evaluating formation rock composition and pore volume.
The second part of the interpretation involves imple-
menting the SNUPAR-based solver for hydrocarbon
characterization. In this step, equations 2, 10, 11, 12,
and 13 are solved such that a SNUPAR-defined inherent
relationship between and hc is implemented in
the analysis for estimation of hc , S hc , and t . The func-
tional relationship between HI and hc is derived
from SNUPAR for oil (hc > 0.25 gcm3 ) and gas
(hc < 0.25 gcm3 ).
Figure 2 summarizes the interpretation workflow
where the Matrix solver loop is interactive as rock
components (e.g., quartz, dolomite, pyrite, etc.) are
chosen to quantify their effects on the calculated neu-
tron-density matrix scale. Additionally, we compare es-
timated m to core measurements wherever available
and appraise the solvers numerical reproduction of
PEF and GR measurements. Based on these compari-
sons, an interpretation decision is made concerning
the most representative formation rock components.
Figure 2. Interactive interpretation workflow for interpreta- Consequent with the Fluid solver of Figure 2, final in-
tion of neutron and density-apparent porosities. terpreted results include total porosity, hydrocarbon

T146 Interpretation / November 2013

 density, water saturation, matrix/grain density, and for- of Figure 4. Hydrocarbons, in comparison to fresh
mation rock components. water, reduce the neutron porosity response because
of lower HI (equation 5). From equations 11 and 12,
Synthetic example the hydrocarbon effect is dependent on hc and is ac-
The interactive interpretation workflow is described centuated in gas-saturated layers when compared to
for matrix and fluid effects on density and neutron- oil-saturated layers. In Figure 4d, layer III shows lower
apparent porosities, using numerically simulated mea- hydrocarbon effects and could be inadvertently inter-
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

surements across a synthetic and simplistic earth preted as a water-filled layer. Consequently, the fluid
model. This model is designed to describe practical sit- solver incorporates the resistivity measurement, Ar-
uations that present challenges to the interpretation of chies model (equation 13), equations 2, 10, 11, and
neutron and density logs. 12 for an inclusive calculation of hc , S hc , and t .
Tables 1 and 2 describe the properties assumed for Figure 4h shows that the estimated hc reliably predicts
the synthetic earth model, while Figure 3 shows the gas and oil densities in gas- and oil-saturated layers II
simulated nuclear and resistivity logs. In Figure 4, we and III, respectively. In Figure 4f, the t approximation
describe the interpretation results obtained with the in- using Gaymard-Poupons formula (equation 1) is valid
teractive analysis workflow. Figure 4a and 4f4h shows in layer II but inaccurate in shaly layers.
that estimated m , t , S w , and hc , respectively, using In layers VI and VII, for oil- and water-saturated
the interactive interpretation, agree well with model dolomite, respectively, the overlay characteristics in
properties in Table 1. It is particularly significant that Figure 4b indicate a matrix crossover. The matrix effect
the calculated hc in Figure 4h distinguishes between in panel d shows that matrix 0.0072 (i.e., 0.72 pu) for
gas- and oil-saturated layers. sh 0. By comparison, SNUPAR-calculated CNT re-
Layers I and IV consist of water-saturated shale of sponse yields apparent thermal neutron porosity of
mixed orthoclase and illite clay, where sh 0.155 0.5 pu in dolomite of 0% pore volume.
and shale density sh 2.738 gcm3 . After correction
for shale-hydroxyl effects, the actual matrix crossover
effect, due to the shale density greater than limestone
density, is shaded in brown in Figure 4b. On the other
hand, layer V consists of gas-saturated shale with 20% Table 2. Summary of assumed Archies parameters and
fluid properties for the synthetic example.
water saturation (refer to Table 1), such that the gas
crossover effect becomes accentuated after correction
for shale-hydroxyl effect. In this layer because gas Variable Value Units
saturation and V sh impose opposite overlay character-
istics, N D experiences a competition between gas Connate water resistivity, Rw at 200F 0.0203 m
and shale-hydroxyl effects. This behavior in neutron- Connate water density, cw 1.11 gcm3
density interpretation is especially common in logs ac- Connate water HI, HIcw 0.936
quired across shale gas formations. Connate water salt concentration 160,000 ppm NaCl
Layers II and III consist of gas- and oil-saturated lime- Archies factor, a 1
stone formations, respectively. The matrix effect is Archies porosity exponent, m 1.95
irrelevant in these layers because limestone is the refer- Archies saturation exponent, n 1.75
ence scale for neutron-density overlay. This behavior is
corroborated by the overlap of Dwf and Nwf in panel c

