Journal of Petroleum Exploration and Production Technology
https://doi.org/10.1007/s13202-023-01670-x
ORIGINAL PAPER - PRODUCTION GEOLOGY
Geological modeling of diagenetic logs of the Sarvak reservoir
in Dezful Embayment, southwestern Iran: implications
for geostatistical simulation and reservoir quality assessment
Vali Mehdipour1 · Ahmad Reza Rabbani1 · Ali Kadkhodaie2
Received: 20 February 2023 / Accepted: 20 June 2023
© The Author(s) 2023
Abstract
Reservoir quality in carbonate reservoirs is significantly influenced by diagenetic processes. Although diagenesis is studied
as a common reservoir quality damaging/enhancing process in many previous studies, literature is limited about the spatial
modeling of diagenesis processes using advanced geostatistical algorithms. In the current study, 3D models of the main
diagenetic processes which affect the reservoir quality of the Sarvak reservoir in an Iranian oilfield located in the north Dezful Embayment were constructed using geostatistics. According to the petrographic studies, a total of 10 microfacies were
identified. In addition, the significant diagenetic processes in this reservoir include dolomitization, cementation, dissolution, and compaction. In this study, diagenetic electrofacies were determined using the “multi-resolution graph clustering”
method based on the quantitative results of the petrographic studies. The results of spatial modeling and provided average
maps were used to investigate the lateral variation of those properties and their relationship with effective porosity. It shows
that trends of the secondary porosity and velocity deviation log (VDL) maps are generally correlatable with the effective
porosity maps confirming the impact of dissolution as a main significant diagenetic process on reservoir quality enhancement. The most impact of the dissolution on porosity is observed in Lower Sarvak-E2 zone where the correlation coefficient
is 0.75. The correlation coefficient between porosity and VDL in some zones is high indicating the effect of diagenesis on
reservoir quality as it exceeds 0.61 in Lower Sarvak-A1 zone. In the occurrence of dolomitization, it has dual constructive
and destructive effects on the reservoir quality. The most constructive and destructive effects of dolomitization were observed
in Lower Sarvak-E1 and Lower Sarvak-F zones in which the correlation coefficients were 0.476 and − 0.456, respectively.
In addition, low porosity zones are correlatable with developing cementation, stylolites, and solution seams.
Keywords Diagenesis · VDL · 3D geological model · Electrofacies · Reservoir quality
List of symbols
DT Acoustic log that displays travel time of P-waves
(µs/ft [or m])
DTf Travel time of P-waves in fluid (µs/ft [or m])
DTmat Travel time of P-waves in matrix (µs/ft [or m])
Vp Elastic body wave or sound wave in which particles oscillate in the direction the wave propagates (m/s)
* Ahmad Reza Rabbani
rabbani@aut.ac.ir
1
Department of Petroleum Engineering, Amirkabir University
of Technology, Tehran, Iran
2
Earth Sciences Department, Faculty of Natural Science,
University of Tabriz, Tabriz, Iran
Vpsyn Synthetic compressional wave velocity derived
from Wyllie equation assuming the known
porosity from neutron, sonic or neutron-density
log (m/s)
Φs Porosity calculated based on sonic log
Abbreviations
AHC Agglomerative hierarchical clustering
AI Acoustic impedance
EFAC Electrofacies
GRFS Gaussian random function simulation
LS Lower Sarvak
MRGC Multi-resolution graph-based clustering
PHIE Effective porosity
SGS Sequential Gaussian Simulation
SIS Sequential indicator simulation
SOM Self-organizing maps
SPI Secondary porosity index
13
Vol.:(0123456789)
Journal of Petroleum Exploration and Production Technology
VDL Velocity deviation log
US Upper Sarvak
Introduction
Carbonate reservoirs comprise more than half of conventional hydrocarbon reservoirs around the world (Dou et al.
2011). The Sarvak carbonate Formation (Middle Cretaceous) has considerable hydrocarbon potential (more than
20% oil-in-place of Iranian oil reserves) in the Zagros
Basin of Iran (Hajikazemi et al. 2010; Bordenave and
Hegre 2010). These types of reservoirs are highly affected
by diagenesis since it causes significant changes in different sediment properties resulting in a higher degree of
heterogeneity (Lahann and Swarbrick 2011; Martyushev
et al. 2023a, b; Rashid et al. 2022). Since diagenesis plays
a significant role in the reservoir quality of carbonate reservoirs (Moore 2001; Qureshi et al. 2023; Bilal et al. 2022;
Moforis et al. 2022; Kontakiotis et al. 2020a; Ahmad et al.
