Fernandez-Ibanezetal.2022 AAPG XSPerm

Excess permeability in the AUTHORS
Fermın Fernandez-Iban ~ ez ~
Brazil pre-Salt: Nonmatrix ExxonMobil Upstream Research Company,
Spring, Texas; fermin.fernandez.ibanez@
types, concepts, diagnostic exxonmobil.com
Fermın Fernandez-Iba~nez is a geoscientist at
indicators, and reservoir ExxonMobil Upstream Research Company.
He specializes in subsurface data integration,
implications reservoir characterization, and modeling to
assess the impact of permeability extremes
Fermın Ferna ndez-Iba
n~ ez, Gareth D. Jones, on reservoir performance. Fermın has 14
Jordan G. Mimoun, Melanie G. Bowen, years of experience working assets in Asia,
J. A. (Toni) Simo, Virginia Marcon, and Australia-Papua New Guinea, and South
America. He received his Ph.D. from Granada
William L. Esch
University (Spain) before working for
GeoMechanics International and ExxonMobil.
He is the corresponding author for this paper.
ABSTRACT Gareth D. Jones ~ ExxonMobil
Excess permeability in carbonate reservoirs is typically related to Upstream Oil and Gas, Spring, Texas;
the presence of dual-porosity multiscale nonmatrix features that [email protected]
can result in early water production or breakthrough of injected Gareth D. Jones is the Brazil subsurface
fluids. Therefore, early recognition of nonmatrix deeply impacts technology manager at ExxonMobil. He has
field development strategies. We use publicly available data more than 20 years of global experience in
from the pre-Salt of Brazil to screen for nonmatrix presence and characterizing and modeling carbonate
reservoirs, with assignments (technical and
dynamic potential. We integrate static (core and wire-line logs)
management) in exploration, development,
and dynamic (drilling losses and well tests) data sets using a con- production, technology, and strategic
sistent workflow to investigate the types and distribution of non- planning. He has a Ph.D. in fluid flow and
matrix excess permeability away from wellbore control. carbonate diagenesis (University of Bristol)
Observations across multiple data sets reveal a multiscale, non- and an M.Sc. in hydrogeology (University of
matrix pore system that is identifiable and important to include Birmingham).
in geologic models when considering reservoir performance. The Jordan G. Mimoun ~ ExxonMobil
main nonmatrix components include faults and fractures, perva- Upstream Integrated Solutions, Spring,
sive networks of touching vugs, centimeter-scale vugs, and caves. Texas; [email protected]
Lost circulation zones (LCZ) encountered while drilling through Jordan G. Mimoun is the principal reservoir
nonmatrix features suggest they are not only open but connected engineer for well testing at ExxonMobil. He has
and with considerable flow capacity. The LCZ dynamic poten- 11 years of experience working assets
tial is further supported by excess permeability observations (up worldwide, from exploration to production. He
to two orders of magnitude greater than matrix) from well tests is a member of the Society of Petroleum
and a set of characteristic inflections in the pressure transients. Engineers (SPE) Journal of Petroleum
The greatest uncertainty remains on the magnitude and extent Technology editorial board and served as
chairperson of the SPE Applied Technology
of nonmatrix away from the borehole. However, a process-
Workshop on well testing in 2017. Jordan
based approach to nonmatrix characterization in combination holds an M.S. degree in petroleum engineering
from The University of Texas at Austin.
Published by the American Association of Petroleum Geologists. Copyright ©2022 ExxonMobil Melanie G. Bowen ~ ExxonMobil
Upstream Research Company. All rights reserved.
Upstream Integrated Solutions, Spring,
Manuscript received October 15, 2020; provisional acceptance December 16, 2020; revised manuscript
Texas; [email protected]
received January 15, 2021; revised provisional manuscript acceptance March 23, 2021; 2nd revised
manuscript received March 29, 2021; final acceptance May 19, 2021.
DOI:10.1306/10042120171
AAPG Bulletin, v. 106, no. 4 (April 2022), pp. 701–738 701

Melanie G. Bowen is a petroleum geologist with the observations presented here provides an opportunity to
with 2 years of development and production define geologically plausible scenarios for nonmatrix types, dis-
experience, with a primary focus on tribution, and properties. Geologically and dynamically con-
carbonate reservoir characterization. She strained conceptual models can then be used for scenario-based
worked the Brazilian pre-Salt oil field,
uncertainty analysis in a play with limited long-term production
Bacalhau (formerly Carcara), and currently
works a carbonate asset in the Middle East. history.
Melanie joined ExxonMobil in 2018 after
completing her B.Sc. at Texas A&M
University. INTRODUCTION
J. A. (Toni) Simo ~ ExxonMobil Carbonate pore systems often exhibit a dual porosity system that
Upstream Business Development, Spring, can be conceptually divided into matrix and nonmatrix compo-
Texas; [email protected]
nents. The matrix component consists of relatively small pores
J. A. (Toni) Simo is a stratigrapher at (radii <0.0625 mm and median pore throats 0.1 to 30 mm [Cho-
ExxonMobil with specialty in carbonate quette and Pray, 1970; Fullmer et al., 2019; and references
sedimentology and diagenesis. Toni is a
within]), distributed somewhat homogeneously and controlled
senior geoscience advisor, integrating
by depositional facies and pervasive diagenesis. We define the
subsurface data and concepts. Currently he
is exploring the pre-Salt of Brazil, and he “nonmatrix” component of a reservoir as a network of non-fabric-
previously worked in Europe, Asia, and selective pores, typically much larger than the matrix pores, such
south Central America. Prior to joining the as fractures and dissolution-enhanced (karst) vugs and caves, dis-
industry 15 years ago, Toni was a professor tributed sporadically and controlled by localized mechanical and
at the University of Wisconsin-Madison. chemical processes (Figure 1A). A common subsurface uncer-
Virginia Marcon ~ ExxonMobil tainty for predicting dynamic performance in carbonate reservoirs
Upstream Business Development, Spring, and aquifers is the presence of nonmatrix features (fractures and
Texas; virginia.m.marcon@exxonmobil. karst) that can result in the phenomena of “excess permeability.”
com Yose et al. (2001) define excess permeability in carbonates as the
Virginia Marcon is a geochemist at portion of effective permeability calculated from dynamic data,
ExxonMobil Upstream Business typically well tests (WTs) (e.g., drill-stem tests or production
Development Company with experience in tests), in addition to production and injection data that exceed
evaluating water-rock interactions. Her the permeability measured from routine core analysis (RCA).
expertise ranges from processes at the Nonmatrix-related excess permeability in reservoirs can be sev-
surface to the shallow subsurface. Her
eral orders of magnitude greater than RCA permeability depend-
current focus is in understanding early and
ing on the type, abundance, and distribution of nonmatrix
late pre-Salt Brazil diagenetic processes. She
received her B.Sc. and M.Sc. from the features (Figure 1B). In reservoirs with a low matrix permeability,
University of Wyoming and Ph.D. from The excess permeability associated with nonmatrix features is often
Pennsylvania State University. critical for economic production (Aguilera, 1982; Bourbiaux,
2010; Sun et al., 2019). However, excess permeability can also
William L. Esch ~ ExxonMobil
result in early breakthrough of injected fluids and poor sweep effi-
Upstream Integrated Solutions, Spring,
Texas; [email protected] ciency that negatively impacts ultimate recovery (Heubeck et al.,
2004; Trice, 2005; Warrlich et al., 2019; Ghosh et al., 2020).
William L. Esch retired as a senior
Failure to recognize or reliably characterize nonmatrix features
subsurface geology advisor at the
ExxonMobil Upstream Integrated Solutions during field appraisal and development is a frequent cause of
Company in 2020. He previously worked in “nonperformance” in carbonate fields that significantly deviate
the ExxonMobil Exploration and Upstream above or below plan for reserves or production.
Research Companies, starting his career in Since the discovery of the Tupi field in 2006, the ultra-deep-
the Exxon Production Research Company. water pre-Salt lacustrine carbonate reservoir play located offshore
His expertise is in low-temperature Brazil has proved to be world class in terms of both resource
geochemistry and predictive diagenesis. He endowment and production (Salom~ao et al., 2015; Rosa et al.,
2018). Reservoir quality is widely recognized as a key subsurface
702 Excess Permeability in the Brazil Pre-Salt

uncertainty in this play (Moczydlower et al., 2012). Compared to received a Ph.D. in geology from Louisiana
their marine counterparts, predictive models of lacustrine carbon- State University in 1995 and a B.A. from the
ate reservoirs are considerably less mature. To address this busi- University of Colorado in 1989.
ness imperative, industry and academia are actively studying the
ACKNOWLEDGMENTS
pre-Salt lacustrine carbonates with a goal of improving reservoir
quality prediction. Recently published pre-Salt studies have We thank ExxonMobil Upstream Research
focused on classification schemes and developing depositional Company for permission to release this
and diagenetic models that can be used for derisking reservoir work. The authors thank P.J. Moore, A.
quality (Muniz and Bosence, 2015; Rezende and Pope, 2015; Nolting, and S. Fullmer for their insightful
discussions while developing some of the
Wright and Barnett, 2015; Herlinger et al., 2017; Farias et al.,
methodologies applied in this paper. We
2019; Lima and De Ros, 2019; Gomes et al., 2020). These studies thank Matt Musick and Eric Bunge for
primarily focus on characterization of the reservoir matrix. By compiling, digitizing, and managing the well
contrast, studies that describe the variety of nonmatrix features test and lost circulation zones databases. We
and their impacts on flow in pre-Salt reservoirs are rare. This rep- also thank Sohan Ahmed for conducting
resents a knowledge gap since nonmatrix features (e.g., fractures), some of the pressure transient analyses and
were specifically identified as a key uncertainty for pre-Salt field Adolfo Enciso for drafting maps. The
development almost a decade ago (Moczydlower et al., 2012; manuscript benefited from the thorough and
constructive reviews of Bj€orn Seyfang, Benoit
Eleson et al., 2014). A notable exception is the paper by de Jesus
Vincent, Ted Playton, Kazuyuki Yamamoto,
et al. (2016), which presents unequivocal examples of karst and AAPG Editor Robert K. Merrill. We thank
(caves and breccia), fractures, and solution enhanced fractures them for their time and suggestions that
evident in borehole image logs from four unidentified pre-Salt helped improve the final version of this
wells that exhibit excess permeability. In a more recent study, the paper. Seismic data shown in this
abundance of open fractures was sufficient to require a discrete manuscript are courtesy of CGG. Wellbore
fracture network model as input to a dual porosity dual perme- data used in this study are courtesy of
ability (DPDK) reservoir simulation to match production history Agencia Nacional do Petroleo. The bulk of
in a pre-Salt field (Correa et al., 2019). this manuscript was written during the
lockdown caused by COVID-19,
The purpose of this paper is to address this reservoir charac-
demonstrating that remote collaboration and
terization imbalance by further investigating the type and abun- integration across disciplines is possible. The
dance of different nonmatrix features that can lead to excess authors thank ExxonMobil for providing the
permeability in the pre-Salt reservoirs of Brazil. The basis for this means to successfully do so.
study is a comprehensive, multifaceted subsurface data set that
includes WTs, mud losses during drilling (lost circulation zones,
[LCZs]), borehole image logs, and conventional core from multi-
ple pre-Salt fields in Brazil. We introduce a four-step nonlinear
multidisciplinary integrated workflow for the identification and
characterization of nonmatrix features that we have previously
applied in several other nonlacustrine carbonate reservoirs (e.g.,
Ibrayev et al., 2016; Ahdyar et al., 2019). Results are reported for
each of the workflow steps related to specific data types to under-
stand diagnostic indicators and identify the advantages and limita-
tions of the different types of data to characterize the variety of
nonmatrix features. A pre-Salt well example is used to demon-
strate the application of this integrated nonmatrix characteriza-
tion workflow. A set of preliminary conceptual geologic models
for different nonmatrix features in the pre-Salt based on our char-
acterization results is introduced. Finally, based on our characteri-
zation findings, we broadly discuss implications of the presence of
nonmatrix features with excess permeability as a key pre-Salt
FERNÁNDEZ-IBÁÑEZ ET AL. 703

subsurface uncertainty across hydrocarbon exploita- intervals to form 1–2-m cycles (Wright and Barnett,
tion business stages (exploration, appraisal, develop- 2015; Herlinger et al., 2017; Farias et al., 2019; Lima
ment, and production). and De Ros, 2019). The carbonate reservoir and
lake-center sediments are overlain by a marine in
origin salt package (Ariri and Retiro Formations,
GEOLOGIC SETTING Figure 2B).
The lacustrine Aptian carbonates of the South Atlan-

tic, commonly referred to as pre-Salt carbonates, are
part of a giant hydrocarbon province offshore eastern METHODOLOGY
South America, where the Santos and Campos are
Robust characterization of nonmatrix features and
the most prolific hydrocarbon basins (Figure 2A).
assessment of their impact on flow requires integra-
The Early Cretaceous lake formed as result of the
tion of multiscale static and dynamic data sets. We
south to north propagation of continental break
briefly review our four-step, nonlinear workflow,
up of Gondwana and opening of the South
which uses borehole image logs, calibration to core,
Atlantic (e.g., Thiede and Vasconcelos, 2010). The
losses while drilling, and WTs that have successfully
Santos and Campos Basins have been interpreted as
been applied in other carbonate reservoirs (e.g.,
resulting from two phases of extension during
Ahdyar et al., 2019).
the prerift stage of South Atlantic opening in the
late Hauterivian–early Barremian (Contreras et al.,
2010). The succeeding synrift phase in the later Step 1. Nonmatrix Identification from
Barremian–Aptian lasted 5–6 m.y. and was character- Borehole Image Logs
ized by tilting of fault-blocks and growth strata in
upper continental crust, followed by the develop- Borehole image logs are a valuable data type for the
ment of a flexural and continuous sag basin (Contre- efficient assessment of nonmatrix presence and type.
ras et al., 2010; Chaboureau et al., 2013). The rifting High-resolution images of the borehole wall can
and sag are diachronous across the basin and are rep- often resolve sedimentary structures, grain types,
resented by several unconformities (e.g., Karner et al., depositional facies, cyclicity, structural features, and
2021). The configuration of Cretaceous horsts and certain types and scales of pore types (Gaillot et al.,
grabens generated many discontinuous highs sur- 2007). For nonmatrix characterization, pore types
rounded by continuous lows (Pietzsch et al., 2018). resolvable in borehole images are either classified as
The deposits in the lows consist of lacustrine fractures, vugs, or caves. A preliminary assessment of
shales (some organic rich) and on the highs corre- the degree of openness is done by interpreting the
spond to lacustrine carbonate (e.g., Dias et al., 1988; photoelectric factor (PEF), density, sonic, and caliper
Macedo, 1989; Moreira et al., 2007; Winter et al., logs. One advantage of image logs is that they provide
2007, de Paula Faria et al., 2017). A simple generali- insights into nonmatrix features that are larger than
zation of the carbonate reservoirs is that they the diameter of a core. Such features that cannot be
are mainly composed of coquinas (Itapema and sampled by core are commonly overlooked from a
Coqueiros Formations) and shrubby-spherulitic car- hydrocarbon pore volume perspective (Fernandez-
bonates (Barra Velha and Macabu Formations, Figure ~ez et al., 2019b). Comparisons of image logs
Iban
2B). The “coquinas” reservoirs correspond to cyclic with corresponding intervals of core can reveal
bioclastic deposits, while the shrubby-spherulitic car- important discrepancies in the presence and abun-
bonates consists of millimeter-to-centimeter-sized dance of nonmatrix features (Laubach et al., 1988;
shrubs and spherulites (Dias et al., 1988; Muniz and Genter et al., 1997, Fernandez-Iban ~ez et al., 2018).
Bosence, 2015; Wright and Barnett, 2015; Herlinger This is primarily because of the remote sensing
et al., 2017; Wright, 2020). The shrubs can be micro- nature of image logs (image quality and resolution).
bial to crystalline in nature (see Wright [2020] for a Therefore, calibration to core is a critical part
detailed review) and alternate with the spherulites of the workflow necessary to validate image log
and clay-rich (stevensite and siliciclastic mud) interpretations.