Table 1. Layer properties assumed in the synthetic example.

Layer Matrix Saturation fluid properties Interpretation comments

I Shale: 80% illite, 20% orthoclase, t 0.10 Sw 1, Shc 0 Shale and matrix effects
sh 2.738 gcm3
II Limestone t 0.28, Sw 0.05, Shc 0.95 (methane, Gas effect
CH4 0.182 gcm3 )
III Limestone t 0.28, Sw 0.05, Shc 0.95 (Liquid hydrocarbon, Hydrocarbon effects
C16 H34 0.757 gcm3 )
IV Shale: 80% illite, 20% orthoclase t 0.05, Sw 1, Shc 0 Shale and matrix effects
V Shale, 80% Illite, 20% Orthoclase t 0.10, Sw 0.20, Shc 0.80 (Methane, Shale and gas effects
CH4 0.182 gcm3 )
VI Dolomite t 0.28, Sw 0.05, Shc 0.95 (liquid hydrocarbon, Matrix and hydrocarbon
C16 H34 0.757 gcm3 ) effects
VII Dolomite t 0.10 Sw 1, Shc 0 Matrix effects
VIII Limestone t 0 Limestone reference

Interpretation / November 2013 T147

 Field examples of application The calculated hc in Figure 5h, with an average
In this section, the interactive interpretation work- value of 0.176 gcm3 , confirms that the reservoir is
flow is implemented for estimation of m , t , S w , and largely saturated with gas. Conclusively, we implement
hc in two field examples: (I) gas-bearing carbonate combined matrix and fluid volumetric analysis with
field of dolomite lithology where m > 2.71 gcm3 , the SNUPAR-based solver, where methane gas of
and (II) oil-bearing shale formation where m < 0.176 gcm3 is assumed as a component of the fluids,
2.71 gcm3 . thus eliminating the water-filled assumption in the inde-
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

pendent matrix analysis. Figure 5i shows cumulative

Field example I, gas-bearing carbonate plots of the volumetric fractions of shale, quartz, calcite,
This field example consists of conventional wireline dolomite, water, and gas, obtained from the SNUPAR-
nuclear and dual induction resistivity logs acquired based solver. The estimated m and t (Figure 5e and 5f,
across a gas-producing dolomite reservoir. Addition- respectively), agree well with core measurements. On
ally, the well includes routine core measurements. the other hand, log-derived S w (Figure 5g) within inter-
Due to low reservoir pressure, deep mud-filtrate inva- val XX08XX32 m is considerably lower than core mea-
sion affects the nuclear logs and even the deep resistiv- surements. This behavior can be attributed to variations
ity log, such that log-derived S hc is considerably lower in Archies parameters for differing rock types along the
than in situ S hc for water-base mud (Xu et al., 2012). well. Furthermore, S w in core samples could increase
Table 3 summarizes the assumed Archies parameters due to quick spurt loss in low-porosity, low-pressure
and fluid properties for the gas-bearing carbonate field. reservoirs (Xu et al., 2012).
Figure 5 shows the field measurements together with
core measurements, compared to results obtained with Field example II, oil-bearing shale example
the interactive interpretation. The neutron-density over- In this example, nuclear and array induction resistiv-
lay in Figure 5b emphasizes the matrix crossover be- ity logs are acquired in a well drilled with oil-base mud
cause the reservoir is primarily of dolomite lithology. across an oil-bearing shale formation from the Eagle
The gas flag in Figure 5j, proportional to hydrocarbon Ford shale play. Table 4 describes the assumed fluid
pore volume, is most pronounced across XX45XX55 m properties and Archies parameters for the oil-bearing
despite the suppressed gas crossover in Figure 5b. shale reservoir. Figure 6 shows field measurements,
Across the interval in Figure 5, the gas flag provides core measurements, and interpreted petrophysical
a qualitative and unequivocal indication of hydrocarbon properties for the oil-bearing shale example. Here,
saturation despite mud-filtrate invasion and matrix the SNUPAR-based matrix analysis assumes kerogen
crossover. (C100 H100 O8 of density 1.4 gcm3 ), calcite, kaolinite,