2022), its study helps to recognize petrophysical variations
and evaluate the reservoir rock quality. In addition, pore
systems in carbonate reservoirs tend to be more complicated than clastic reservoirs because of the facies diversity and the combination of several diagenetic processes
(Rahim et al. 2022; Bilal et al. 2023; Ahmad et al. 2022;
Janjuhah et al. 2021; Kontakiotis et al. 2020b). Therefore,
in carbonate reservoirs, it is necessary to consider the
control of diagenesis and depositional facies on reservoir
characterization.
The main purpose of this study is to construct a 3D
geological model indicating the role of diagenesis in reservoir quality. Due to the lack of sufficient cored intervals, the petrophysical logs after calibration with core
data are used for the diagenetic modeling because of their
continuity and availability in all wells of the area. The
diagenetic processes, such as cementation, dissolution,
compaction, dolomitization, fracturing, micritization,
recrystallization, and bioturbation, were observed based
on the petrographic studies. However, among all the above
diagenetic processes affecting the reservoir quality of the
Sarvak reservoir, dolomitization, dissolution, cementation,
and stylolite data were modeled using advanced geostatistics algorithms. Mechanical compaction has reduced the
micro-porosity in mud-supported facies, and interparticle
porosity in grain-supported facies, due to the intense packing of the grains. Cementation has reduced the porosity by
precipitation of authigenic minerals in the pore spaces of
Sarvak carbonate. Porosity is considerably enhanced by
dissolution through eliminating all or part of earlier existing minerals and reducing the pore spaces in the rocks.
13
Geological and stratigraphical settings
There are many significant hydrocarbon reservoirs of the
Cretaceous age in the Arabian Plate and the Zagros Basin
(Hollis 2011). The Zagros Basin includes both carbonate
and siliciclastic reservoirs. The late Albian, Cenomanian,
and Turonian (89–98.9 Ma) stratigraphic record in the
Arabian Plate and the Zagros Basin corresponds to the
Sarvak Formation in Iran (Motiei 1993) and is equivalent
to Natih Formation in Oman, and Mishrif, Ahmadi, and
Rumaila Formations in Saudi Arabia (Van Buchem et al.
2002). It should be noted that the equivalent of the Sarvak
Formation is Mishrif Formation in the west of Iran (Iraq)
which was investigated by some researchers (Al-Dujaili
et al. 2023a, 2023b). Sediments of the Sarvak Formation,
a member of the Bangestan Group, were deposited on a
platform and within the intrashelf basin on the passive
margin of the Arabian Plate (Martin 2001).
The studied field is located in the northern part of the
Dezful Embayment (Fig. 1). The Sarvak Formation in the
type section (Bangestan anticline) overlies the Kazhdumi
Formation and underlies the Gurpi Formation. However,
in other parts of the Zagros Basin, particularly in the Dezful Embayment, underlies the Ilam Formation (James and
Wynd 1965; Rahimpour-Bonab et al. 2012). The Sarvak
Formation has two pelagic (wackestone to packstone with
abundant planktonic foraminifera) and neritic (mudstone
to packstone with abundant benthic foraminifera) facies
(James and Wynd 1965). Reservoir characteristics of the
Sarvak Formation were influenced by some local and
regional disconformities (Rahimpour-Bonab et al. 2013).
The lithology of this formation is mainly made up of carbonate rocks (limestone and argillaceous limestone, with
some dolomite and shale, interbeds). Sarvak Formation in
the studied field is divided into three main zones including
Upper-Sarvak, Ahmadi, and Lower Sarvak (Fig. 2). Both
the Upper Sarvak and the Lower Sarvak (as the main target
reservoir) are divided into nine reservoir units (Fig. 3).
The thickness of the Sarvak reservoir in the studied field
is about 600 m based on constructed isochore map.
Materials and methods
The first phase of this study began with the thin section
study, core description, and core analysis for determining
diagenesis features. In addition, the evaluated petrophysical logs were used to construct diagenetic-related logs
(such as dissolution and dolomitization) as main input data
for the spatial diagenetic model.
Journal of Petroleum Exploration and Production Technology
Fig. 1 The main structural
zones of the Zagros and the
position of the studied zones
(After Casini et al. 2011) (the
approximate location of the
studied field is shown by the
black circle)
Fig. 2 Generalized chronostratigraphy of the Cretaceous successions in the Zagros region and nearby regions (after Sharland et al. 2001)
Petrographic study
A total of 1350 thin sections of the Sarvak Formation from
four wells were used for petrographic studies. Petrographic
studies involve identifying microfacies study and diagenetic
processes from thin-sections (from cores), discriminating
genetic diagenesis types, and evaluating their effects on
reservoir quality in 3D geological models. Description of
the microfacies was done based on Folk and Dunham (and
Embry and Klovan modification) classification schemes
(Dunham 1962; Embry and Klovan 1971; Folk 1962). Apart
from cored wells, petrophysical logs of 20 drilled wells were
used for calibration with petrographic results and creating
synthetic diagenesis logs. Four parameters including visible secondary porosity, dolomite volume fraction, cement
volume fraction, and existing stylolite, were detected and
13
Journal of Petroleum Exploration and Production Technology
Petrophysical logs interpretation
Petrophysical interpretation is a process of using petrophysical raw data (such as gamma ray, sonic log, density log,
neutron log, photoelectric factor log, and resistivity logs) for
obtaining the total and effective porosity, water saturation,
and mineral volumes through petrophysical log interpretation. Afterward, petrophysical parameters including, effective porosity, volume of shale, volume of dolomite, velocity deviation log (VDL), secondary porosity, and digenesis
electrofacies were populated in a 3D geological grid.