Figure 1. Nonmatrix and excess permeability. (A) Ternary conceptual diagram showing host rock (in black) and porosity (in white) com-
ponents of a dual-porosity carbonate pore system (modified after Grimes, 2006). Concentric circles (not to scale) conceptually represent
pore system permeability investigated by a routine core plug analysis (blue) versus a well test (green). (B) Box and whisker plots showing
ranges of measured permeability from routine core plug analysis for two pre-Salt wells (in blue) and effective permeability derived from
well test (drill-stem tests and production tests) pressure transient analysis (green) conducted on multiple wells in the pre-Salt. Lines extend-
ing from the boxes (whiskers) represent variability outside the upper and lower quartiles, excluding outliers. The median is represented by
the horizontal line in the box. Excess permeability is defined as the difference between permeability based on core analysis data and effec-
tive permeability calculated from well test pressure transient analysis. n 5 number of samples.
Step 2. Calibration to Core compared to those from whole core is a useful indica-
tor to identify whether nonmatrix features that are
Core represents the physical reservoir, whereas all resolved at the core scale are associated with excess
other forms of data in our workflow rely on some permeability (Ehrenberg, 2007). No whole-core per-
form of remote sensing. Core provides explicit infor- meability data were available for the wells investi-
mation about the nonmatrix types and their physical gated in this study; however, differences in pre-Salt
properties. Nonmatrix features observed in core are reservoir permeability for these two scales of mea-
key for interpreting geologic processes that created surement exist (Chitale et al., 2015).
and modified them and their paragenetic sequence.
This information is critical for developing process- Step 3. Dynamic Potential from LCZs
based conceptual models that are foundational for
interpreting scenarios of nonmatrix distribution away Drilling parameters respond to nonmatrix features as
from well control. Core is critical for validating they are encountered. In particular, mud losses while
nonmatrix interpretations from image logs and devel- drilling can help identify the presence of hydrauli-
oping rules for interpretation in the absence of cally effective nonmatrix features and provide
overlapping core. Our experience with carbonate res- important insights into their permeability and flow
ervoirs has consistently demonstrated the importance ~ez et al., 2019a). There-
significance (Fernandez-Iban
of this calibration exercise to avoid overpicking natu- fore, deliberate collection and careful analysis of
ral fractures and misinterpreting vuggy zones mud loss data is a fundamental step in our workflow.
~ez et al., 2018).
(Ibrayev et al., 2016, Fernandez-Iban In our experience, the use of mud losses to help char-
Differences in permeability measured in core plugs acterize nonmatrix features by other authors is

Figure 2. (A) Map showing location of Campos and Santos Basins. Main pre-Salt fields are highlighted in green. (B) Generic strati-
graphic column for the pre-Salt formations covered in this study (after Winter et al., 2007; Moreira et al., 2007; Wright, 2020).
generally overlooked or underused for reservoir 2018) and some attempts to predict potential excess
characterization. permeability zones ahead of the bit (Duarte et al.,
Lost circulation materials (LCMs) introduced to 2018). The VvR plot is an empirical nomogram
the borehole to mitigate drilling losses often irrevoca- based on LCZs experienced during conventional dril-
bly modify reservoir properties near the wellbore ling in nondepleted reservoirs; however, more recent
such that production logging tools (PLT) might never techniques of managed pressure drilling may have an
capture the predrill, dynamic properties of many impact on the total volume lost that might not yet be
nonmatrix features (Ahdyar et al., 2019). In this fully accounted for in this method.
respect, the dynamic behavior interpreted from dril-
ling losses can in some cases be more reliable for res- Step 4. WTs
ervoir characterization than other types of dynamic
measurements. In other instances, well control opera- Representative sampling of nonmatrix features in a
tions can lead to gaps in data acquisition. Often, poor wellbore is challenging because of spatial bias. During
to no core recovery and poor image log quality exists a WT, the well is flowed to surface and the corre-
over LCZs. In these situations, we use the volume sponding pressure response (known as “pressure
and rate of drilling fluids lost to the formation to help transient”) yields insights into the properties of the
identify and characterize nonmatrix features. Cross- reservoir investigated during the test. Unlike core,
plots of drilling mud volume lost versus rate (VvR) logs, and formation testers, which are limited to the
plots (Fernandez-Iban~ez et al., 2019a) show several wellbore and its immediate vicinity, the WTs investi-
distinct categories of flow behavior that can be used gate deeper into the reservoir. The character of the
to identify dominant nonmatrix features. In this pressure derivative reflects information about
paper, we test the applicability of the VvR method reservoir heterogeneity (spatial distribution and mag-
for the identification and characterization of nonma- nitude of permeability) within the investigated
trix features in the pre-Salt reservoirs of Brazil. Losses volume. Significant structural or stratigraphic hetero-
while drilling is a frequent challenge in the pre-Salt geneities such as barriers, baffles, flow conduits, and
that has led to the use of special drilling techniques in in certain instances, dual-porosity systems, can be
recent years (Yaqoob et al., 2017; Pontes et al., interpreted and an effective permeability calculated.

If PLT data are also available to ascertain the zonal e Biocombustiveis (ANP) public database (i.e., avail-
contribution across the formation, a more reliable able for purchase). Approximately 10% of the study
estimate of the magnitude of excess permeability can data are currently proprietary; however, these data
be calculated. In carbonate reservoirs, the permeabil- will shortly become publicly available under the cur-
ity calculated from a WT commonly and substan- rent ANP data guidelines. The majority of the data
tially exceeds that measured from RCA. This excess were obtained from wellbores and include four main
permeability is typically because of the presence types (Figure 3A): (1) core data; (2) wireline data;
of nonmatrix features. Consequently, WTs are (3) drilling data; and (4) WT data. To preserve confi-
extremely valuable for reservoir characterization and dentiality, neither individual wells nor specific fields
reducing dynamic uncertainty in heterogeneous car- are identified.
bonates like the pre-Salt reservoirs of Brazil. Core data refer to observations made in person
The theory and methods of pressure transient with a hand lens “on the rock” and not by means of
analysis (PTA) are well established (Lee, 1982; core photographs. For this study, we evaluated a total
Kamal, 2009). However, in executing this study, we length of 1000 m of conventional core from 12
identified opportunities to improve the traditional wells. All cores examined for this study were col-
analysis and integration of WT data in two important lected in the Santos Basin from the Barra Velha and
ways. First, we recognized that greater collaboration Itapema Formations.
between the geoscience and reservoir engineering dis- Wireline data specifically refer to wellbores with
ciplines is crucial for the accurate characterization of available image logs (predominantly acoustic image
nonmatrix features and maximizing the value of WT logs) and accompanying standard wire-line logs (e.g.,
data. For example, the dynamic response of hetero- PEF, caliper, gamma-ray, and neutron porosity). The
geneous carbonates may not always neatly conform data set used in this study consists of a total length in
to the analytical solutions described by “textbook” excess of 10,000 m, from 35 wells. Approximately
type curves. Since PTA is nonunique, it is necessary 85% of the wells are located in the Santos Basin,
to develop alternative credible geological scenarios of whereas the remaining 15% of wells are located in
reservoir heterogeneity that are consistent with avail- the Campos Basin.
able static and dynamic data. Second, but related, is Drilling data specifically refer to information on
the early and efficient use of numerical-forward mod- mud losses while drilling. Information of LCZs was
els that simulate pressure responses as a technique to obtained from three main sources: daily drilling
help decipher excess permeability architecture and reports, end of well reports, and composite well logs.
evaluation of alternative geological scenarios. The Drilling data are available for 65 wells, with 80% of
objective of this approach is to directionally repro- the wells located in the Santos Basin and 20% located
duce the character of the pressure derivative curve, in the Campos Basin (all selected wells penetrate
but not necessarily to obtain a perfect history match. pre-Salt reservoirs). Approximately half of these
Whereas solutions remain nonunique, this approach wells (33 total) reported losses (Figure 3B).
allows for testing alternative geologic concepts as a The WT data refer to the shut-in derivatives
basis for discarding implausible scenarios. Further- obtained from the raw pressure and rate data collected
more, the fast and iterative nature of this approach during flowing and shut-in conditions. These data
provides new insights in terms of understanding the include tests conducted with temporary completions
parameters most sensitive to flow in pre-Salt carbo- in exploration and appraisal wells (i.e., drill-stem
nates where a range of reservoir architectures and tests) and test conducted with a permanent comple-
heterogeneities, covering a variety of nonmatrix fea- tion in production wells (i.e., production tests). We
tures, exists. consider WT performed preacid and postacid stimula-
tion as two separate tests. Shut-in periods last, on
average, 60 hr. We compiled derivatives from 147
DATA tests conducted in 96 wells across the pre-Salt (78
wells in the Santos Basin and 18 wells in the Campos
Data used in this study is predominantly sourced Basin [Figure 3A]). Approximately 75% of the tested
from the Agencia Nacional do Petroleo, Gas Natural intervals were conducted in the Santos Basin and 25%

in the Campos Basin. This data set was then filtered is consistent with other studies of pre-Salt reservoirs
for WTs that exhibit excess permeability, where a that report at the thin section scale extensive silica
total of 79 of the tests met this criterion between the cementation in micro fractures and vugs (Lima and
Santos (61) and Campos (18) Basins. De Ros, 2019). We classify silica into three distinct
types based on crosscutting relationships observable
in core with the naked eye and a hand lens:
RESULTS
Silica I occurs as continuous layers that span the
The following section summarizes nonmatrix obser- width of the core and are typically 1 to tens of
vations made in numerous wells from the main well- centimeters in thickness. These silica bands tend
bore data sets used in this study: core, borehole image to occur as a concordantly replacing phase
logs, LCZs, and WTs. We describe the different types around the transition from shrub boundstones to
of nonmatrix observed as well as their static and mudstones and/or laminated spherulite facies
dynamic signature. that has previously been interpreted as the top/
base of a depositional cycle in the Barra Velha
Core Observations
Formation pre-Salt reservoirs (sensu Wright and
Generally, the overall core integrity and recovery in Barnett, 2015). The replacing occurrences tend
the wells used for this study is very good. However, to have irregular to diffuse layer boundaries.
core integrity abruptly degrades as nonmatrix fea- Other occurrences of silica I, however, display
tures become more abundant. As the size of these sharp contact characteristics at the layer bound-
features nears the core diameter, core recovery drops aries that appear to be consistent with primary
abruptly to the point where the void space is greater chemical sedimentation. In facies dominated by
than the core diameter. This is the case for large, reworking such as rudstones, grainstones, and
open, meter-scale cavities evident in some image rarely, breccias, we observe broken, transported,
logs. The range of nonmatrix features typically and rotated clasts that are sourced from silica
observed in core encompasses multiscale vugs, frac- layers. We therefore interpret these silica I as
tures, and breccias with varying degrees and styles of contemporaneous with lacustrine sediment
diagenetic overprinting. deposition and likely the earliest phase of silica
Nonmatrix features in pre-Salt reservoirs mostly precipitation. The mechanical properties of these
show some open pore space, although the amount of silicified intervals make them especially prone to
pore occluding cement can range from none (i.e., develop fractures. Examples of silica I are shown
completely open) to 100% filled (i.e., completely in Figures 4A and B; 5B; and 7.
occluded). Whereas the paragenetic sequence of Silica II is chalcedony that postdates silica I. It
cements filling nonmatrix pores is complex and can crosscuts stratigraphy as a pore lining cement
vary from well to well, and between fields, there are that partially fills nonmatrix pores. Silica II is also
some common first-order core observations, possible found as intraclasts. Silica II is abundant in areas
using a hand lens, that shows some consistency. of increased fracture and vug density that exhibit
excess permeability. Silica I and silica II are com-
Silica monly closely associated. Examples of silica II are
One of the most striking observations from core (all shown in Figures 4C and 6C, E.
from the Barra Velha Formation) is the strong spatial Silica III describes clear crystalline euhedral
association of silicification and density of nonmatrix quartz. It is a later pore filling cement phase that
features that commonly exhibit excess permeability. partly occludes nonmatrix features and develops
Silica is found as a replacing phase (that can favor in the form of druses along fracture and vug walls.
nonmatrix development because of its brittleness) An example of silica III is shown in Figure 6C.
but also as a pore filling cement in nonmatrix features
that are ubiquitously dominated by silica with rare Other minor (in volume) phases typically lining
occurrence of carbonate or other cement types. This nonmatrix pores include bitumen and pyrite crystals

silicified layer or occasional dolomite layer in which
they occur, ranging from approximately 0.5 to
25 cm. Their spacing can roughly equal the layer
thickness but is commonly many times smaller.
Where BBFs contact unfractured, muddier layers
above and below, they sometimes extend a short dis-
tance into the adjacent lithology along a sinuous
tapering path. In other cases, prior to being lithified,
material from adjacent unfractured units has
deformed plastically into the open tips of BBFs. They
can exhibit large apertures with exceptionally high
width to height ratios, and a filling cement phase is
mostly absent (Figure 4B). When multiple silicified
layers occur in close proximity (typically within a
meter), BBFs commonly show a later growth phase
(that can include a shear component) where they
propagate through the less competent interbedded
layers and connect to the set of BBFs above or below.
Non-BBFs—Non-BBFs (otherwise referred to as