Figure 3. Simulated well logs across the syn-

thetic multilayer model. (a) GR log, (b) neu-
tron and density-apparent porosities on a
limestone scale, (c) array induction apparent
resistivity logs, and (d) PEF log. Refer to
Table 1 for a description of assumed layer
properties.

T148 Interpretation / November 2013

 and illite as components of the matrix. Figure 6e com- curve) is significantly larger than the core m (blue
pares m from the interactive analysis to core measure- circle points). This result is attributed to the exclusion
ments and elemental capture spectroscopy (ECS) of the low-density kerogen matrix from the ECS analy-
lithology analysis. The ECS-derived m (dashed blue sis. The matrix density m from SNUPAR-based matrix
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 4. Interpretation results for the synthetic example using the interactive interpretation workflow. (a) Interpreted matrix
density from SNUPAR-based matrix solver, (b) neutron-density overlay showing shale-corrected neutron log, matrix, and fluid
crossover characteristics, (c) neutron and density-apparent water-filled logs from SNUPAR-based matrix solver, (d) interactive
flag indicators showing matrix effect and gas flag, (e) corrected neutron-density overlay, (f) estimated total porosity, (g) estimated
water saturation, and (h) estimated hydrocarbon and fluid densities. Refer to Table 1 for a description of layer properties.

Table 3. Summary of assumed fluid properties and Archies parameters for field example I, gas-bearing carbonate.

Variable Value Units

Connate water resistivity, Rw at 96F 0.04 m

Connate water density, cw 1.12 gcm3
Connate water HI, HIcw 0.932
Connate water salt concentration 170,000 ppm NaCl
Mud-filtrate water resistivity, Rmf at 96F 0.84 m
Mud-filtrate water density, mf 1 gcm3
Mud-filtrate HI, HImf 1
Mud-filtrate water salt concentration 5147 ppm NaCl
Archies factor, a 1
Archies porosity exponent, m 1.96
Archies saturation exponent, n 1.83

Interpretation / November 2013 T149

 analysis (red curve) agrees well with core measure- (Figure 6f and 6g respectively), agree well with core
ments. The resulting fluid crossover, in Figure 6b, after measurements.
matrix-hydrogen and shale-hydroxyl corrections, is due
to the combined effects of m (less than 2.71 gcm3 of Fluid zone identification
limestone), fluid density, and fluid HI. It is found that Conventional methods for fluid contact identification
the interactive analysis yields a relatively constant include interpretation of pressure gradients due to fluid
hc of 0.747 gcm3 for the interval in Figure 6h. Further- density differences in the reservoir hydrostatic column.
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

more, estimated t and S w from the interactive analysis Additionally, impermeable sealing or geological barriers,

Figure 5. Interpretation results for field example I, gas-bearing carbonate reservoir, using the interactive analysis workflow.
(a) GR log, (b) neutron and density porosities on limestone scale, (c) dual-induction resistivity logs, and (d) PEF log. (e) Matrix
density, (f) total porosity, and (g) water saturation from core measurements and interactive analysis. (h) Calculated fluid densities
showing a gas cutoff of 0.25 gcm3 . (i) Volumetric concentrations of rock and fluid components from the SNUPAR-based solver.
(j) Gas flag from interactive analysis workflow.