Secondary porosity calculation
Secondary porosity resulting from dissolution was calculated based on the difference between two porosity logs
calculated from sonic and neutron-density porosities. The
sonic-based porosity is compared with the total porosity
obtained from density and neutron logs. Therefore, the total
porosity can be subdivided into two “primary porosity”
obtained from the sonic log and “secondary porosity” calculated from the difference between the sonic porosity and
total porosity (Schlumberger 1974). Although the calculated
log is uncertain, it can be used as secondary porosity index.
VDL calculation
Fig. 3 Stratigraphic chart and reservoir zonation of the Sarvak Formation in the studied oilfield
digitized for spatial modeling. To distinguish dolomite from
calcite, all studied samples were stained with alizarin red S
solution (Dickson 1965). The percentage of the cement was
estimated based on the petrographic studies of thin section.
The VDL is a synthetic log obtained from applying an average time equation addressed by Wyllie et al. (1956). In this
study, the VDL log was calculated for the Sarvak reservoir
using the sonic log and calculated petrophysical porosity.
In this regard, it can be stated that the synthetic VDL log
is dependent on the lithological characteristics, sedimentary textures, diagenetic features, and pore systems of the
Sarvak reservoir. For this purpose, the synthetic travel time
DTsyn, was derived by calculating through re-ordering the
Fig. 4 Histograms of normalized input data along with matrix plot showing the relationship between input logs in different clusters
13
Journal of Petroleum Exploration and Production Technology
Fig. 5 Histograms of the final diagenesis electrofacies
Table 1 Statistical properties of eight electrofacies clusters in the
Sarvak reservoir of the studied oilfield; SPI, dolomite volume, and
PHIE units are fractions
Facies
Weight
VDL
SPI
Vol-Dolomite
PHIE
Facies1
Facies2
Facies3
Facies4
Facies5
Facies6
Facies7
Facies8
1006
2212
3148
813
2023
2308
2510
2398
494.49
117.16
123.80
− 608.56
− 96.94
129.74
375.91
646.09
0.04
0.01
0.00
0.00
0.00
0.01
0.02
0.05
0.49
0.35
0.00
0.00
0.00
0.00
0.01
0.10
0.07
0.02
0.00
0.01
0.04
0.09
0.03
0.09
where VDL is velocity deviation log, Vp, is the real velocity,
and Vpsyn is the calculating velocity.
The pore types can be divided into three categories based
on the velocity deviation log including positive, zero, and
negative deviations (Anselmetti and Eberli 1999; Wang and
Nur 1990). Fracturing, caving, anomalies of the borehole,
and a high percentage of the free gas can cause a negative
VDL value (Nur and Simmons 1969). Positive VDL indicates zones having velocity higher than expected, such as
zones in which moldic and vuggy pores are developed. Zero
VDL represents intervals having dominant micro-porosity
or interparticle porosity.
Diagenetic electrofacies analysis
Wyllie formula to calculate synthetic velocity. The difference between the two porosity logs (the real sonic log
(DTreal) with the artificial sonic log ( DTsyn)) is considered a
VDL log (Anselmetti and Eberli 1999). Therefore, the velocity deviation was simply obtained through the calculating
difference between Vpreal and Vpsyn.
ΦS =
DTlog − DTmat
DTf − DTmat
(1)
where Φs in secondary porosity, DT log is the sonic log,
DTmat, is the sonic value of the matrix, and D
Tf is the sonic
value of the fluid.