“fractures” throughout the text) are volumetrically
the most common type of structural discontinuity
observed in core and image logs. Based on their
characteristics we distinguish at least three differ-
ent subtypes.
Figure 3. Data types and distribution by basin. (A) Bar graph
showing number of wells evaluated by data set and by basin; Locally abundant, opening-mode structures with
(B) number of wells with reported losses while drilling and num-
steeply inclined (>60 ) to vertical dips, commonly
ber of discrete lost circulation zones (LCZs) by basin.
with irregular nonmatching surfaces and widths
that always postdate silica I and silica II cements. mostly ranging from 0.1 to 20 mm in core. These
Some occurrences exist of euhedral barite crystals fill- joints usually occur as networks of branched to
ing vugs. crossing segments that show evidence of corrosion
along the fracture walls, and are commonly filled
Fractures with silica II and silica III cements (Figure 4C).
Fracture types observed in cores include both open- These features are partly open (50%) and are
ing mode (i.e., joints) and shear fractures (i.e., faults) more abundant in areas of increased silicification
(Ramsay and Huber, 1987). We observe two types of where they can achieve a local maximum of areal
opening-mode fractures based on their terminations: fracture density of 0.03 m/m2.
bed-bounded fractures (BBFs) and non-BBFs. A less common type of non-BBF is characterized
by limited vertical extent (10 cm) and high sinu-
BBFs—The BBFs are opening-mode structures typi- osity that results from early sediment compaction
cally restricted to silicified beds (silica I) with thick- (Figure 4D). These fractures are closed, not well
nesses on the order of tens of centimeters. They are connected, and with narrow widths (<1 mm).
near vertical (>75 ) with straight matched walls and In other instances, we observe meter-tall, centime-
widths up to 2 mm, but they are commonly less than ter-wide (up to 10 cm) megafractures partly filled
0.5 mm (Figure 4A). They are relatively common in with multiple cement generations, but with the
packages of thin-bedded, fine-grained, silicified grain- distinctive characteristic of transported carbonate
stones but are generally absent in other facies. The grains and is therefore one of the earliest fill stages
height is commonly equal to the thickness of the recognizable in the fractures (Figure 4E).

Shear Fractures—Shear fractures consist of small- mostly open with minor silica cements (silica II
scale faults that impose very different degrees of or III) occluding it, and it rarely occurs in grain domi-
deformation in the surrounding rocks. These are nated fabrics and mudstones.
sliding-mode structures with steeply inclined (>50 )
to near vertical, surfaces that generally exhibit nor- Centimeter-Scale Vugs—This class includes centimeter-
mal offsets of approximately 2 to 10 cm where mea- scale, non-fabric-selective pores with an area greater
surable (Figure 5A). The small, relatively straight, than 1 cm2 (Figure 6B–D). Here, we define the area
fault segments extend from 5 to 20 cm in a dip direc- of a vug as the product of the long and short axis.
tion, ranging in width from 0.1 to 1.0 mm, and gen- The average area of a vug in core is 10 cm2 with
erally appear as black traces (likely bitumen lined) measured maximum of 55 cm2. Vugs typically occur
with occasional clasts derived from the host rock as isolated features (at the core scale) that in certain
assimilated in the fault gauge (Figure 5B). Layer cases can touch or connect with other vugs via natu-
smearing is sometimes observed as well as local fold- ral fractures. They are common in silicified and
ing (drag) of adjacent strata (Figure 5C). Shear frac- fractured intervals. Vugs also develop along natural
tures commonly occur in grain-supported carbonates, fractures. Less common are vugs that develop along
and in places, in mud-supported carbonates and bedding planes (Figure 6F). Vugs are mostly open
boundstones. Shear fractures rarely exhibit any asso- (>70%) and common pore lining cements include
ciated open pore space, which is a limiting factor for laminated silica II, silica III, and in places bitumen
fluid flow contribution because they are not volumet- and pyrite, which represent the latest pore lining
rically significant. phases observed in vugs (Figure 6B–E). The upper
boundary for vug size observable in core is set by the
Vugs core diameter. This results in an inherent core sam-
Vugs include non-fabric-selective multiscale pores pling bias toward the smaller size vugs. Borehole
that are generated or enhanced by dissolution. To image logs provide an opportunity to address this
minimize interpretation bias toward any processes, scale bias and characterize large size (greater than
we choose to classify vugs based on their size and core diameter) vugs, including caves.
visual connectivity.
Breccias
Millimeter-Scale (Touching) Vugs—This touching vugs Another variety of nonmatrix with potential for
pore system with associated excess permeability is excess permeability are breccias, which are not indi-
defined by a non-fabric-selective solution-enlarged vidual features like those described above, but rather
interconnected framework pore system that can a combination that typically includes core-scale vugs
extend the full width of the core and up to 100 cm and fractures. A wide variety of breccias exist,
vertically (Figure 6A). This touching vug pore system depending on the process behind them, which deter-
is most abundant in shrub dominated facies that are mines the nature and angularity of the clasts and
interpreted by Wright and Barnett (2015) to form in whether they are matrix or clast supported. Clasts of
the upper part of an idealized depositional cycle in shrubby boundstones, spherulitic grainstones, and
the Barra Velha Formation. Shrub boundstones can silica compose the breccias and conglomerates
exhibit high RCA permeability from hundreds of (Figure 7) that we have observed in core. Early diage-
millidarcys to occasionally greater than 1 d, depend- netic clays and dolomites as well as carbonate mud
ing on the connectivity of the pore network con- act as matrix that fills the space between clasts.
trolled by the range of depositional textures and Depending on grain size and shape, clast-supported
diagenetic overprint observed in these facies conglomerates and breccias may have high amounts
(Rezende and Pope, 2015; Herlinger et al., 2017). of interclast porosity that appear as centimeter-scale
However, the presence of touching vugs and their vugs in core. Fractures occur along clast boundaries
heterogeneity relative to the dimensions of a core or can be confined to individual clasts (Figure 7). Sil-
plug make intervals of shrub facies a more subtle can- ica I (as clasts) and silica II (as cements) are com-
didate for excess permeability compared to other monly observed in breccias with open visual porosity
nonmatrix features. This touching vug pore system is suggesting they are syndepositional in nature. In this

Figure 4. Examples of opening mode fractures in core from the Barra Velha Formation. (A) Silica-filled bed-bounded fractures (BBF) in
5-cm-thick silicified (silica I) dipping layer; (B) wide-open joints in a silicified layer; the black matter coating the fractures is bitumen; (C)
centimeter-wide open fracture with nonmatching walls resulting from dissolution enhancement. Note a light blue coating along the frac-
ture walls corresponding to silica II; the beige-colored infill is drilling mud; (D) sinuous and filled fractures interpreted as an early feature
now compacted; silica I layer noted at the bottom of the picture; (E) meter-scale partly open fracture filled with carbonate debris (CD) and
host rock angular fragments, as wells as later large white calcite crystals (CC).
paper the term syndepositional is used in a broad rules for interpreting nonmatrix away from core con-
sense to refer to processes that occur during deposi- trol. We identified six distinct classes of nonmatrix
tion or shortly after (i.e., before burial). features with characteristic image log signatures
(Figures 8, 9). These classes include the dominant
Image Log Observations features observed in core and one additional class cor-
responding to large-scale cavities that are too large
Electric and acoustic borehole image logs are avail- for sampling with conventional core.
able for most wells drilled in the pre-Salt. In our
experience, acoustic image logs are best for imaging BBFs
nonmatrix features in carbonates. This also proved to Depending on bed thickness, BBFs can sometimes
be the case for the pre-Salt reservoirs where most be identified in image logs. They are characterized
nonmatrix observations were made on amplitude by not fully developed sinusoids because they are
and traveltime borehole images. However, when truncated at the bed boundaries and the thickness of
both data sets are available, our nonmatrix character- such layers rarely exceed 25 cm. As a result, the
ization integrates observations from acoustic and expression of these fractures is that of linear traces
electric tools. Nonmatrix features observed in core that abut against the bedding planes (Figure 8A, B).
are consistent with observations made in image logs. Their most common acoustic signature is character-
Through the integration of core and image logs, we ized by low amplitude, high traveltime features sug-
calibrated our interpretation of the latter, and defined gesting they are open. Multiple traces can be

Figure 5. Examples of shear fractures (i.e., small-scale faults) in cores from the Barra Velha Formation. (A) A 60 dipping normal fault
with ~10 cm offset (yellow marker for reference) and no visual porosity along fault zone; (B) 55 dip normal fault with offset larger than
core sample, with observable grain size reduction within fault gauge zone and brittle deformation of silica I layers on footwall block;
(C) 60 dipping nonplanar fault trace with associated drag structures and growth packages suggesting syndepositional timing of the fault.
Figure 6. Examples of millimeter-scale and centimeter-scale vugs in core from the Barra Velha Formation. (A) Millimeter-scale touching
vugs in shrubby boundstones; (B) centimeter-scale open vug with silica II and bitumen lining; (C) broken-open centimeter-scale vug coated
with silica II, bitumen (Bit), and pyrite (Py) crystals; (D) centimeter-scale vugs coated with botryoidal silica II, silica III, and bitumen;
(E) detail of silica II (blue) lined by silica III (citrine); (F) centimeter-scale vugs developed along bedding planes suggesting preferential fluid
flow and dissolution enhancement.

Millimeter-Scale (Touching) Vugs
These acoustic facies are characterized by low ampli-
tude and higher than background travel times that
result in a mottled texture suggesting they are filled
with drilling fluids and therefore, at least partially
open. The touching vugs commonly appear as clus-
ters parallel to bedding (Figure 9A, B). Mud invasion
can be significantly higher than background in these
facies commonly resulting in an increase in PEF
because of the presence of barite as a weighting agent
in the drilling mud. In highly permeable, touching
vug intervals, a thick mud cake can develop resulting
in poor data quality regions or areas with observable
tool marks.
Centimeter-Scale Vugs
Vugs have an acoustic signature characterized by dis-
crete low amplitude and high traveltime events that
commonly appear as discrete anomalies often inter-
preted, and calibrated to core, as isolated vugs.
Locally, in areas of increased density of features, the
vugs can touch (Figure 9C, D) or connect to other
centimeter-scale vugs via fractures. As a result, these
high vug density regions are characterized by a local
elongation in the borehole diameter reflected in cali-
per excursions and sometimes increased PEF. Our
definition of vugs includes features as tall as 0.5 m
Figure 7. Example of matrix-supported conglomerate with sil- but with a lateral extent smaller than the wellbore
ica I subrounded clasts from the Barra Velha Formation; note that diameter.
fractures are confined to the brittle clasts.
Meter-Scale Vugs (Caves)
observed within one layer as fractures are closely These are meter-scale features characterized by low
spaced (2–6 cm). Through integration of core and amplitude/high traveltime events, with heights typi-
image logs we learned that these BBFs correspond to cally in excess of 0.5 m, associated caliper excursions
greater than 2.54 cm (Figure 10A, B) and increased
silicified (silica I) intervals, which have distinctive
PEF. This class is used to describe cavities that are
character in multimineralogy logs.
larger than the wellbore diameter and have associated
lost circulation events suggesting meter-scale dimen-
Non-BBFs sions. Because open caves cannot be cored, they are
Non-BBFs are characterized by fully developed sinus- only observable in image logs.
oids that commonly correspond to low amplitude/
high traveltime events, suggesting they are open Breccias
(Figure 8C, D). Locally, there are instances of high Breccias are characterized by chaotic image log facies
amplitude and low traveltime sinusoids that suggest with highly variable amplitude and traveltime values.
fractures are cemented. These types of fractures occur No primary stratification (in the form of sinusoids)
in all pre-Salt reservoir lithology types but are far is observed. Individual clasts can sometimes be
more abundant in regions that exhibit a high degree of observed, as well as the interclast open pore space
silicification. In this context, it is common to observe (centimeter-scale vugs) in clast-supported breccias.
these fractures coexisting with centimeter-scale vugs. Shrubby boundstones and silicified clasts are best

Figure 8. Examples of nonmatrix features interpreted from acoustic borehole image logs (amplitude). Dark colors correspond to low
amplitude values interpreted to be open pore space. To the left of each image log the filled texture represents borehole elongations; to
the right, a gray bar highlights the interval of interest. (A, B) Bed-bounded joints in ~10- to 20-cm-thick silicified layers; note that silicified
layers occur as high amplitude events; (C) 60 dip natural open fracture represented by a continuous sinusoid; (D) 80 dip open
natural fracture with irregular nonmatching walls and associated type IIIa losses. Type losses as defined by Fernandez-Iba~nez et al.
(2019b). BS 5 bit size; CALI 5 caliper.
imaged because their texture and high amplitude val- LCZs

ues, respectively, are highlighted by most acoustic
image logs (Figure 10C). Fractures within breccia The LCZs in the reservoir section were reported in
do not develop full sinusoids and often are difficult 30 out of the 65 wells examined (Figures 3B, 11).
to trace because of the limited extent and small Several of the wells had multiple LCZs and it was
apertures (Figure 10D). Clast-supported breccias possible to characterize 60 LCZs using the total vol-
with open centimeter-scale vugs and fractures ume of drilling fluids lost and the rate of loss. An
(Figure 10D) commonly show associated high PEF additional 13 LCZs were reported for which only the
values and caliper excursions (i.e., high traveltime volume lost or the loss rate were available. Reported
intervals) because of local wellbore enlargement. loss rates range from 1 to 476 m3/h (6 to 3000 bbl/h)
Integration of core and image logs have proven that and are characterized by a mean value of 32.5 m3/h
breccias are not always identifiable on image logs, (204 bbl/h) and a median of 12.6 m3/h (79 bbl/h).
and they do not have a unique petrophysical signa- Total volumes lost to individual LCZs range from 4
ture because of their variable composition. Also, our to 2250 m3 (25 to 14,150 bbl) and are characterized
core calibration has shown that steeply dipping sinus- by a mean value of 232 m3 (1460 bbl) and a median
oids within breccias commonly correspond to clast of 60 m3 (377 bbl).
boundaries and are commonly misinterpreted as The VvR plot, introduced by Fernandez-Iban ~ez
fractures. et al. (2019a), is used to visualize distinct types of