Table 4. Summary of assumed fluid properties and Archies parameters for field example II, oil-bearing shale.

Variable Value Units

Connate water resistivity, Rw at 215F 0.019 m

Connate water density, cw 1.077 gcm3
Connate water HI, HIcw 0.901
Connate water salt concentration 165,000 ppm NaCl
Archies factor, a 1
Archies porosity exponent, m 2.1
Archies saturation exponent, n 2

T150 Interpretation / November 2013

 often at residual saturations, prevent hydraulic commu- gradients. Water saturation, S w , shown in Figure 7e,
nication between fluid zones such that a higher density estimated using the dual-water resistivity model, indi-
fluid resides above a lower density fluid in the hydro- cates that the hydrocarbon column exhibits a complete
carbon column. capillary transition with an aquifer below X660 m. Pres-
Occasionally, when pressure measurements are un- sure gradients identify three fluid zones, i.e., gas of
available or expensive to acquire, well logs are used to 0.263 gcm2 density, oil of 0.647 gcm3 density, and
infer fluid zones. In this section, the estimated hc log an aquifer at residual hydrocarbon saturation with con-
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

yielded by the interactive interpretation is used for a nate water of 1.005 gcm3 density, distinguished by
quick-look qualitative identification of hydrocarbon red, green, and black intervals, respectively. Figure 7g
zones and fluid contacts along the reservoir column shows the fluid densities, fluid zones, and fluid con-
in two more field examples. tacts, where estimated fluid densities, f ;p , from pres-
sure gradients are juxtaposed with f estimated with
Field example III, identification of hydrocarbon the interactive interpretation.
fluid contacts in a North Sea siliciclastic reservoir Qualitatively, f (Figure 7g) and t (Figure 7h) from
This example consists of a siliciclastic reservoir the interactive interpretation agree well with pressure
located in the central North Sea, where rock formations and core measurements, respectively, except across
consist of noncalcareous mudstones interbedded with the interval between X550 and X600 m. This interval
shaly sand deposits (Heidari et al., 2012). Figure 7a7d consists of highly interbedded sand-shale sequences;
shows nuclear and array induction resistivity measure- evident from the GR log in Figure 7a, whereby log-
ments acquired in a vertical well drilled with oil-base derived t and f are significantly influenced by
mud, while Table 5 summarizes the assumed properties shoulder-bed effects, and depth-by-depth analysis is
and Archies parameters for the siliciclastic reservoir. In inadequate. Note that the estimated f from the interac-
addition, available pressure data in Figure 7f describe tive interpretation agrees well with f ;p across the thick
three distinct and approximately constant pressure bed layers in the gas zone. Nonetheless, assuming no

Figure 6. Interpretation results for field example II, oil-bearing shale reservoir, using the interactive analysis workflow. (a) GR
log, (b) neutron and density porosities on limestone scale, (c) array induction resistivity logs, and (d) PEF log. (e) Matrix density,
(f) total porosity, and (g) water saturation from core measurements and interactive analysis. (h) Calculated fluid densities showing
a gas cutoff of 0.25 gcm3 .

Interpretation / November 2013 T151

 reservoir compartmentalization and good hydraulic with an estimated oil density of 0.43 gcm3 is above
communication, the gas-oil contact is located at the gas-saturated compartment at X817X819 m. The
X600 m, while the oil-water contact is located at compartmentalization of the gas layer is possible be-
X655 m where the water zone is at residual hydrocar- cause hydraulic communication is severed between
bon saturation. the oil and gas zones by the interleaving imperme-
able nonnet shale barriers. This example verifies the
Field example IV, identification of reservoir
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