VDL = Vp − Vpsyn
(2)
After petrophysical interpretation and calculating the VDL
and secondary porosity, diagenetic electrofacies were calculated. Electrofacies, as introduced by Serra and Abbot,
are defined as the set of petrophysical well log responses,
which characterizes a stratum that can be distinguished from
the adjacent ones (Serra and Abbot 1982). In this regard,
various data clustering approaches including multi-resolution graph-based clustering (MRGC), self-organizing maps
(SOM), and AHC (agglomerative hierarchical clustering)
were employed. Examining the results revealed that the
MRGC method outperforms other clustering algorithms for
the extraction of diagenetic electrofacies. It is a multi-dimensional algorithm based on the non-parametric and K-nearest
neighbor method with graph data representation (Ye and
Rabiller 2000). MRGC method gives classes based on the
13
Journal of Petroleum Exploration and Production Technology
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
Fig. 6 Thin-section photos of the facies in Well-Z8: facies1 (4060.37 m), facies 2 (4076.67 m), facies 3 (4093.94 m), facies 4 (4102.43 m), facies
5 (4106.10 m), facies 6 (4126.63 m.), facies 7 (4312.68 m), facies 8 (4470.15 m), facies 9 (4441.17 m), and facies 10 (4510.83 m)
distribution of natural points with various scales (Asgari and
Sobhi 2006). It allows the user to gain the ideal cluster numbers for defining electrofacies clusters from conventional
petrophysical logs. Several diagenetic-related logs (dolomite
volume, secondary porosity, and VDL logs) were chosen
for electrofacies analysis as input data. An unsupervised
13
algorithm has been applied after training and normalizing
the input data. The optimum number of diagenetic clusters
was identified after considering different clusters suggested
via the MRGC algorithm and then similar clusters were
merged to find the optimum number of clusters. Therefore,
after considering the statistical distribution input parameters
Journal of Petroleum Exploration and Production Technology
Table 2 Properties of ten facies classes in Sarvak reservoir of the studied oilfield
Facies code Facies
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
Sub environment
Dunham (1962), Embry and Klovan (1971)
Folk (1959, 1962)
Wackestone to packstone with abundant benthic foraminifera
Floatstone to rudstone with abundant benthic
foraminifera
Grainstone with abundant bioclast and peloid
Packstone (dolomitized)
Wackestone with sponge spicules
Packstone with abundant benthic foraminifera
Wackestone to packstone with abundant planktonic foraminifera
Wackestone to packstone with abundant
foraminifera
Floatstone to rudstone with abundant rudist
and echinoid
Grainstone with abundant rudist debris
Pel-biomicrite with abundant benthic
foraminifera
Pel- biomicrite with abundant benthic
foraminifera and rudist
Bio-pelsparite
Biosparite (dolomitized)
Biomicrite with sponge spicules
Biomicrite with abundant benthic foraminifera
Biomicrite with abundant planktonic
foraminifera
Biomicrite with abundant foraminifera
Proximal inner shelf
Biomicrite with abundant rudist and echinoid
Middle shelf
Pel-biosparite with abundant rudist debris
Distal inner shelf to middle shelf
Fig. 7 Histogram of ten facies in Sarvak reservoir of the studied oilfield
including secondary porosity, effective porosity, dolomitization, and VDL, a total of 20 electrofacies were extracted
for the Sarvak reservoir (Fig. 4). After checking with sedimentological facies and petrophysical properties, 20 clusters were merged until the final number of clusters reached
eight electrofacies (Fig. 5). Table 1 outlines the geological/
petrophysical properties of the final diagenetic electrofacies.
Geological modeling
The 3D geological modeling of the Sarvak reservoir was
started with structural modeling. The structural model captures the structural elements such as interpreted horizons,
faults, and reservoir zones. Essentially, the construction of
the structural model includes importing and defining the
Distal inner shelf
Distal inner shelf to middle shelf
Proximal middle shelf
Outer shelf
Proximal inner shelf
Outer shelf
Distal inner shelf
depth maps and considering the faults affecting the reservoir. The structural model was created using the corner
point approach. A 3D fine geological grid was constructed
with 100*100 m cell increments considering the area of the
project. The horizon-making is the first step in defining the
vertical layering for the 3D grid in the geological model.
Afterward, all interpreted horizons in the depth domain
were imported into the model. Three main seismic interpreted depth surfaces were used for horizon-making in the
model, including Sarvak, Ahmadi, and Kazhdumi surfaces.
Isochore maps of each subzone were prepared using well top
data. The layering was done as the final step of framework
construction. The layering of the constructed model (proportional method) was designed according to the thickness
analysis and vertical variogram analysis of reservoir facies.
Accordingly, the cell thickness of pay zones was assigned
as 0.5–1.5 m. Thus, the vertical changes of the reservoir
changes will be captured and the number of whole cells will
be optimized. The final layering framework of the model
consists of 647 layers with 7,612,602 total cells.