amplitude values interpreted to be open pore space. To the left of each image log, the filled texture represents borehole enlargements as
measured by single arm mechanical caliper; to the right, a gray bar highlights the interval of interest. (A) Millimeter-scale touching vugs;
(B) centimeter-scale vugs along bedding planes; (C) centimeter-scale vugs and faint fractures in silicified interval; (D) centimeter-scale
vugs. BS 5 bit size; CALI 5 caliper.
flow behavior from LCZs that are often indicative of vugs or fractures, both styles of nonmatrix features
different nonmatrix features (Figure 11). Results that have been observed in core and borehole images
show that 87% of the LCZs fall below a 50 m3/h (cf. Figures 4, 6). Type III losses represent a total of
(315 bbl/h) threshold, which based on our experi- 13% of the LCZ population and are exclusively
ence is the upper limit for mud losses in natural frac- indicative of natural fractures. When natural frac-
tures or touching vugs (Figure 11). The remaining tures are enhanced by dissolution the end result is
13% of LCZ events are typified as type IV that are fractures with a very wide aperture (up to 1 m
interpreted as meter-scale vugs (cave) related losses; observed by Ibrayev et al. [2016]) that can result in
for example, type IV losses were reported while dril- large volumes of drilling fluid losses in excess of
ling through the cave shown in Figure 10A. 300–400 m3 (1885–2515 bbl). Losses of this magni-
Type I losses, with relatively small volumes tude are classified as type IIIa and represent 10% of
(<60 m3, 377 bbl) of fluid lost, represent 46% of the the data population in this pre-Salt reservoir study.
LCZ events. Losses that plot in the type I domain Type IIIa losses occurred while drilling the 0.5 m
can be related to natural fractures, touching vugs, wide enlarged fracture shown in Figure 8D.
small cavities (e.g., Figure 10B) or drilling induced Based solely on the drilling losses data presented
fractures and require integration with borehole but corroborated with other types of data discussed
images or core to confidently assign the nonmatrix in this study, the pre-Salt reservoirs of Brazil have a
features responsible. Type II losses, which represent relatively high abundance of nonmatrix features that
18% of the LCZ events, are indicative of touching will impact their dynamic behavior (Figure 11). This

amplitude values interpreted to be open pore space. To the left of each image log, the filled texture represents borehole enlargements as
measured by single arm mechanical caliper; to the right, a gray bar highlights the interval of interest. (A) Meter-scale cave with associated
type IV losses; (B) meter-scale open cave with associated type I losses; (C) chaotic facies with high amplitude events interpreted as silicified
clasts within a cemented breccia; (D) chaotic mix of open fractures and vugs interpreted as a breccia within a silicified interval. Type losses
as defined by Fernandez-Iba~nez et al. (2019b). BS 5 bit size; CALI 5 caliper.
is especially true when considering spatial bias, and multiple test intervals in addition to tests performed
the odds of intersecting nonmatrix features with a before and after wellbore acid stimulation targeting
vertical wellbore (Fernandez-Iban ~ez et al., 2019b). a given interval. Of these WTs, 76 showed evidence
Furthermore, it is evident that fractures are the lead- of excess permeability. Identification of excess
ing type of nonmatrix features responsible for losses permeability was based on the magnitude of the inter-
while drilling in the pre-Salt although meter-scale preted WT permeability relative to the typical range
vugs (caves) are also present (Figure 11). This result observed in cored intervals and the behavior of the
reinforces that drilling loss data, which in our experi- derivative curve during PTA. The majority (80%) of
ence tends to be underused, is extremely valuable for all the derivatives were classified as one of four type
carbonate reservoir characterization especially since curves, i.e., derivatives with similar shapes and poten-
it often arrives ahead of other data types used to iden- tially comparable underlying geologic concepts
tify and characterize nonmatrix features. (Figure 12). The four type curves identified in this
review and their respective abundance are as follows:
WTs
Linear flow (30% of WTs with excess permeabil-
A total of 147 WTs from 96 wells in pre-Salt carbon- ity signature): The derivative exhibits a half-slope
ate reservoirs from the Santos and Campos Basins at early time, that is, a line that moves upward by
were used in this study. This includes wells with one log cycle for every two cycles along the time

upward trend can commence very early on,
within minutes. This response indicates a reduc-
tion in effective flow capacity away from the
wellbore.
A “V-shape” response in the derivative is the

most widely recognized paradigmatic indicator for
the presence of dual-porosity behavior and the pres-
ence of nonmatrix features (Warren and Root, 1963;
Freites et al., 2019). However, of the pre-Salt WT
data set showing evidence of excess permeability
(corroborated with wellbore observations of nonma-
trix and LCZs), only five wells exhibited a V-shape
Figure 11. Volume versus rate plot with characteristic regions response (3.4% of all WTs). The remaining subset of
and type losses. Red dots correspond to lost circulation zone WTs in our study that exhibits excess permeability
events reported in wells from this study. Gray dots correspond to (48% of all WTs examined) displays a variety of
a global database from carbonate reservoirs (Fernandez-Iba~nez
type curve responses. This finding is consistent with
et al., 2019b). Characteristic regions and type losses according to
Fernandez-Iba~nez et al. (2019b). Type I: mixed; type II: touching the emerging view that nonmatrix features and asso-
vugs (TV) or natural fractures; type III: natural fractures; type IIIa: ciated excess permeability are not solely restricted to
dissolution enhanced natural fractures; type IV: caves/karst. exhibiting the V-shape type curve (Kuchuk and Bir-
Max. 5 maximum. yukov, 2015; Freites et al., 2019; Wan et al., 2020).
This serves as a reminder that the integration of dif-
(horizontal) axis (Figure 12A). This type of deriv- ferent multiscale static and dynamic data sets is a best
ative is characteristic of stimulated wells; the practice for the reliable characterization of nonma-
excess permeability is interpreted to result from trix features and excess permeability. To illustrate
the development acid-induced wormholes this point, we describe the application of the work-
extending only a moderate distance away from flow presented to a pre-Salt well.
the wellbore (Bourdet, 2002). It does not appear
to be derived from the presence of natural, pre-
dominantly open, nonmatrix features. EXAMPLE OF DATA INTEGRATION AND
Sharp roll-over (20%): The derivative curve INTERPRETATION
drops two or more log cycles indicating an
extreme increase in effective flow capacity away, Well A is from a pre-Salt field in the Santos Basin,
but near, the wellbore (Figure 12B). Our obser- Brazil, with excess permeability derived from non-
vations suggest there is a common association matrix features. This well was selected because it has
between this type of derivative and the occur- a comprehensive data set including drilling data,
rence of heavy mud losses while drilling (types acoustic borehole images, conventional wire-line
III and IIIa) and increased degree of silicification. logs, elemental logs, conventional core, and two WTs
Sustained downward trend (17%): This type of (Figure 13). Partial overlap exists between the test
curve is characterized by a derivative that intervals and core. This is informative for a rock-
decreases at a steady rate over the entire duration based interpretation of the dynamic response
of the shut-in period (Figure 12C). In some cases observed in each of the tested intervals.
for a total of 100 hr. The typical rate of decrease
is up to two log cycles over the duration of the Drilling Experience: LCZs
buildup period, indicating a gradual increase in
effective flow capacity away from the wellbore. First, we reviewed the drilling experience at this well.
Sustained upward trend (15%): The derivative Losses were reported while coring the reservoir sec-
curve increases in a steady upward fashion, typi- tion of well A with an onset of severe losses starting
cally up to 1.5 log cycles (Figure 12D). This at 4950 m (Figure 13A). Shortly after this, coring

Figure 12. Log-log plots of time versus pressure showing the four most common pressure transient type curves with associated excess
permeability observed in the pre-Salt. The solid line corresponds to the pressure difference curve and the dashed line to the derivative
curve of an idealized pressure transient analysis plot. Type curves: (A) linear flow; (B) sharp roll-over; (C) sustained downward trend;
(D) sustained upward trend.
operations were terminated because of heavy losses. above the onset of losses. The top 85 m of the core is
The well was drilled to total depth with continuous dominated by grainstones and shrub boundstone
heavy losses that could not be stopped. A total of facies organized in typical depositional cycles. The
1350 m3 (8500 bbl) of drilling fluids were lost to the thicker (>20 cm [>0.65 in.]) packages of shrub
formation in the lower half of the reservoir section. boundstone facies, interpreted to occur at cycle tops,
Well A did not report data on loss rate and therefore, exhibit a well-connected and pervasive network of
on this occasion, it was not possible to predict non- open millimeter-scale vugs (Figure 13B). In contrast,
matrix types using the VvR plot alone. However, the silicified (silica I) layers (<0.5 m thick) that exhibit
presence of heavy losses is already a strong indicator closely spaced open-to-partly-open BBFs are present
of nonmatrix features and the potential for excess toward the base of depositional cycles (Figure 13B).
permeability at well A. The bottom 10 m of the core is characterized
by a superimposed nonmatrix pore system that
Core and Image Log Observations consists of fractures, BBFs, and centimeter-scale vugs
that obliterates the primary stratigraphic fabric
Second, we reviewed the core and image log interpre- (Figure 13B, C). This is coincident with the onset of
tation. More than 90% of the core was collected losses reported at well A and a qualitative visual

increase in the silica content observed in core. Ele- robust set of rules, a training data set, to improve
mental logs quantify silica as 20% to 25% of image log interpretation where core calibration is not
the total rock volume in the lower part of the core available. With this information in hand, acoustic
(Figure 13A). Core observations reveal that whereas borehole images at well A were interpreted for non-
silica I is the dominant silica phase, silica II is also pre- matrix features. Results from image log interpretation
sent and appears as a common cement lining the at well A suggests the upper part of the reservoir
open nonmatrix pore space. Nonmatrix features interval (<4950 m) is dominated by matrix with
observed in this core interval show evidence of disso- interbedded layers (up to 1 m thick) of millimeter-
lution predating at least silica II and III cements. scale touching vugs and some BBFs (Figures 13A, B).
Wide dissolution enhanced fractures with sinuous Nonmatrix features in high density are visible in
nonmatching walls (cf. Figure 4B, C) and non-fabric- image logs over the lower half of well A, with a com-
selective centimeter-scale vugs (Figure 13C) domi- bination of fractures and centimeter-scale vugs
nate the nonmatrix system here. Bitumen and pyrite and some millimeter-scale touching vugs intervals
are also common pore lining phases in the nonmatrix (Figures 13A, C). The distinctive character of the
features observed in this core interval (Figure 13C). upper and lower sections of well A is also reflected by
The next step is to interpret acoustic borehole the caliper response. The upper section shows an
images. Core provides an opportunity to ground-truth in-gauge borehole, whereas the lower section reveals
image log interpretations as well as to evaluate any non- a rough and enlarged borehole caused by the
matrix features below image log resolution. Nonmatrix presence of abundant and open nonmatrix features
features were identified and assigned to one of six (Figure 13A).
classes described in the sections above. The BBFs, frac-
tures, millimeter-scale touching vugs, and centimeter- WTs Interpretation
scale and meter-scale vugs were all interpreted on the
acoustic image logs and validated with core (where Two WTs were performed in well A (Figure 13A).
possible). Integration of both data sets also revealed the Test 1 was performed in the upper section over a zone
potential for misinterpretation. For example, some 66 m thick where the main features observed in
steeply dipping features interpreted as fractures, were both core and image logs are millimeter-scale touch-
in fact beds in high-relief mounds or boundaries ing vugs and BBFs. A second test (test 2) was con-
between large clasts. Also, there were a few instances ducted over a shorter interval, 30 m, overlapping
where interpreted vugs corresponded with incipient the lowermost part of the core, and the onset of losses.
borehole breakouts. The vast majority of nonmatrix Test-derived permeability in both cases would be in
features not imaged by acoustic logs but visible in the the order of several darcys, thus significantly greater
core corresponded to vugs and small aperture fractures. than the RCA permeability (cf. well A in Figure 1B).
The BBFs that were filled with cements or occurred in Although both tests in well A suggest the presence of
thin silica layers were seldom captured by the acoustic excess permeability, they yield values that are an order
image because their properties are similar to the of magnitude apart and have distinct derivative type
matrix. Overall, acoustic image logs captured approxi- curves.
mately 30% of the nonmatrix features observed in The derivative of test 1 shows a short-lived stabi-
core. Whereas this 30% represents the largest and lization of approximately 0.1 hr (with the overprint
probably most important features for flow under a pri- of nonreservoir effects from 0.05 to 0.1 hr) yielding
mary depletion scenario, some if not all of the 70% of an apparent permeability averaged over the test inter-
the features below image log resolution may be an val of 1600 md. From there, the derivative exhibits a
important factor during secondary oil recovery. “sustained upward trend” (Figure 14A), implying a
Lessons learned from the integration of core and reduction in flow capacity away from the wellbore,
image logs were used to define rules and awareness which a standard interpretation would assign to the
around possible misinterpretations away from core presence of baffling heterogeneities in the near-
control. This is not a one-well exercise, rather some- wellbore region (Kamal, 2009). An alternative, more
thing that is consistently done in all wells with over- geologically plausible interpretation, especially given
lapping core and image logs. The intent is to build a the unequivocal presence of nonmatrix features at

Figure 13. Integrated nonmatrix characterization at well A. (A) Wire-line logs from well A (from left to right): depth, caliper (C1), bit size
(BS), acoustic image log (amplitude), cored interval, nonmatrix interpretation, well test intervals, losses (total volume lost), mineralogy log.
(1) Interval with observed vuggy shrubs; (2) interval with observed bed-bounded fractures (BBF); (3) downward increase in silica; (4) onset
of losses; (5) interval with observed centimeter-scale vugs and dissolution enhanced fractures; (6) interval interpreted as part of fault dam-
age zone. (B) Examples of nonmatrix features observed in the upper half of the wellbore: millimeter-scale vugs (or vuggy shrubs) in core
and image logs (left) and bed-bounded fractures (right); (C) examples of nonmatrix features observed in the lower part of the borehole
coinciding with the onset of losses and increase in silica content: vugs and dissolution enlarged fractures as seen in image logs (left) and
core (right). DST 5 drill-stem test; TV 5 touching vugs; Vol 5 volume.
the wellbore, is a transition from nonmatrix to average permeability over the WTs is likely underesti-
matrix-dominated flow. Two candidates for excess mating the actual effective permeability of the
permeability in test 1, both present in the overlapping contributing intervals that are likely to be in part con-
core, are (1) millimeter-scale touching vugs and (2) centrated in small nonmatrix dominated intervals.
BBFs in silicified layers. If these stratigraphically con- Well A is located on a local structural high that
cordant features are of limited lateral extent this may have provided the antecedent topography for
would indeed result in an effective loss of excess per- the shrub dominated reservoir facies to develop and
meability away from the well, hence a decrease of the experience diagenetic modification through repeated
overall flow capacity away from the well, consistent exposure or focusing of diagenetic fluids. This reser-
with the upward trending derivative that picks up at voir facies becomes less abundant away from the
0.3 hr (Figure 14A) and remains as such for 20 hr crest toward the flank of the structure, so their con-
(i.e., for approximately two full log cycles). The tribution to flow capacity decreases (Gomes et al.,