compartments in the deepwater Gulf of Mexico Table 5. Summary of assumed properties and Archies
In this example, the reservoir consists of channel parameters for field example III, North Sea
levees located in the deepwater Gulf of Mexico, siliciclastic reservoir.
where formations consist of unconsolidated shaly sand
intervals and are primarily saturated with oil. Table 6 Variable Value Units
summarizes the assumed properties used in the
interactive interpretation with a dual-water resistivity Connate water resistivity, Rw at 254F 0.025 m
model. In Figure 8, the panels describe well logs and Connate water density, cw 1.005 gcm3
interpretation results across a hydrocarbon-saturated Connate water HI, HIcw 0.997
interval in the Gulf of Mexico reservoir. Figure 8e Connate water salt concentration 77,600 ppm NaCl
shows that average total porosities in the clean and
Archies factor, a 1
shaly sand layers are 0.2721 and 0.1724, respectively.
Archies porosity exponent, m 1.89
In Figure 8c and 8g, we observe a gas-saturated reser-
voir compartment between X817 and X819 m, where Archies saturation exponent, n 1.92
gas density is 0.144 gcm3 and the neutron-density Shale porosity, sh 0.10 vv
overlay exhibits significant gas crossover. The pri- Shale resistivity, Rsh 1.50 m
mary oil-saturated zone, between X778 and X802 m,

Figure 7. Fluid zone interpretation results for field example III, North Sea siliciclastic reservoir. (a) GR log, (b) PEF log, (c) neu-
tron and density porosities on limestone scale, and (d) array induction resistivity logs. (e) Estimated water saturation, (f) pressure
measurements, and (g) fluid densities from interactive analysis and pressure gradients. (h) Total porosity from core measurements
and interactive analysis.

T152 Interpretation / November 2013

 capability of the interactive interpretation workflow to Conclusions
distinguish between oil- and gas-saturated layers, irre- The interactive interpretation workflow rescales the
spective of formation lithology. The workflow also neutron-density overlay with corrected neutron and
provides an efficient qualitative method for identifica- density-apparent porosities in a variable matrix scale,
tion of reservoir compartments separated by sealing for independent characterization of fluid effects. It
barriers. was found that the SNUPAR-based matrix analysis, as-
suming freshwater-filled formations, renders accurate
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

estimations of matrix density even across hydrocar-

Table 6. Summary of assumed properties and Archies bon-saturated intervals. Such a result is due to the fact
parameters for field example IV, deepwater Gulf of that formation fluids have negligible or no effect on PEF
Mexico reservoir.
and GR logs. One limitation of the SNUPAR-based ma-
trix analysis is that a priori qualitative knowledge of
Variable Value Units matrix components, i.e., lithology, clay mineral, etc.,
is essential for accurate estimation of matrix density.
Connate water resistivity, Rw at 150F 0.030 m
This is achieved by preliminary lithology or matrix iden-
Connate water density, cw 1.098 gcm3 tification crossplots, e.g., PEF-b , thorium-potassium,
Connate water HI, HIcw 0.9441 and PEF-potassium crossplots. Furthermore, the work-
Connate water salt concentration 140,000 ppm NaCl flow assumes minimal shoulder-bed effects such that
Archies factor, a 1 depth-by-depth analysis is adequate for SNUPAR calcu-
Archies porosity exponent, m 1.92 lations. The uncertainty in estimated hc increases in
Archies saturation exponent, n 2.00 thinly bedded intervals with pronounced shoulder-
Shale porosity, sh 0.15 vv bed effects.
Shale resistivity, Rsh 1.0 m
The merits of the SNUPAR-based interactive inter-
pretation workflow and its contributions to the practice
of interpretation include the following:

Figure 8. Fluid zone interpretation results for field example IV, deepwater Gulf of Mexico reservoir. (a) GR log, (b) PEF log,
(c) neutron and density porosities on sandstone scale, and (d) array induction resistivity logs. (e) Total porosity, (f) water sat-
uration, and (g) fluid density logs estimated using the interactive analysis.