Application of geostatistical modeling methods
Conditional simulation [such as SGS (Sequential Gaussian Simulation) and GRFS (Gaussian Random Function
Simulation)] and deterministic algorithms such as ordinary
kriging were applied for the porosity modeling. Kriging
was originally considered for property estimation based on
D.G. Krige’s idea (Krige 1951; Lake et al. 2007). Kriging
and cokriging interpolation methods as forms of univariate and multivariate linear regression-based models are
13
Journal of Petroleum Exploration and Production Technology
Fig. 8 Partially and complete pore-filling cement (calcite and dolomite) of the Sarvak Formation in Well Z8. A Depth: 4438.70 m, xpl B Depth:
4502.18 m, ppl; C Depth: 4063.11 m, D Depth: 4446.52 m
usually used for property distribution in static modeling.
There are different kriging types such as ordinary kriging, simple kriging, indicator kriging, universal kriging,
probability kriging, cokriging, etc. (Basbug and Karpyn
2007). In this study, both simple and ordinary kriging were
used for initial property estimation. Gaussian simulation
methods such as SGS and SIS are conditional geostatistical methods utilizing the kriging mean and variance to
produce a Gaussian distribution of parameters (Evans
Annan et al. 2019). All these stochastic methods provide
more reliable results, which can improve the quality of
the predictions (Zhao et al. 2014). In complex reservoirs
such as the Sarvak Formation, the deterministic kriging
methods may not represent an accurate level of heterogeneities, while the stochastic simulation methods can
provide more reliable results (Sahin and Al-Salem 2001;
Zhao et al. 2014).
13
Results
In this study, after determining the major facies (according to Flügel 2010) and main diagenetic features based on
petrographic studies, they were correlated with petrophysical logs to provide necessary diagenetic input data for 3D
geological modeling.
Microfacies analysis
The microfacies analysis of the carbonate rocks was
defined according to Dunham’s (1962) classification
scheme, which was later modified by Flügel (2010).
Microfacies analysis of the Sarvak reservoir was done
based on the observed sedimentological characteristics
during the thin-section study, and through comparison
with additional data obtained from the Sarvak reservoir in
Journal of Petroleum Exploration and Production Technology
Fig. 9 Physical compaction between bioclastic grains. A Depth:
4474.76 m, ppl; Chemical compaction (stylolite) into the limestone
parts of Sarvak Formation. B Depth: 4432.33 m, ppl; Dolomite with
abundant inter-crystalline porosity in upper Sarvak Formation of Well
Z8 C Depth: 4120.83 m, ppl; and D Depth: 4098.25 m, pp
nearby fields and also from the literature (Kontakiotis et al.
2020a, b; Bilal et al. 2022; Ali et al. 2021; Mehmood et al.
2023). Therefore, a total of 10 facies were distinguished
in the Sarvak Formation coded as F1–F10 (Fig. 6). The
brief properties of each facies are summarized in Table 2.
In classical facies models, a carbonate shelf is separated
into an inner shelf, middle shelf, and outer shelf (Burchette and Wright 1992; Kontakiotis et al. 2020a; Flügel
and Munnecke 2010). The Sarvak cores in the studied
wells contain shallow (inner shelf) to deep water (middle
to outer shelf) carbonates. It shows open marine, patch
reef, and lagoon sub-environments indicating an extended,
carbonate shelf platform. The inner shelf of the Sarvak
Formation includes the lagoon sub-environment, while the
middle shelf includes the margin platform and reef barrier. The outer shelf environment was recognized in the
studied cores of the Sarvak Formation based on abundant
planktonic foraminifera (belonging to the lower parts of
the upper Sarvak Formation). Based on this study, Biomicrite with abundant benthic foraminifera facies (F6) are the
shallowest facies, then toward the lagoon and middle parts
of the inner shelf the three facies distinguished including Pel-biomicrite with abundant benthic foraminifera
(F1), Pel-biomicrite with abundant benthic foraminifera
and rudist (F2), and Biomicrite with abundant foraminifera (F8). Toward the open marine, three facies have been
determined as platform margins containing: Bio-pelsparite
(F3), Biomicrite with abundant rudist and echinoid (F9,)
and Pel-biosparite with abundant rudist debris (F10). Figure 7 indicates histograms of the mentioned ten facies in
the Sarvak reservoir from the studied oilfield.
Diagenetic features in the Sarvak reservoir
According to the thin-section studies of four cored wells in
this field, a widespread variety of diagenetic processes were
determined in the studied Sarvak Formation. However, four
diagenetic processes, comprising cementation, compaction,
dissolution, and dolomitization show significant impacts on
reservoir quality.
Cementation
Various types of cement are developed in the Sarvak carbonate Formation. Equant, drusy mosaic, and blocky cement
are the most dominant cement in the studied core samples.
Based on the petrographic description, two main types of
cement (calcite and dolomite) were observed in this reservoir (Fig. 8).