2020). Based on our observations, we argue that pre-
Salt wells in crestal positions, when away from faults,
and with flanks within the volume investigated by
the test are likely to develop “sustained upward
trend” derivative type curves.
Test 2 (Figure 14B) shows a “sharp roll-over”
derivative type curve suggesting substantial increase
in flow capacity immediately adjacent to the well-
bore, which goes on for 7 hr before the derivative
deviates upward with an overprint of tidal effects.
Although the derivative did not stabilize at any point,
its relative position compared to that of test 1 indi-
cates a permeability approximately an order of mag-
nitude greater (17,600 vs. 1600 md). Test 2 was
conducted over the interval where the well experi-
enced the onset of heavy losses during drilling; both
core and logs show the presence of a highly silicified
interval with increased density of open centimeter-
scale vugs and fractures (Figure 13). The lower part
of well A was drilled near the tip of a seismically
resolvable fault (Figure 15); the nonmatrix features
in this section and associated pressure transient dis-
playing excess permeability are interpreted to be part
of a fault damage zone (Figure 13). We have made Figure 14. Log-log plots for test 1 and test 2 in well A. (A) Test
similar observations regarding the degree of silicifica- 1 shows a short-lived derivative stabilization 0.1–0.3 hr and a sus-
tion, increased density of nonmatrix features, losses tained upward trend afterwards suggesting a reduction in overall
while drilling, and “sharp roll-over” derivatives in flow capacity away from the wellbore. Permeability-thickness at
other pre-Salt wells in close proximity to seismically the well is estimated to be 347,000 mdft (k 5 1600 md, h 5 66
resolvable faults. In these cases, the response of a typ- m [217 ft], in the absence of a production logging tool [PLT]). (B)
ical pre-Salt WT with a 72-hr maximum flow period Test 2 shows a sharp roll-over of the derivative in the early time
region with no stabilization of the pressure derivative.
and less than 60-hr buildup period may not be suffi-
Permeability-thickness is estimated to be 1,720,000 mdft (k 5
cient to investigate beyond the fault damage zone 17,600 md, h 5 30 m [98 ft] in the absence of a PLT) at 1–5 hr.
and characterize uniquely matrix properties. An At late time, the derivative is exhibiting the overprint of tidal
extended WT (EWT) may be required to further effects resulting in the sharp, oscillating response.
de-risk reservoir properties and ascertain the ability
of the matrix to recharge nonmatrix features during challenging to identify and discriminate between
production (Rosa et al., 2018). credible geological scenarios that are consistent with
a nonunique test signature. An integral part of our
PTA Model-Assisted Validation of workflow is the use of forward numerical models
Geologic Concepts that simulate pressure transient responses. This
approach provides a model-assisted quantitative vali-
As illustrated by well A, the pre-Salt reservoirs of dation of alternative plausible geological scenarios.
Brazil can be extremely heterogeneous and exhibit For example, is the “sustained upward trend”
several different styles of nonmatrix features that type derivative in test 1 (Figure 14A) consistent with
have a wide range of permeability contrast with the the geologic concept of stratigraphically concordant
matrix. Juxtaposition of the reservoir with imperme- beds of vuggy shrub boundstone decreasing in abun-
able overlying salt (anhydrite and halite) by faults can dance away from the well? To address this question,
further complicate the reservoir architecture and a three-dimensional numerical model with a thick-
associated pressure response. Therefore, it can be ness of 66 m was used to simulate the test 1 response

In summary, conducting a well-by-well analysis
and integration of observations across multiple scales
and data sets provides the most robust static and
dynamic geologic conceptual model or set of models.
This approach helps constrain key uncertainties but,
most importantly, identifies implausible geologic sce-
narios that can be discarded, thus minimizing
the number of iterations (and cycle time) with the res-
ervoir simulation model. Test 1 required a stratigraph-
ically concordant model of excess permeability
Figure 15. Seismic line across structure showing location well (touching vugs or BBFs), where lateral extent and con-
A along with a gamma-ray log (green) displayed over the reser- nectivity were key uncertainties. However, the lower
voir section and interpreted faults (white dashed lines). Note that
section of the well was dominated by stratigraphically
well A is located in a structural high and the tip of an interpreted
fault intersects the lower part of the wellbore coinciding with the discordant nonmatrix features, likely related to a fault
onset of losses, increase in silica, and observations of nonmatrix zone, with implications for vertical connectivity. Test
(fractures and vugs) that we interpret as part of the fault damage 2 signature here was completely dominated by the
zone. Scale ratio is 1:3. nonmatrix, thus providing no information on matrix
properties, which according to wireline and core data
were low because of a high degree of silicification.
for different specified reservoir heterogeneities
Matrix permeability and its ability to feed into the
(Figure 16). Ten vuggy shrub boundstone layers, 1
nonmatrix in the lower section of well A were key
m in thickness with a permeability of 2000 md
uncertainties that can impact in-place volumes,
(based on PLT derived permeability in analogue
sweep, and estimated ultimate recovery.
wells) and of variable lateral extent (200 to 2000 m)
away from well A as depicted in Figure 16A and B,
were superimposed on a reservoir matrix permeabil- DISCUSSION
ity of 100 md. Permeability was modeled as isotropic
and the porosity was uniformly set to 20 porosity Silica in the Pre-Salt
units (p.u.) in the matrix and vuggy layers (Figure
16B). The simulated pressure transient yields a Well observations suggest a close association between
“sustained upward trend” derivative that supports silica and excess permeability in some scenarios.
the viability of this proposed geologic concept (Fig- Although silica is seemingly ubiquitous across both
ure 16C). The BBFs in silicified beds observed in the Santos and Campos Basins, it is currently chal-
core and image logs were identified as another, but lenging to predict, a priori, because the source of
distinctly different, type of nonmatrix feature with silica and mechanisms of formation are not well
excess permeability in test 1 interval (Figure 13). understood or constrained for the ancient lacustrine
Would a test, using identical design specifications, system. Currently, several hypotheses exist to explain
be able to differentiate between these two different the presence of silica in these carbonate reservoirs,
types of nonmatrix features? This was investigated which are often challenging to consider in complete
by reducing the porosity of the 1-m-thick layers (Fig- isolation from one another.
ure 16B), which in this scenario represent silicified
beds, to 2 p.u. Permeability in these silicified layers Silica I
was specified as 2000 md to represent the open Silica I, the earliest occurrence of silica, is strata-
BBFs. The simulated pressure derivative overlies the bound and commonly found near the base of the
prior vuggy shrub layer model response (Figure cycle. Because of the nature of silica I, it is thought to
16C). Therefore, it would not be possible to differ- form syndepositionally as a result of changes in lake
entiate between these two different end-member water chemistry. Here, we consider four different
nonmatrix geologic scenarios, vuggy shrubs versus sources of silica for silica I: (1) direct precipitation of
silicified fracture layers, with a WT alone. amorphous silica from supersaturated lake waters

Figure 16. Forward numerical model of laterally limited layers of excess permeability (perm) (vuggy layers and bed-bounded frac-
tures). (A) Map view of areal extend of the three sets of layers modeled. (B) Schematic cross section at the well showing vertical arrange-
ment of three sets of layers with properties of vuggy shrubs (i.e., 20% porosity (20 p.u. [porosity units]) and 2000 md permeability). In
gray, background matrix properties corresponding to 20% porosity (20 p.u.) and 100 md permeability. Excess permeability layers are 1 m
thick. (C) Pressure transient analysis showing model results (pressure difference curve [top] and its derivative [bottom]) for vuggy layers
(pink) and bed-bounded fractures (blue). Both model results overlap in log-log space, suggesting minimal impact of porosity difference
between conceptual models.
(e.g., Behr, 2002); (2) destabilization of Mg-rich consistent with the supersaturation of silica phases
clays (e.g., Tosca and Wright, 2018); (3) destabiliza- like chalcedony (Tosca and Wright, 2018), yet chal-
tion of a magadiite precursor (Eugster, 1967; Behr, cedony is kinetically limited and unlikely to abioti-
2002); and (4) hydrothermal waters (temperatures cally precipitate at surface conditions (e.g. Harder,
>10 C above ambient formation temperature) carry- 1965). However, if biota such as diatoms or cyano-
ing silica-rich waters along permeable fluid pathways bacteria were present in the pre-Salt during the mid-
(e.g., Mercedes-Martın et al., 2019). dle Cretaceous (Sims et al., 2006), then silica could
Present-day analogs for the pre-Salt include have been biotically mediated (e.g. Medlin, 2002;
extensional basins such as the East Africa Rift (Tier- Tang et al., 2014). Relationships between diatoms
celin et al., 1994), Ounianga Basin (Tosca and and carbonate microbialites have been observed in
Wright, 2018), and the Great Basin, United States other hypersaline lakes such as the Laguna Negra in
(Demott et al., 2019), where large lakes, underlain Catamarca, Argentina (Gomez et al., 2018) that
by basalt, have formed as a result of rapid extension could provide some insight into the pre-Salt. Yet, to
(e.g., Lake Tanganyika, Lake Yoa [Tosca and Wright, date, there is no direct evidence of diatoms in the
2018] and Lake Lahontan [Benson 1994; Vice et al., Barra Velha Formation and, therefore, diatoms are a
2007]). Present-day water chemistry in Lake Yoa is questionable source of silica I. Cyanobacteria,

however, are prevalent in the pre-Salt and have been silica II: (1) remobilization of silica I or the conver-
identified to deposit silica on extracellular polymeric sion of amorphous silica to chalcedony; (2) hydro-
substance in marine settings and are hypothesized as thermal (temperatures >10 C above ambient forma-
a potential source of silica in the deep ocean (Tang tion temperature) fluids moving through faults and
et al., 2014). If a similar process has occurred in fractures; and (3) movement of silica-rich fluids
the pre-Salt, much of the strata-bound silica could downward (e.g. silcrete) as a result of exposure and
be explained by a symbiotic relationship with weathering (e.g. Murray 1990).
cyanobacteria. The first obvious source of silica II would be the
Silica I could also be sourced from the destabili- remobilization and reworking of silica I. This could
zation of stevensite, a Mg-rich clay that is prevalent be associated with the conversion of amorphous silica
in the pre-Salt (e.g., Tosca and Wright, 2018; Sartor- to chalcedony, a reaction that results in a reduction
ato et al., 2020). Stevensite is stable in highly saline of molar volume (Bettermann and Liebau, 1975).
lakes at pH >9. When a freshening event occurs as This process could create localized fractures, with
lake levels rise or if stevensite is exposed to meteoric localized fluid movement that would result in the
conditions, the decrease in pH will result in destabili- redistribution of silica I. Another potential source of
zation of stevensite and release SiO2(aq) into solution silica II is, again, hydrothermal fluids sourced from
(following equation 1). This process would result in a within basement. In this case, silica II would only be
cyclical pattern of silica associated with changes in expected near fault zones and the related fracture
lake level. Similarly, destabilization of stevensite network that are open to fluid flow. It is challenging
could occur during early shallow burial if fresher to constrain this hypothesis because many faults are
waters are circulated through the subsurface. How- likely subseismic and the lateral extent and distribu-
ever, when considering mass balance and molar vol- tion of silica associated with deep-seated basement
umes, it would take 2.5 times more clay than the faults is uncertain. The last hypothesis we consider
observed silica. here is the potential for silica-rich fluids to be trans-
ported downward from the destabilization of overly-
Na0:66 Mg2:67 Si4 O10 ðOHÞ2 + 2:67 CaCO3 + 3:33 H+
ing clays that create silcrete layers associated
+ 2:67 HCO3 ! 2:67 ðCa, MgÞðCO3 Þ2 + 4 SiO2
with exposure and weathering (similar to what is
+ 0:66 Na+ + 4 H2 O described in Murray, 1990). Again, this would
(1) require 2.5 as much clay as observed in silica II, but
this process cannot be ruled out.
Another potential source of silica I is from hydrother-
mal fluids that are sourced from within the basaltic
basement and are brought to the lake waters via Silica III
faults and fractures (i.e., Mercedes-Martın et al., Silica III, also referred to as megaquartz by other
2019). When these warmer waters with higher silica authors (e.g., Sartorato et al., 2020), is the last occur-
solubility (i.e., Bettermann and Liebau, 1975) and rence of silica observed in core. It is fracture-filling
associated higher concentrations of silica interact euhedral quartz and is commonly associated with
with the lake waters, silica may become strongly large pyrite framboids and bitumen. Silica III is
supersaturated and precipitate from solution in local- thought to be the result of a much later burial-
ized regions around hydrothermal springs (e.g., Sar- related diagenetic event associated with charge
torato et al., 2020).The observed cyclicity could be (either coeval or just prior) (e.g., Sartorato et al.,
explained by pulses of discharge from the spring or 2020). However, in some cases, it is challenging to
the concentration of lake waters undergoing evapora- constrain the timing of silica III because we cannot
tion (Mercedes-Martın et al., 2019). discount the potential of an early (post-silica II) for-
mation. One potential hypothesis for the early for-
Silica II mation of megaquartz could be the transition from a
The next phase of silica observed is silica II that cross higher state of supersaturation to a lower state of
cuts strata and is commonly associated with silica I. supersaturation, reducing the number of nucleation
Here we can consider three potential sources for sites, and resulting in larger quartz crystals as