Interpretation / November 2013 T153

 unequivocal identification of hydrocarbon-satu- V sh = Volumetric concentration of shale (vv)
rated zones, = Capture cross section (cu)
estimation of model-consistent/lithology-indepen- = Density (gcm3 )
dent formation porosity, and = Departure
calculation of hydrocarbon density for gas/oil- = Apparent porosity (vv)
zone identification. m = Matrix density (gcm3 )
t = Total porosity (vv)
It was shown that the workflow incorporates inter-
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

= Neutron fluid effect parameter ( )

active matrix corrections such that Gaymard-Poupons
formulation for lithology-independent porosity and hy-
drocarbon identification can be implemented for any Subscripts
neutron-density matrix scale and lithology (clean and/
b = Bulk
or shaly), especially in wells with limited data.
corr = Corrected
Synthetic and field examples of application indicate
cw = Connate water
that lithology-independent porosity and hydrocarbon
e = Electron
density can be efficiently estimated from conventional
f = Fluid
nuclear and resistivity logs for reliable and quantita-
g = Gas
tive detection and appraisal of hydrocarbon-saturated
hc = Hydrocarbon
sweet spots and nonviable zones. Furthermore, identi-
i = Component index
fication of fluid types in the reservoir column provides a
ker = Kerogen
qualitative means for determining fluid contacts and
mf = Mud-filtrate
reservoir compartments.
N = Neutron
nsh = Nonshale
Symbols and nomenclature
sh = Shale
a = Winsauer factor in Archies equation ( ) t = Total
AmBe = Americium-beryllium wf = Water-filled
AO10 = 25.4-cm (10-in) array induction one-foot
resistivity ( m)
AO30 = 76.2-cm (30-in) array induction one-foot Acknowledgments
resistivity ( m) The work reported in this paper was funded by the
AO90 = 228.6-cm (90-in) ray induction one-foot University of Texas at Austins Research Consortium
resistivity ( m) on Formation Evaluation, jointly sponsored by Afren,
CNT = Schlumberger-compensated neutron
Anadarko, Apache, Aramco, Baker-Hughes, BG, BHP
tool
Billiton, BP, Chevron, ConocoPhillips, COSL, ENI,
gd = Neutron-porosity calibration function,
ExxonMobil, Halliburton, Hess, Maersk, Marathon Oil
in dolomite units
Corporation, Mexican Institute for Petroleum, Nexen,
gl = Neutron-porosity calibration function,
in limestone units ONGC, OXY, Petrobras, PTT Exploration and Produc-
GR = Gamma ray American Petroleum Insti- tion, Repsol, RWE, Schlumberger, Shell, Statoil, Total,
tute Weatherford, Wintershall, and Woodside Petroleum Lim-
gs = Neutron-porosity calibration function, ited. We are indebted to Shell Oil Company for providing
in sandstone units the core and well-log measurements used in this study.
HI = Hydrogen index ( )
ILD = Deep induction resistivity ( m) References
ILM = Medium induction resistivity ( m) Ellis, D. V., and J. M. Singer, 2007, Well logging for earth
Lm = Neutron migration length (cm) scientists: Springer.
m = Archies porosity exponent Fertl, W. H., and D. J. Timko, 1971, A distinction of oil and
M = Number of components gas in clean and shaly sands as derived from well logs:
n = Archies saturation exponent
The Log Analyst, 12, 2132.
NMR = Nuclear magnetic resonance
Gaymard, R., and A. Poupon, 1968, Response of neutron
PEF = Photoelectric factor (be)
and formation density logs in hydrocarbon bearing for-
Rt = True resistivity ( m)
mations: The Log Analyst, 9, 312.
Rw = Water resistivity ( m)
Spherically focused resistivity ( m) Heidari, Z., C. Torres-Verdn, and W. Preeg, 2012, Improved
SFLU =
SNUPAR = Schlumberger nuclear parameter calcu- estimation of mineral and fluid volumetric concentra-
lator tions in thinly bedded and invaded formations: Geophys-
Sw = Water saturation (%) ics, 77, no. 3, WA79WA98, doi: 10.1190/geo2011-0454.1.
UTAPWeLS = The University of Texas at Austin petro- Mao, Z.-Q., 2001, The physical dependence and the corre-
physical and well-log simulator lation characteristics of density and neutron logs: Pet-
Vi = Volumetric concentration (vv) rophysics, 42, 438443.