13
Journal of Petroleum Exploration and Production Technology
Fig. 10 Plot showing the layout of diagenesis logs in the Sarvak reservoir. The tracks from left to right indicate the effective porosity log
(PHIE), cement log (based on the petrophysical log), SPI log (secondary porosity based on petrophysical interpretation), SPI-index model
(secondary porosity based on petrographic study), lithology (limestone, dolomite, and shale) VDL (velocity deviation log), stylolite
index log (based on petrographic study) and diagenesis electrofacies
(EFAC)
Dolomitization
lot of the dissolution pores filled by dolomitic cement. Scattered, unimodal, fine-to-medium-grained dolomite rhombs
formed within the mud-supported facies (bioclastic wackestones) in both shallow and deep-water environments, and
its heterogeneous nature is associated with dissolution seams
and stylolite surfaces. Their occurrence along stylolite and
In general, the dolomitization process has dual constructive and destructive effects on the reservoir quality. In the
uppermost parts of the upper Sarvak, there is a dolomitic
zone showing abundant inter-crystalline porosity and also a
13
Journal of Petroleum Exploration and Production Technology
Fig. 11 Workflow of diagenesis parameters determination, and 3D geological modeling in Sarvak reservoir of the studied field
solution-seams surfaces indicate a diagenetic origin, particularly during shallow burial diagenesis. In the nearby area,
they are laterally equivalent facies, in which dolomitization is related to the late diagenetic and fracture-controlled
diagenesis process (Sharp et al. 2010). The dolomite crystals
are seen in the form of fine crystalline, subhedral to euhedral dolomite rhombs (Fig. 9). However, the saddle coarse
dolomite crystals belonging to the deep-burial diagenesis
(Sarg 1988) were not recognized in the studied intervals.
The percentage of the dolomite minerals in the thin-section
study was estimated visually for generating related logs for
spatial modeling.
Compaction
Chemical compaction was observed in solution seams and
stylolites. Most of the solution seams and stylolites were
oil-stained. They are observed in both outer-shelf and innershelf facies with abundant micrite. In addition, the solution
seams and stylolite extend the dolomitization process into
the upper and lower Sarvak Formation. It means that in the
upper and lower Sarvak Formation, as observed in Well
Z8, the dolomitization process is related to the stylolites
(Fig. 9). The dolomite crystals formed after stylolitization.
The existence of the stylolites was detected and converted
to a discrete log for 3D modeling.
Dissolution
The dissolution process is considered the main significant
diagenetic phenomenon in this reservoir, improving the
reservoir quality. Generally, it is an essential mechanism in
developing secondary porosity, including moldic, vuggy, and
interparticle porosity (Janjuhah et al. 2021). Dissolution of
skeletal grains produced abundant moldic and vuggy pores
mostly into the shallow water facies (particularly in rudistbearing facies) belonging to the lower Sarvak. In the shallow water facies of the upper Sarvak particularly in Well
Z8, moldic pores are the common pore type having poor
reservoir quality due to the disconnected nature and occurring later cementation (Fig. 9). The best and high porosity
belongs to the top of the lower Sarvak (especially related
to the rudist-bearing facies) due to developing interparticle, moldic, and vuggy porosities. Moreover, karstification
features have also been developed in some intervals. The
feature is recognizable over most of the middle Cretaceous
carbonate rocks all over the Middle East (Sharland et al.
2001). The percentage of the secondary porosity (moldic
and vuggy), which is visible from the thin-section study, was
estimated visually for generating diagenetic logs for spatial
modeling.
Diagenetic petrophysical logs
As discussed in “Materials and methods” section, VDL and
secondary porosity logs were calculated to have complete
diagenesis logs for all drilled wells having petrophysical
logs. Moreover, diagenesis electrofacies logs were prepared
to distinguish the different clusters from, the diagenesis
point of view. Those logs, along with effective porosity log
(PHIE), shale, and dolomite volume logs from petrophysical
interpretation were imported into the 3D geological model.
13
Journal of Petroleum Exploration and Production Technology
3D diagenetic parameter modeling
The aim of conventional 3D geological reservoir modeling
as an essential stage in field development is to get a 3D spatial distribution of reservoir types, reservoir property, and
reservoir anisotropy via integrating all available data and
geological concepts (Branets et al. 2009). Generally, 3D geological modeling includes two structural and property modeling. The main purpose of property modeling is to propagate
reservoir properties between the wells matching the well
data, which can be done after structural modeling. Property
modeling was done based on stochastic methods producing more realizations, while deterministic approaches (such
as Kriging) resulted in one realization. Generally, property
modeling includes facies (discrete) and petrophysical modeling (continuous). Therefore, all the discrete properties were
populated using facies modeling and applying “Sequential
Indicator Simulation” (SIS). Continuous ones (such as
Fig. 12 Cross-section view of effective porosity and VDL models with related histograms (C: Histogram of the effective porosity, D: Histogram
of the VDL)
13
Journal of Petroleum Exploration and Production Technology
Fig. 13 Cross plot between effective porosity and VDL (A), SPI-log (B), dolomite volume (C), and cement index (D)
porosity) were propagated using the stochastic approach of
“Gaussian Random Function Simulation” (GRFS) via the
“petrophysical modeling” module.