hypothesized in paleosilcrete horizons in northcen- across stratigraphy) (Figure 17). We describe the
tral Texas (Murray, 1990). main processes and resulting features that drive
The distribution and type of silica is challenging excess permeability as well as some key diagnostic
to constrain, yet, in our experience can be crucial for indicators that in our experience are indicative of
predictability of excess permeability regions in the each geologic concept. Additional potential indica-
pre-Salt. In particular, predictability of silica I is key tors as well as distribution and main uncertainties
for distributing nonmatrix, because volumetrically it associated to each excess permeability concept are
is the most significant phase and we observe a strong addressed that can be used by any practitioner as a
correlation between silica I and presence of opening- first step to develop a credible geologic scenario and
mode fractures (BBFs and fractures), centimeter- implement such concepts in a reservoir model (Saka
scale vugs, and caves. Faults are one particular case et al., 2019).
where the presence of silica appears to enhance the
hydraulic properties making them susceptible to Stratigraphically Concordant Nonmatrix Concepts
losses while drilling and exhibiting a sharp roll-over Stratigraphically concordant nonmatrix concepts are
derivative type curves in WTs that yield multi-darcy subdivided into two types: (1) millimeter-scale
permeability. Silica II and silica III are pore lining touching vugs in shrub facies and (2) silicified layers
cements that can impact the magnitude of the excess with bed-bounded open fractures (Figures 17, 18).
permeability because they can control the amount of To predict the spatial distribution of these nonmatrix
open pore space and overall network connectivity. features, for example in a reservoir model for flow
simulation, requires careful consideration of alterna-
Preliminary Nonmatrix Conceptual tive depositional models and the diagenetic processes
Geologic Models that generated the pore system of interest.
Several alternative depositional and diagenetic con- Vuggy Shrubs—The vertical stacking pattern of depo-
ceptual models for the pre-Salt reservoirs of Brazil sitional cycles (e.g., Wright and Barnett, 2015), and
exist. However, these published models almost the lateral extent of shrub dominated beds, with an
exclusively focus on the reservoir matrix with little average thickness 0.51 m in the Tupi field (Wright,
description or incorporation of nonmatrix features. 2020), is a first-order constraint on the spatial distri-
We address this disparity by introducing some pre- bution of the millimeter-scale touching vug intervals.
liminary conceptual models of nonmatrix features Shrub dominated beds do not always have associated
and associated processes in the pre-Salt reservoirs excess permeability. Locally, the intershrub pore
(Figures 17, 18). We acknowledge that these models space is occluded with Mg-silicate clay (stevensite),
are immature and require future refinement as more and commonly there is evidence of pore filling
pre-Salt subsurface data (static and dynamic) and Mg-silicate dissolution in addition to partial non-fab-
applicable analogue data are integrated. Despite this ric-selective dissolution of the calcite shrubs (Wright
limitation, there is an imperative to evaluate the and Barnett, 2015; Lima and De Ros, 2019). Pore
impact of nonmatrix features on reservoir perfor- space between shrubs can also be filled with silica
mance to support pre-Salt field development invest- cement or dolomite (Wright and Barnett, 2015).
ment decisions. We contend that for this purpose Typically, excess permeability in shrub facies is asso-
alone, the conceptual models presented serve as a ciated with mm-scale touching vugs that from a
useful guide for developing credible scenarios of non- solution-enlarged framework pore system.
matrix features with excess permeability that can be In situ shrub facies are more abundant on paleo-
superimposed with compatible depositional and dia- highs (Gomes et al., 2020) (concept 1 in Figures 17
genetic concepts for the matrix. and 18). Shrub facies on paleohighs above wave base
We distinguish two broad groups of geologic also have a lower likelihood of intershrub pore space
concepts for nonmatrix features in pre-Salt reservoirs being occluded with Mg-silicate clay (Tosca and
as (1) stratigraphically concordant (i.e., for concepts Wright, 2018). Therefore, paleohighs represent a
that follow stratigraphy) and (2) stratigraphically dis- “sweet spot” for shrub facies with potentially a well-
cordant (i.e., for concepts where nonmatrix cuts developed touching vug pore system. The downdip

Figure 17. Summary chart of concepts for nonmatrix and excess permeability in the pre-Salt. From left to right: Sketch of geologic con-
cept, main nonmatrix types associated to each concept, process(es) governing development of such features, two diagnostic indicators,
secondary indicators, conceptual distribution of high probability regions affected by this type of excess permeability, and main uncertain-
ties from a model implementation perspective. BHI 5 borehole image logs; PEF 5 photoelectric factor; PTA 5 pressure transient analysis;
VvR plot 5 volume versus rate plot (Fernandez-Iba~nez et al., 2019b).
extent of shrub facies and the presence or absence focused in more permeable beds (Jones and Xiao,
of pore filling Mg-silicate is more uncertain with sev- 2013; de Luca et al., 2017; Lima and De Ros, 2019).
eral alternative depositional and diagenetic models For this concept, we would predict the distribution
related to paleostructure, lake bathymetry, and lake of dissolution to be more localized in shrub beds that
water chemistry (Wright and Barnett, 2015; Tosca intersect faults. By convolving the depositional and
and Wright, 2018; Wright, 2020). Significant diagenetic controls described and their relative uncer-
reworking of shrub boundstones is commonly tainty, it is possible to define credible alternative sce-
observed and this can further complicate the spatial narios for the spatial distribution of solution-enlarged
prediction of shrub dominated facies and any associ- framework pore system that results in nonmatrix
ated touching vug pore system (Gomes et al., 2020; vuggy intervals with associated excess permeability
Wright, 2020; Ferreira et al., 2021). in shrub facies especially in regions of limited well
The diagenetic process of dissolution that acts as control.
a second-order constraint on the distribution of high
permeability vuggy intervals in shrub facies has sev- BBFs—Opening-mode fractures, confined to silica I
eral alternative modes of origin. Exposure-related layers, that are commonly open and in cases exhibit
meteoric dissolution, including the mixing of wide apertures are characteristic of intervals with
rainwater with evolved lake waters, is one subset of associated excess permeability. The BBFs likely form
plausible diagenetic processes that could be laterally as the result of vertical stacking of layers of contrast-
extensive. Alternatively, dissolution may be related ing mechanical properties with an increasing over-
to the geothermal circulation of groundwater and/or burden load. Core observations of material from
ascending hydrothermal fluids from faults with flow adjacent unfractured units that has deformed

Figure 18. Schematic conceptual cross section (not to scale) of the pre-Salt reservoirs illustrating main type of concepts for nonmatrix
and excess permeability under present-day reservoir conditions. The inset on the figure shows an idealized stratigraphic section at the
core scale. Numbers on the figure point to locations for different excess permeability concepts (cf. summary chart on Figure 17):
(1) millimeter-scale touching vugs; (2) bed-bounded fractures; (3) fault zones; (4) fractures and faults because of margin instability;
(5) freshwater lens and meteoric exposure-related caves; (6) slope breccias. BDF 5 boundstone-dominated fabric; GDF 5 grainstone-
dominated fabric; MDF 5 mudstone-dominated fabric.
plastically, prior to being lithified, into the open tips combination of the latter process and additional dis-
of BBFs, suggest some of these fractures formed very solution enhancement
early. This process likely starts as soon as layers are Understanding the controls on the distribution of
silicified but continues through burial. When multi- silicified layers (silica I) is a first-order constraint for
ple silica layers are in close vertical separation, some the spatial prediction of stratigraphically concordant
fractures grow out of the silica layer, propagate BBFs (concept 2 in Figures 17 and 18). The preferen-
through the less competent interbedded layers, and tial replacement of fine grain rocks by silica I is
connect to the set of BBFs above or below. Such pro- demonstrably early; however, the geological pro-
cesses can have implications for the ratio of vertical cesses responsible are yet not well understood. This
to horizontal permeability. makes it challenging to reliably predict the lateral
We argue that the process of stabilization of silica extent of individual silica I layers observed in core or
I from its amorphous form (opal A) to quartz can image logs and the relative abundance and continuity
lead to additional fracturing of the silicified layers, as of these layers containing BBFs along depositional dip
well as widening of existing fractures, because and strike.
this conversion involves a volume reduction of
21.7% (29 to 22.7 cm3 mol1 [Geochemist Work- Stratigraphically Discordant Nonmatrix Concepts
bench database data0.ymp.R2]). Any preexisting The second class of nonmatrix in the pre-Salt is dis-
fractures would preferentially accommodate the cordant to the stratigraphy and it is related to frac-
strain generated from such volume reduction thus, tures, often solution enhanced, that are commonly
explaining the unusually wide BBFs sometimes associated with non-fabric-selective centimeter-scale
observed in core. The final wide apertures are likely a vugs and caves.

Faults—Highly silicified carbonates that encompass long-lived structures, and during burial they are com-
replacement of precursor carbonate and cementation monly reactivated in shear (Nolting and Fernandez-
(silica I and II) typically host this fracture-vug non- ~ez, 2021), potentially linking up to further
Iban
matrix pore system that is commonly spatially related enhance reservoir connectivity, particularly if they
to some, but not all, seismically resolvable faults are later enhanced by dissolution. This stratigraphi-
(concept 3 in Figures 17 and 18). We therefore inter- cally discordant syndepositional fracture system
pret this nonmatrix pore system as associated to silici- would only be compatible with a platform margin
fication around fault zones and potentially hydraulic interpretation. Other authors have argued for pre-
brecciation at multiple scales in fault damage zones Salt carbonate deposition (Barra Velha Formation) in
(Menezes et al., 2019). Recurrent fault slip events a shallow lake with no significant margin relief and
would form new fractures within an already brittle propose that the present-day structures of the Barra
(silicified) damage zone combined with dissolution Velha reservoirs are predominantly related to tec-
from ascending fluids. Faults resolvable in seismic tonic deformation after deposition (Wright and Bar-
data and their subseismic scale equivalents are a first- nett, 2015; Wright, 2020). These two alternative
order constraint for spatial prediction. Fault throw end-member depositional models (Wright, 2020)
can be used to predict the width of the damage zone reinforce the critical importance of interpreting pale-
and the fracture intensity based on empirical rules ostructure for predicting the spatial distribution of
(Solum and Huisman, 2017). Two observations com- certain types of nonmatrix features in addition to
plicate this approach. First, there is no guarantee that depositional facies in the pre-Salt reservoirs of Brazil.
all seismically resolvable faults are associated with
this nonmatrix pore system that has excess perme- Karst—There are two additional interpretations for
ability. Second, the degree of silicification, which for the meter-scale vugs or caves observed in image logs
the most part is associated with this nonmatrix pore and interpreted from the rate of mud losses while
system, is variable both vertically and laterally in pre- drilling. One hypothesis is that they are the result
Salt reservoirs. of hypogenic karst processes. Hypogenic caves are
porous structures that are solution-enlarged by
Margin Failure—Meter-scale fractures in core partially upward-flowing groundwater (Ehrenberg, 2007;
filled with reworked carbonate grains (Figure 4E) can Klimchouk, 2007; Jones and Xiao, 2013). Cave pat-
be interpreted as syndepositional features that terns are strongly guided by contrasts in permeability
develop in the proximity of platform margins (Frost such as fractures and bedding planes (Cazarin et al.,
and Kerans, 2009; Ibrayev et al., 2016; Nolting et al., 2019; Balsamo et al., 2020). This would be consistent
2018) (concept 4 in Figures 17 and 18). The exis- with the occurrence of some of these features in the
tence of platform margins during the deposition of proximity of seismically resolvable faults (e.g., Oli-
the pre-Salt lacustrine reservoirs is interpreted from veira et al., 2019).
seismic data that show well imaged clinoforms with Another hypothesis is that some of the caves are
hundreds of meters of bathymetry and relatively generated in an epigenic karst system in a version of
steep margins up to 30 (Simo et al., 2019; Minzoni the island karst model (Mylroie and Mylroie, 2017)
et al., 2020; Ferreira et al., 2021). Similar processes modified for lacustrine environments (concept 5 in
have been postulated for fractures observed in core Figures 17 and 18). These caves can later be buried
and image logs in the Campos Basin (Wennberg et al., and remain open at present-day reservoir depths
2019). The spatial prediction of these stratigraphi- (Nolting et al., 2021). Exposed paleohighs during
cally discordant syndepositional fractures is informed lake level lowstands would have resulted in intralake
by the mapping of the platform margin in time and islands with the potential to support a freshwater
space and empirical rules derived from process-based lens. Such freshwater lenses could be thicker or thin-
analogues (Frost and Kerans, 2009). These syndepo- ner atop lake water in comparison to seawater
sitional fractures would be predominantly orientated depending on the lake water salinity (all other factors
parallel to platform margins with occasional perpen- equal for example, climate, permeability, and area of
dicular fracture sets (Collins et al., 2006; Narr and exposure). The migration of such a freshwater lens
Flodin, 2012; Ibrayev et al., 2016). Such fractures are and associated mixing zone with evolved lake waters

in response to fluctuations in lake level could have log and particularly core data are required to deter-
resulted in dissolution and the generation of a strati- mine if these slope breccias are dominantly deposited
graphically discordant nonmatrix karst pore system as discrete tongues or back filled aprons in a dip pro-
especially prolific along the margins of the paleo- file. Based on the conceptual models of Playton et al.
highs. Nonmatrix karst in certain heavily silicified (2010), we predict slope breccias to be generally
pre-Salt reservoirs with excess permeability in the strike continuous as evidenced by pervasive scallops
Campos Basin have in part been interpreted as the along pre-Salt platform margins imaged in seismic
result of a modified island karst model (Hunt et al., data. When penetrated by a wellbore, these rocks
2019). First-order predictions for the spatial distribu- may cause operational issues including core jams and
tion of this nonmatrix pore system requires knowl- lost returns (type I, II, and III) and can show a gentle
edge of paleostructure to identify regions of pre-Salt roll-over on the pressure derivative with darcy-scale
reservoirs that were more likely to be exposed. This permeability (concept 6 in Figure 17).
is not trivial to reconstruct because of the often- Core observations also show breccias with vari-
complex polyphase faulting and post salt deforma- able reservoir quality found at some cycle caps and
tion (Minzoni et al., 2020; Wright, 2020). Since are interpreted to be the product of subaerial expo-
many of the pre-Salt paleohighs are interpreted to be sure (Eleson et al., 2014). Although these breccias
structurally controlled, it is likely that a freshwater related to exposure are relatively thin, they can have
lens will intersect numerous subvertical faults that substantial lateral extent. Other types of breccias are
could further modify fluid flow and the distribution spatially more restrictive. This is the case of hydraulic
of dissolution in the intralake island scenario (combi- breccias that develop in fault zones (e.g., Menezes
nation of concepts 3, 4, and 5 in Figure 17). More et al., 2019; concept 3 in Figure 17) or chaotic and
sophisticated quantitative spatial predictions of fresh- crackle breccias associated with the partial collapse of
water lens and mixing zone thickness and residence caves in karst settings (e.g., Loucks, 1999; associated
times are possible by adapting methods designed to with concepts 3, 4, and 5 in Figure 17).
investigate freshwater lens diagenesis in marine car-
bonates (Budd and Vacher, 1991; Paterson et al., Implications for pre-Salt Exploration,
2008). Appraisal, Field Development,
and Production
Breccias—Multiple processes form breccias and con-
glomerates through physical weathering, brittle fail- The weight of evidence from the data and interpreta-
ure, and gravitational collapse causing fragmentation tions presented indicates that nonmatrix features
of the original rock. Carbonate fields producing from with the potential for excess permeability are a rela-
brecciated intervals show evidence of excess perme- tively common occurrence in the pre-Salt reservoirs
ability and can be especially prolific (Collins et al., of Brazil. Generally good matrix reservoir properties,
2006). Breccias are present in the pre-Salt, and they but with noted variability, would lead to not classify
can be the generated by several different processes. the pre-Salt play as traditional “fractured” or “karst”
Concept 6 in Figures 17 and 18 illustrates one of reservoirs (Sun and Sloan, 2003). Instead, we consider
these processes, which involves gravitational collapse these carbonate lacustrine reservoirs, with a highly
of high-relief platform or mound margins, as inter- heterogeneous matrix reservoir mineralogy and pore
preted in the Buzios field (Ferreira et al., 2021). This system, as modified by fractures and karst. This hybrid
slope margin failure causes resedimentation from the reservoir type creates some distinct challenges with
structural highs to slopes forming poorly sorted brec- respect to hydrocarbon exploitation business stages.
cias along with other rock types (Playton et al.,
2010). Slope breccias can vary from clast-supported Exploration
to mud-supported depending on the slope angle From an exploration perspective, in an undrilled
(Kenter, 1990). In our pre-Salt experience, heteroge- prospect, some, but not all, types of nonmatrix fea-
neous clast-dominated breccias commonly exhibit tures may be imaged with seismic data, the quality of
high amounts of interclast porosity and correspond which depends on survey design and processing
to intervals of observed excess permeability. More methods. Faults and by association their fracture