T154 Interpretation / November 2013

 McKeon, D. C., and H. D. Scott, 1989, SNUPAR A resistivity measurements. His research interests include
nuclear parameter code for nuclear geophysics applica- well-log modeling, inversion, and petrophysical interpreta-
tions: IEEE Transactions on Nuclear Science, 36, 1215 tion. He is a member of SEG, SPE, SPWLA, and IEEE.
1219, doi: 10.1109/23.34634.
Passey, Q. R., S. Creaney, J. B. Kulla, F. J. Moretti, and J. D.
Stroud, 1990, A practical model for organic richness Carlos Torres-Verdn received a
Ph.D. (1991) in engineering geosci-
from porosity and resistivity logs: AAPG Bulletin, 74,
Downloaded 05/29/14 to 128.83.167.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

ence from the University of California

17771794.
at Berkeley. During 19911997, he
Quintero, L. F., and Z. Bassiouni, 1998, Porosity determina- held the position of research scientist
tion in gas-bearing formations: Presented at SPE Permian with Schlumberger-Doll Research.
Basin Oil and Gas Recovery Conference, SPE 39774. From 19971999, he was a reservoir
Schlumberger Limited, 2009, Log interpretation charts, specialist and technology champion
http://www.slb.com/resources/publications/books/ with YPF (Buenos Aires, Argentina).
log_charts.aspx. Since 1999, he has been affiliated with the Department
Spears, R. W., 2006, Lithofacies-based corrections to den- of Petroleum and Geosystems Engineering of the Univer-
sity-neutron porosity in a high-porosity gas- and oil- sity of Texas at Austin, where he is currently a full profes-
bearing turbidite sandstone reservoir, Erha field, OPL sor, holds the Zarrow Centennial Professorship in
Petroleum Engineering, and conducts research on bore-
209, Deepwater Nigeria: Petrophysics, 47, 294305.
hole geophysics, formation evaluation, well logging, and
Voss, B., C. Torres-Verdn, A. Gandhi, G. Alabi, and M. Lem-
integrated reservoir characterization. He is the founder
kecher, 2009, Common stratigraphic framework to sim- and director of the Research Consortium on Formation
ulate well logs and to cross-validate static and dynamic Evaluation at the University of Texas at Austin, which is
petrophysical interpretation: Presented at Transactions currently sponsored by 32 companies. He has published
of the SPWLA, 50th Annual Logging Symposium. more than 115 refereed journal papers and 130 conference
Xu, C., Z. Heidari, and C. Torres-Verdn, 2012, Rock clas- papers, has served as a guest editor for Radio Science, as
sification in carbonate reservoirs based on static and an associate editor for the Journal of Electromagnetic
dynamic petrophysical properties estimated from con- Waves and Applications, SPE Journal, and Petrophysics
ventional well logs: Presented at SPE Annual Technical (SPWLA), and is currently associate editor for GEOPHYSICS
Conference and Exhibition, SPE 159991. and an editorial board member of The Leading Edge
(SEG). He is the recipient of the 2006 Distinguished Tech-
nical Achievement Award from the SPWLA, the 2008 For-
mation Evaluation Award from the SPE, the 2003, 2004,
Olabode Ijasan received a B.S. 2006, and 2007 Best Paper Awards in Petrophysics by
(2006) in electrical and electronics the SPWLA, the 2006 Best Presentation Award, and the
engineering from the University of 2007 Best Poster Award by the SPWLA.
Lagos, Nigeria, and an M.S. (2010) in
petroleum engineering from the Uni-
versity of Texas at Austin, where he
is currently a Ph.D. candidate devel- Biographies and photographs of the other authors are
oping modeling-based techniques for not available.
interpreting borehole nuclear and

Interpretation / November 2013 T155