Generally, the routine geological reservoir modeling
workflow includes integrating seismic data, and geological and petrophysical log data to construct the geological
model. Understating the diagenesis phenomena from petrophysical logs and petrographic study has uncertainty which
needs some trend maps and secondary variables obtained
from seismic cubes. Therefore, due to the uncertainty in
the diagenesis modeling, seismic-derived trend maps and
secondary variable (AI) with stochastic propagation methods have been applied. Applying geostatistical methods for
diagenetic modeling has been investigated by some authors
(Doligez et al. 2011; Barbier et al. 2012).
After structural modeling, provided petrophysical diagenetic logs (secondary porosity, effective porosity, VDL, diagenetic electrofacies, dolomite volume, and shale volume),
along with petrographic-based index logs were scaled up
after detail analyzing for selecting the best methods (Kadkhodaie-Ilkhchi et al. 2019). The “Most of” method was
used for scaling up the discrete logs (diagenetic electrofacies and stylolite index), and the “Median” or “Arithmetic” methods were selected for other continuous logs after
comparing the resulting histograms. Figure 10 shows the
layout of different input diagenetic logs considered for 3D
modeling. The data analysis process is the further stage after
scaling up, which is used in preparing the data and controlling data quality for geostatistics distribution (Iske and
Randen 2005). Therefore, all scaled-up data were analyzed
via the “data analysis” module for data transformation (truncating abnormal values and normalizing the distributions)
and variography. Data proportion analysis and variography
were made by facies and by zones. The variogram analysis
is a critical process for geostatistical studies. Data variability
between the wells for each zone was considered individually,
by the variogram analysis module in three main directions
(vertical, major, and minor). The major direction used in
variography was obtained from the variogram map created
based on the variance of the acoustic impedance. Figure 11
illustrates the workflow for 3D modeling of different diagenetic parameters.
PHIE modeling
Effective porosity distribution was obtained after scaling up
the effective porosity log by applying the “median” method
and variography analysis. After that, it was distributed
using the GRFS algorithm and co-kriging with the available
13
Journal of Petroleum Exploration and Production Technology
Fig. 14 Cross-section view of shale volume, and cementation models with related histograms (C: Histogram of the shale volume, D: Histogram
of the cementation)
acoustic impedance cube. Because of the heterogony effect
on porosity distribution (Grader et al. 2009), seismic data
is necessary to convert the heterogeneous reservoir into
more homogeneous reservoirs. The correlation coefficient
between them is 64%. The 3D model of effective porosity in
the Sarvak reservoir in cross-section view with a related histogram of the well log, scaled-up log, and model are shown
in Fig. 12.
13
VDL modeling
VDL log was propagated after scaling up and variography analysis. GRFS algorithm was applied for distribution through co-kriging with an effective porosity model
as secondary data. Figure 12 indicates a cross-section view
of the VDL model and histogram of the original well log,
scaled-up log, and model log. After constructing the VDL
Journal of Petroleum Exploration and Production Technology
Table 3 Effective porosity and
correlation coefficient between
effective porosity and shale
volume, dolomite volume,
secondary porosity index,
and VDL in reservoir zones
of Sarvak reservoir (negative
correlation coefficient signifies
inverse correlation)
Zones
PHIE (%) Volume (%)
The correlation coefficient between
porosity and these parameters
Shale Dolomite Stylolite Cement Shale
Upper Sarvak
USB1
USC2b
USC1a
LSA1
LSC-D
LSE1
LSE2
LSF
LSG
3.70
5.40
6.50
4.00
8.00
9.00
3.50
4.30
7.00
2.50
3.330
4.410
2.200
1.700
2.800
2.730
2.800
2.200
1.680
2.400
model, the VDL facies model was constructed using the
classification of VDL ranges (code1 (VDL < − 500), code
2 (− 500 < VDL < 0), code 3(0 < VDL < 500), and code4
(VDL > 500)) as shown in Fig. 17.
Cement‑volume modeling
The cement model was propagated after scaling up and
variography analysis. GRFS algorithm along with PHIE as
secondary data was considered for co-kriging and propagation. There is an inverse relationship between porosity and
cement logs (Fig. 13). Figure 14 shows a cross-section view
of the cement model and related histogram of the original
well log, scaled-up log, and model log.
Shale volume modeling
Although shale is supposed to be fine-grained and clastic
sediment, it was distributed in the 3D geological model to
check the effect of primary depositional factors on the reservoir quality while investigating the impact of diagenesis.