damage zones (concept 3 in Figures 17 and 18) are associated silica can severely limit core recovery and
the most discernible. Exposure-related karst surfaces prove costly with repeat attempts in deep-water
and caves (concepts 1 and 5 in Figures 17 and 18) are wells. This can make the value proposition of includ-
more rarely imaged and the interpretation confidence ing a WT as part of an exploration well (success case)
is considerably lower than fault damage zones. Slope data-acquisition strategy attractive, considering its
deposits are relatively well imaged; however, delinea- ability to help identify the presence of excess perme-
tion of slope breccias with potential nonmatrix flow ability and nonmatrix flow behavior (Figure 12). The
behavior (concept 6 in Figures 17 and 18) are more presence of nonmatrix features, if connected, can
challenging to confidently interpret predrill (e.g., fig- result in the rapid transmission of pressure transients
ure 10B in Ferreira et al., 2021). In any case, an over several kilometers (e.g., Saurbayev et al., 2018).
exploration well generally prioritizes testing other Installing a long-term downhole pressure monitoring
play elements, for example trap, charge, reservoir system in a successful exploration well to take advan-
presence, and matrix reservoir quality ahead of non- tage of future appraisal WTs to reduce uncertainty in
matrix. In fact, most exploration wells deliberately interwell reservoir connectivity may also be prudent
avoid penetrating seismically interpreted nonmatrix (Champion and Puntel, 2015).
features to reduce operational risk. An optimal
exploration well is therefore one that satisfies the
Appraisal
evaluation of key play elements and has seismically
From a field appraisal perspective, the presence of
resolvable nonmatrix features interpreted in the
nonmatrix features in the pre-Salt reservoirs of Brazil
radius of investigation of a WT that are not present at
increases subsurface complexity and dynamic uncer-
the well location.
The risk of encountering nonmatrix features in tainty that is already high in this play with limited
pre-Salt reservoirs during drilling operations is high applicable analogues (Moczydlower et al., 2012).
because they are abundant and most are subseismic The location of candidate appraisal wells should in
in scale. This has several implications. First, in antici- part be targeted toward evaluating credible scenarios
pation of likely drilling losses in the reservoir section, of nonmatrix features and prioritizing dynamic data,
the appropriate amount of drilling fluids and LCM for example routine collection of WT and PLT data.
needs to be available to cure any mud losses. It is recommended that appraisal wells with the pri-
Although most losses experienced in the wells we mary objective of derisking other subsurface uncer-
examined were relatively moderate, a subset associ- tainties, for example oil-water contact, reservoir
ated with solution enhanced fractures and caves were compartmentalization, fluid variability, and matrix
more severe (e.g., type IIIa in Figure 11) to the point reservoir quality, consider an ancillary objective of
that nonproductive time (NPT) increased exponen- appraising for nonmatrix features. The number, loca-
tially and a sacrificial fluid system was required to tion, and timing of appraisal wells is determined by
handle the uncurable losses (Yaqoob et al., 2017). the field development concept and strategy, for
Managed pressure drilling techniques in these situa- example phased versus full field, and should be
tions have proven to successfully reduce NPT in value-based.
nonmatrix-dominated wells (e.g., Sweep et al., 2003; The presence of nonmatrix features necessitates
Pontes et al., 2018). Second, in a situation where the additional design criteria and increases operational
matrix is poor quality, often because of silica replace- complexity and risk. Nonmatrix can lead to the
ment and cementation, the associated nonmatrix inability to cement production casing strings or liners,
pore system will be difficult to interpret from side- resulting in erratic pressure response during WTs and
wall cores and logs alone, which will tend to grossly potential flow behind casing that complicates the
underestimate reservoir quality (Hunt et al., 2019). interpretation of PLTs. For WT programs that
Conventional coring may have more chance of iden- involve acid stimulation in the presence of nonma-
tifying certain types of nonmatrix features but is a trix, the delivery of acid must be sufficiently engi-
greater financial commitment in an exploration well. neered with advanced diversion, so all the acid is not
In addition, poor recovery related to jamming associ- pumped into the nonmatrix features and the matrix
ated with certain nonmatrix features and commonly is stimulated.

Early production pilots and EWTs that investi- handling capacity, opting for smart well completions
gate a much larger swath of the reservoir, have also to mitigate preferential flow in nonmatrix features
been used in the pre-Salt to further reduce dynamic and selecting appropriate flexibility in a subsea layout
uncertainty, including the long-term effect of nonma- (Moczydlower et al., 2012; Turazzi Naveiro and
trix features at a larger scale, although these data are Haimson, 2015; Rosa et al., 2018).
not currently publicly available (Moczydlower et al.,
2012; Turazzi Naveiro and Haimson, 2015). Con- Production
ventional core collection should be prioritized to Upon production startup, risks associated with early
help calibrate other data types and develop rules for breakthrough of injected fluids and/or aquifer-
interpretation away from core control, to help derived water through nonmatrix features with
mature predictive models of nonmatrix-related excess permeability should be an integral part of any
excess permeability. Deployment of long-term moni- mitigation strategy. Production history matching will
toring gauges in wells within reasonable distances of likely be more complicated for these reservoirs, and
future appraisal wells with testing programs should certain types of nonmatrix features will require more
also be considered. sophisticated DPDK reservoir models (Correia et al.
2011; Correa et al., 2019; Stephens et al., 2020).
Development Producer wells completed in caves, although they
Following appraisal, the presence of nonmatrix fea- can deliver high rates, can also present challenges
tures with excess permeability in the pre-Salt can from a casing integrity and reservoir management
have several important ramifications for selecting a perspective (Pe~ na et al., 2019; Firme et al., 2021).
field development concept and its design. First, a From a reservoir monitoring perspective, the extent
comprehensive integration across multiple data sets to which four-dimensional seismic data and other
and disciplines is required to adequately characterize technologies, for example distributed acoustic sens-
a carbonate reservoir with nonmatrix features. Sec- ing, will be able to identify differential sweep related
ond, it is possible to generate multiple credible reser- to the presence of nonmatrix features in pre-Salt res-
voir scenarios consistent with appraisal data given the ervoirs is yet to be determined but could be impor-
variety of nonmatrix features observed in the pre- tant for reservoir management (Lopez et al., 2017;
Salt and the conceptual models proposed (Figures Cypriano et al., 2019). Finally, depending on the ini-
17, 18). In the fast-paced pre-Salt development proj- tial reservoir stresses, nonmatrix features (especially
ects, it is only practically feasible to evaluate a subset
fractures) might be more stress sensitive, resulting
of these scenarios with reservoir model realizations.
in a permeability reduction in the near wellbore
Third, reservoir model design is generally more com-
(mechanical skin) or across the reservoir (Crawford
plicated in a reservoir with nonmatrix features (Ring-
et al., 2018) as a function of drawdown and
rose and Bentley, 2015). Sophisticated reservoir
depletion, respectively. This results in a pressure-
model strategies that include discrete fracture net-
dependent nonmatrix compressibility that remains a
work models (Dershowitz et al., 1998; Vaughan,
key uncertainty in pre-Salt reservoirs (e.g., Moczy-
2015), effective property models of karst (Jones,
dlower et al., 2012).
2015) and DPDK simulation (Tankersley et al.,
In summary, the abundance of nonmatrix fea-
2010; Correia et al., 2011; Correa et al., 2019)
tures in the pre-Salt reservoirs of Brazil has the net
extend the modeling timeline and are computation-
effect of increasing subsurface complexity and uncer-
ally expensive to simulate. In contrast, more
“efficient” reservoir models that ignore or simplify tainty. This can increase project capital expenditure,
the incorporation of nonmatrix features run the risk through both acquiring additional appraisal data to
of predicting tanklike dynamic performance that reduce nonmatrix-related uncertainties and paying
overestimates recovery and breakthrough times of for additional flexibility to mitigate their effects
injected fluids. Since partial-to-full gas injection and (Moczydlower et al., 2012). However, as demon-
water alternating gas are common development con- strated by numerous global reservoir examples and
cepts in pre-Salt reservoirs, reliably predicting flow is the world-class production performance in many of
an important value driver related to sizing gas the Brazil pre-Salt fields, the challenges associated

with nonmatrix features and excess permeability in an emphasis on drilling data, conventional core,
carbonate reservoirs are not insurmountable (Turazzi borehole image logs, and WTs for the effective
Naveiro and Haimson, 2015; Rosa et al., 2018). The characterization of the diverse multiscale nonma-
integrated workflows and preliminary conceptual trix features present in the pre-Salt reservoirs of
models of nonmatrix features in pre-Salt reservoirs Brazil.
presented are designed to enhance operational effi- Preliminary conceptual models of nonmatrix
ciency, support scenario-based analysis to manage features provide a process-based framework to
spatially predict their distribution. When superim-
uncertainty and value-of-information-based appraisal
posed with compatible depositional and diagenetic
strategies, reduce and optimize characterization and
models these nonmatrix concepts formulate credi-
modeling cycle time, support the selection, and
ble alternative subsurface scenarios to evaluate for
design of robust depletion plans and help formulate
flow significance. Antecedent topography during
reservoir management strategies to maximize hydro-
the deposition of the pre-Salt lacustrine reservoirs
carbon recovery. is required for several of these conceptual models
and reliable paleostructure reconstructions are
important for predicting the distribution of non-
CONCLUSIONS
matrix features.
Drilling experience is commonly underused for
Nonmatrix features are a common occurrence in
reservoir characterization. The VvR plot of losses
the pre-Salt reservoirs of Brazil. Many, but not
while drilling provides a rapid first-order assess-
all, exhibit excess permeability that can be orders
ment to identify the presence and type of non-
of magnitude greater than that measured from
matrix features and their relative significance for
RCA.
fluid flow.
Multiple types and scales of nonmatrix features
Particularly valuable for identifying flow parti-
have been observed and characterized, including
tioning along the borehole, as well as the pres-
touching vugs in shrub boundstones; BBFs in silica
ence and magnitude of excess permeability that
layers; fractures that are commonly solution
can be related to nonmatrix features, are WTs,
enhanced with associated vugs in pervasively silici-
especially when combined with PLTs. Approxi-
fied carbonates in close proximity to seismic mately 50% of WTs show some evidence of
scale faults; syndepositional fractures filled in part excess permeability related to a dual-porosity
with transported carbonate clasts, meter-scale system. The majority of WTs with nonmatrix
caves, and a variety of breccia types, including excess permeability had one of four characteristic
slope deposits that include fractured cobble size pressure derivative type curves. These type
clasts with open intraclast porosity and karst curves, in combination with other diagnostic indi-
breccias. cators, can be used to discriminate between alter-
Based on core observations, we recognize three native nonmatrix geologic concepts. The textbook
silica types: silica I (a concordantly replacing “V-shape” pressure derivative type curve indica-
phase, most abundant and prone to develop non- tive of nonmatrix flow behavior is rare in the pre-
matrix), silica II (chalcedony that develops as a Salt WTs examined for this study.
pore lining cement and it is abundant in areas of Evaluating alternative nonmatrix scenarios using
increased nonmatrix that exhibit excess perme- WT forward numerical models proved highly
ability), and silica III (late pore filling cement informative for interpreting pressure derivative
that partly occludes nonmatrix). Although the curves and “seeing the geology” in the heteroge-
source of different types of silica remains poorly neous pre-Salt reservoirs with complex architec-
constrained, observations support a first-order ture. Heterogeneity interpreted from different
relationship between silicified intervals and pres- pressure derivative type curves described in clas-
ence of excess permeability. sic well testing reference books should be used
The multidisciplinary integrated workflow pre- with caution because the incorporation of non-
sented uses a suite of subsurface data types with matrix features can result in viable alternative