GRFS method was used for shale volume distribution by
considering the effective porosity model as secondary data
through co-kriging. The correlation coefficient between the
shale volume and effective porosity is reversed in many
zones, indicating the negative impact of the shale volume on
reservoir quality (Table 3). Figure 14 shows a cross-section
view of the shale volume model and related histogram of the
original well log, scaled-up log, and model log.
Secondary porosity index (SPI) modeling
Secondary porosity log propagation (from petrophysical
interpretation) was done after scaling up and variography analysis. GRFS method along with co-kriging with
8.500
41.900
3.700
11.7
2.300
4.390
8.420
12.710
11.980
10.250
31.000
47.000
34.000
57.000
42.000
22.000
14.000
43.000
26.000
43.000
6.400
10.000
6.300
3.000
4.500
8.700
5.590
5.400
5.800
3.770
Dolomite Secondary
porosity
0.140
0.100
− 0.045
0.330
− 0.090 − 0.276
− 0.099
0.010
− 0.050
0.084
− 0.144 − 0.118
− 0.290
0.476
− 0.158
0.154
− 0.060 − 0.456
− 0.300
0.100
0.350
0.190
0.677
0.348
0.690
0.580
0.620
0.750
0.137
0.415
VDL
0.560
0.290
0.600
0.445
0.610
0.547
0.540
0.420
− 0.019
0.065
effective porosity log was used for SPI distribution. The
correlation between effective and secondary porosity is
73%, therefore porosity model has been used as secondary data for propagation via co-kriging. In addition, the
secondary porosity log (SPI-index) which was provided
based on visual petrographic study, was modeled using
the same procedure. The correlation coefficient between
both SPI data is 43% indicating that the SPI data from
petrophysical interpretation has different resolution and
accuracy. Figure 15 shows the cross-section view of both
secondary porosity models and histograms of the original
well logs, scaled-up logs, and model logs.
Dolomite volume modeling
Two dolomite volume logs derived from the petrophysical
interpretation and petrographic study were spatially modeled. VDL model was used as secondary data for propagation through co-kriging and GRFS algorithm. The correlation coefficient between both volumes is about 40%,
indicating the different resolutions of both data. Since the
dolomite volume from petrophysical interpretation is more
accurate and continuous than the petrographic study, it was
selected as a final dolomite volume property. Figure 16
shows a cross-section view of both dolomite volume models and related histograms of the original secondary porosity
logs, scaled-up logs, and model logs.
Stylolite index modeling
The stylolite index, which was provided from a visual petrographic study was distributed using the “SIS” algorithm.
Due to the low density of the stylolite index, the model’s
accuracy is low. Figure 17 indicates a cross-section view of
13
Journal of Petroleum Exploration and Production Technology
Fig. 15 Cross-section view of A SPI-log model (based on petrophysical interpretation), B SPI-index model (secondary porosity based on petrographic study and related histograms) C Histogram of petrophysical based-SPI model, D Histogram of the petrographic based-SPI-index model)
the stylolite index model and histogram of the original well
log, scaled-up log, and model log.
Diagenetic electrofacies modeling
The diagenetic electrofacies log (EFAC) was distributed in
the 3D model after scaling up and variography analysis. It
13
was populated in the geological grid using the “Facies modeling” module and applying the SIS algorithm individually
for each facies code. Figure 18 displays the cross-section
view of the diagenetic electrofacies model and related histogram of the original well log, scaled-up log, and model log.
Journal of Petroleum Exploration and Production Technology
Fig. 16 Cross-section view of: A dolomite volume model (based on
the petrophysical log), B dolomite volume index model (based on petrographic study) with related histograms) C histogram of petrophysi-
cal based-dolomite model, D histogram of the petrographic baseddolomite model)
Discussion
between provided maps was investigated in ten pay zones
from the diagenetic point of view (Table 3). To understand
After constructing the 3D models for all properties, the
average maps of each zone were produced. The relationship
13
Journal of Petroleum Exploration and Production Technology
Fig. 17 Cross-section view of discrete models, including the stylolite index model (based on petrographic study), and VDL facies model with
related histograms (C: Histogram of stylolite index model, D: Histogram of the VDL facies model)
how diagenetic phenomena control the reservoir quality, all
related maps were provided and compared together in all
ten reservoir zones.
In the Upper Sarvak zone (USA), according to the effective porosity map (average porosity is 3.7%), the northeast
of the field has good reservoir quality (Fig. 19). In poor
13
porosity areas, the shale volume fairly increases. Dissolution is the most effective diagenesis phenomenon, influencing the reservoir quality (correlation coefficient is 0.35).
The dolomite and cement volumes are low (less than 10%).
However, dolomitization has a positive e