interpretations with material implications for res- Budd, D. A., and H. L. Vacher, 1991, Predicting the thickness
ervoir characterization. of fresh-water lenses in carbonate paleo-islands: Journal
of Sedimentary Research, v. 61, no. 1, p. 43–53.
The presence of nonmatrix features with excess
Cazarin, C. L., F. H. R. Bezerra, L. Borghi, R. V. Santos, J.
permeability in the pre-Salt reservoirs of Brazil Favoreto, J. A. Brod, A. S. Auler, and N. K. Srivastava,
creates distinct, but not insurmountable, chal- 2019, The conduit-seal system of hypogene karst in Neo-
lenges across hydrocarbon exploitation business proterozoic carbonates in northeastern Brazil: Marine
stages. In general, these nonmatrix features have and Petroleum Geology, v. 101, p. 90–107, doi:10.1016/
j.marpetgeo.2018.11.046.
the net effect of increasing subsurface uncer- Chaboureau, A.-C., F. Guillocheau, C. Robin, S. Rohais, M.
tainty. This impacts project capital expenditure Moulin, and D. Aslanian, 2013, Paleogeographic evolu-
both through acquiring additional appraisal data tion of the central segment of the South Atlantic during
to reduce nonmatrix-related uncertainties or pay- Early Cretaceous times: Paleotopographic and geody-
namic implications: Tectonophysics, v. 604, p. 191–223,
ing for additional flexibility to mitigate their
doi:10.1016/j.tecto.2012.08.025.
effects. Champion, B. P., and E. A. Puntel, 2015, Reducing reservoir
uncertainty during appraisal and development - Novel
applications of a new wireless monitoring technology in
REFERENCES Santos Basin Pre-Salt: Society of Petroleum Engineers
Annual Technical Conference and Exhibition, Houston,
Aguilera, R., 1982, Relative permeability concepts for predict- Texas, September 28–30, 2015, SPE-175075-MS, 16 p.,
ing the performance of naturally fractured reservoirs: doi:10.2118/175075-MS.
Journal of Canadian Petroleum Technology, v. 21, no. 5, Chitale, V. D. V., G. Alabi, P. Gramin, S. Lepley, and L. Pic-
p. 41–48, doi:10.2118/82-05-03. coli, 2015, Reservoir characterization challenges due to
Ahdyar, L. O., R. P. Setki, S. R. Mohammad, F. Fernandez- the multiscale spatial heterogeneity in the Presalt carbon-
Ibanez, and P. J. Moore, 2019, Integrated carbonate non- ate sag formation, North Campos Basin, Brazil: Petrophy-
matrix characterization in Banyu Urip field: Proceedings, sics, v. 56, no. 6, p. 552–576.
Indonesian Petroleum Association 43rd Annual Conven- Choquette, P. W., and L. C. Pray, 1970, Geologic nomencla-
tion and Exhibition, Jakarta, Indonesia, September 4-6, ture and classification of porosity in sedimentary carbo-
2019, IPA19-G-35, 17 p. nates.: AAPG Bulletin, v. 54, no. 2, p. 207–250.
Balsamo, F., F. H. R. Bezerra, A. B. Klimchouk, C. L. Cazarin, Collins, J. F., J. A. M. Kenter, P. M. Harris, G. Kuanysheva,
A. S. Auler, F. C. Nogueira, and C. Pontes, 2020, Influ- D. J. Fischer, and K. L. Steffen, 2006, Facies and
ence of fracture stratigraphy on hypogene cave develop- reservoir-quality variations in the late Visean to Bashkir-
ment and fluid flow anisotropy in layered carbonates, NE ian outer platform, rim, and flank of the Tengiz buildup,
Brazil: Marine and Petroleum Geology, v. 114, 104207, Precaspian Basin, Kazakhstan, in P. M. Harris and L. J.
Weber, eds., Giant hydrocarbon reservoirs of the
14 p., doi:10.1016/j.marpetgeo.2019.104207.
world: From rocks to reservoir characterization and
Behr, H. J., 2002, Magadiite and Magadi chert: A critical anal-
modeling: SEPM Special Publication and AAPG Memoir
ysis of the silica sediments in the Lake Magadi Basin,
88, p. 55–95, doi:10.1306/1215874M881469.
Kenya, in R. W. Renaut and G. M. Ashley, eds., Sedi-
Contreras, J., R. Z€uhlke, S. Bowman, and T. Bechst€adt, 2010,
mentation in continental rifts: Tulsa, Oklahoma, SEPM
Seismic stratigraphy and subsidence analysis of the south-
Special Publication 73, 257–273, doi:10.2110/pec.02.73.
ern Brazilian margin (Campos, Santos and Pelotas
0257. basins): Marine and Petroleum Geology, v. 27, no. 9,
Benson, L., 1994, Carbonate deposition, Pyramid Lake subba-
p. 1952–1980, doi:10.1016/j.marpetgeo.2010.06.007.
sin, Nevada: 1. Sequence of formation and elevational Correa, R., C. Pereira, F. Cruz, S. Lisboa, M. Junior, V. Souza,
distribution of carbonate deposits (Tufas): Palaeogeogra- C. Rocha, F. Araujo, and B. Melo de Carvalho, 2019,
phy, Palaeoclimatology, Palaeoecology, v. 109, no. 1, Integrated seismic-log-core-test fracture characterization
p. 55–87, doi:10.1016/0031-0182(94)90118-X. and modelling, Barra Velha formation, Pre-Salt of Santos
Bettermann, P., and F. Liebau, 1975, The transformation of Basin (abs.): AAPG Annual Convention and Exhibition,
amorphous silica to crystalline silica under hydrothermal San Antonio, Texas, May 19–22, 2019, accessed May 20,
conditions: Contributions to Mineralogy and Petrology, 2020, http://www.searchanddiscovery.com/abstracts/
v. 53, no. 1, p. 25–36, doi:10.1007/BF00402452. html/2019/ace2019/abstracts/356.html.
Bourbiaux, B., 2010, Fractured reservoir simulation: A chal- Correia, M., C. Maschio, D. J. Schiozer, and M. Santos, 2011,
lenging and rewarding issue: Oil & Gas Science and Tech- Sensitivity analysis for dual porosity and dual permeabil-
nology—Revue d’IFP, v. 65, no. 2, p. 227–238, doi:10. ity systems to small and large scale heterogeneities in a
2516/ogst/2009063. naturally fractured carbonate reservoir: Society of Petro-
Bourdet, D., 2002, Well test analysis: The use of advanced leum Engineers Brasil Offshore Conference and Exhibi-
interpretation models: Amsterdam, Elsevier Science, v. tion, Macae, Brazil, June 14–17, 2011, SPE-143149-MS,
3, 416 p. 7 p., doi:10.2118/143149-MS.

Crawford, B., J. Homburg, F. Fernandez-Ibanez, R. D. Myers, well, BM-S-22, Santos Basin, offshore Brazil: AAPG
and B. Gao, 2018, Predicting pore volume and transmis- Annual Convention and Exhibition, Houston, Texas,
sibility multipliers for simulating geomechanical effects April 6–9, 2014, accessed May 19, 2019, https://www.
in naturally fractured reservoirs using stress-sensitive dis- searchanddiscovery.com/abstracts/html/2014/90189ace/
crete fracture network modeling: 2nd International Dis- abstracts/1836145.html.
crete Fracture Network Engineering Conference, Seattle, Eugster, H. P., 1967, Hydrous sodium silicates from Lake
Washington, June 20–22, 2018, ARMA-DFNE-18- Magadi, Kenya: precursors of bedded chert: Science, v.
0429, 11 p. 157, no. 3793, p. 1177–1180, doi:10.1126/science.157.
Cypriano, L., Z. Yu, D. Ferreira, B. Huard, R. Pereira, F. 3793.1177.
Jouno, A. Khalil, et al., 2019, OBN for Pre-Salt imaging Farias, F., P. Szatmari, A. Bahniuk, A. Barros França, 2019,
and reservoir monitoring—Potential and road ahead: Evaporitic carbonates in the pre-salt of Santos Basin—
16th International Congress of the Brazilian Geophysical Genesis and tectonic implications: Marine and Petroleum
Society, Rio de Janeiro, Brazil, August 19–22, 2019, 6 p. Geology, v. 105, p. 251–272, doi:10.1016/j.marpetgeo.
de Jesus, C. M., A. L. Martins Compan, and R. Surmas, 2016, 2019.04.020.
Permeability estimation using ultrasonic borehole image Fernandez-Iban ~ ez, F., J. M. DeGraff, and F. Ibrayev, 2018,
logs in dual-porosity carbonate reservoirs: Petrophysics, Integrating borehole image logs with core: A method to
v. 57, no. 6, p. 620–637. enhance subsurface fracture characterization: AAPG Bul-
de Luca, P. H. V., H. Matias, J. Carballo, D. Sineva, G. A. letin, v. 102, no. 6, p. 1067–1090.
Pimentel, J. Tritlla, M. Esteban, et al., 2017, Breaking Fernandez-Iban ~ ez, F., J. M. DeGraff, P. J. Moore, L. Ahdyar,
barriers and paradigms in presalt exploration: the P~ao de and A. Nolting, 2019a, Characterization of non-matrix
Açucar discovery (offshore Brazil), in R. K. Merrill and type and flow potential using lost circulation information:
C. A. Sternbach, eds., Giant fields of the decade Journal of Petroleum Science Engineering, v. 180, p. 89–
2000–2010: AAPG Memoir 113, p. 177–194, doi:10. 95, doi:10.1016/j.petrol.2019.05.023.
1306/13572007M1133686. Fernandez-Iban ~ ez, F., P. J. Moore, and G. D. Jones, 2019b,
Demott, L. M., C. A. Scholz, and C. K. Junium, 2019, 8200-
Quantitative assessment of karst pore volume in carbon-
year growth history of a Lahontan-age lacustrine tufa
ate reservoirs: AAPG Bulletin, v. 103, no. 5, p. 1111–
deposit: Sedimentology, v. 66, no. 6, p. 2169–2190, doi:
1131, doi:10.1306/10261818061.
10.1111/sed.12579.
Ferreira, D. J. A., H. P. L. Dutra, T. M. de Castro, and W. M.
de Paula Faria, D. L., A. T. dos Reis, and O. G. de Souza, Jr.,
Lupinacci, 2021, Geological process modeling and geo-
2017, Three-dimensional stratigraphic-sedimentological
statistics for facies reconstruction of presalt carbonates:
forward modeling of an Aptian carbonate reservoir
Marine and Petroleum Geology, v. 124, 14 p.
deposited during the sag stage in the Santos basin, Brazil:
Firme, P. A. L. P., R. J. Quevedo, D. Roehl, L. C. Pereira, and
Marine and Petroleum Geology, v. 88, p. 676–695, doi:
C. L. Cazarin, 2021, Mechanical behavior of carbonate
10.1016/j.marpetgeo.2017.09.013.
reservoirs with single karst cavities: Geomechanics for
Dershowitz, W., T. Eiben, E. Wadleigh, and T. Cladouhos,
Energy and the Environment, v. 25, 100209, doi:10.
1998, Discrete feature network approaches for enhanced
oil recovery: International Journal of Rock Mechanics 1016/j.gete.2020.100209.
Freites, A., P. Corbett, S. Geiger, and J. P. Norgard, 2019,
and Mining Sciences, v. 35, no. 4–5, p. 550, doi:10.1016/
S0148-9062(98)00150-8. Macro insights from interval pressure transient tests:
Dias, J. L., J. Q. de Oliveira, and J. C. Vieira, 1988, Sedimen- Deriving key near-wellbore fracture parameters in a light
tological and stratigraphyc analysis of the Lagoa Feia For- oil reservoir offshore Norway: Society of Petroleum Engi-
mation, rift phase of Campos Basin, offshore Brazil: neers Europec featured at 81st EAGE Conference and
Revista Brasileira de Geociencias, v. 18, no. 3, p. 252– Exhibition, London, June 3-6, SPE-195435-MS, 32 p.,
260, doi:10.25249/0375-7536.1988252260. doi:10.2118/195435-MS
Duarte, S. B., C. M. De Jesus, V. F. Da Silva, M. C. Arouca Frost, E., III, and C. Kerans, 2009, Platform-margin trajectory
Sobreira, R. A. L. Cristofaro, L. de Lima, F. G. De Mello as a control on syndepositional fracture patterns, Canning
E Silva, et al., 2018, Artificial intelligence use to predict Basin, Western Australia: Journal of Sedimentary
severe fluid losses in Pre-Salt carbonates: Society of Pet- Research, v. 79, no. 2, p. 44–55, doi:10.2110/jsr.2009.
rophysicists and Well Log Analysts 59th Annual Logging 014.
Symposium, London, June 2–6, 2018, 23 p. Fullmer, S., H. Al Qassab, A. Buono, B. Gao, B. Kelly, and
Ehrenberg, S. N., 2007, Whole core versus plugs: scale depen- P. J. Moore, 2019, Carbonate pore-system influence on
dence of porosity and permeability measurements in hydrocarbon displacement and potential recovery, in
platform carbonates: AAPG Bulletin, v. 91, no. 6, D. F. McNeill and P. M. Harris, eds., Carbonate pore sys-
p. 835–846, doi:10.1306/01090706093. tems: New developments and case studies: Tulsa, Okla-
Eleson, J. W., K. Bohacs, B. Carter, J. DeGraff, T. M. Demko, homa, SEPM Special Publication 112, p. 268–284. doi:
M. Fabijanic, M. French, et al., 2014, Conventional core 10.2110/sepmsp.112.14
as a Rosetta Stone for Presalt RQ: Insights and observa- Gaillot, P., T. Brewer, P. Pezard, and Y. En-Chao, 2007, Bore-
tions within Upper Sag Presalt microbialites from con- hole imaging tools—Principles and applications: Scientific
ventional core of the 3-ESSO-4A-SPS (Guarani-1ST) Drilling, v. 5, no. 5, p. 1–4. doi:10.2204/iodp.sd.5.07.2007.

Genter, A., C. Castaing, C. Dezayes, H. Tenzer, H. Traineau, March 8–11, 2015, SPE-172648-MS, 8 p., doi:10.2118/
and T. Villemin, 1997, Comparative analysis of direct 172648-MS
(core) and indirect (borehole imaging tools) collection of Jones, G. D. and Y. Xiao, 2013, Geothermal convection in
fracture data in the Hot Dry Rock Soultz reservoir South Atlantic subsalt lacustrine carbonates: Developing
(France): Journal of Geophysical Research, v. 102 no. B7, diagenesis and reservoir quality predictive concepts with
p. 15419–15431, doi:10.1029/97JB00626. reactive transport models, AAPG Bulletin, v. 97, no. 8,
Ghosh, B., S. A. Ali, and H. Belhaj, 2020, Controlling excess p. 1249–1271.
water production in fractured carbonate reservoirs: Kamal, M. M., 2009, Transient well testing: Richardson,
chemical zonal protection design: Journal of Petroleum Texas, Society of Petroleum Engineers Monograph
Exploration and Production Technology, v. 10, no. 5, p. Series, v. 23, 850 p.
1921–1931, doi:10.1007/s13202-020-00842-3. Karner, G., C. Johnson, J. Shoffner, M. Lawson, M. Sullivan, J.
Gomes, J. P., R. B. Bunevich, L. R. Tedeschi, M. E. Tucker, Sitgreaves, J. McHarge, J. Stewart, and P. Figueredo,
and F. F. Whitaker, 2020, Facies classification and pat- 2021, Tectono-magmatic development of the Santos and
terns of lacustrine carbonate deposition of the Barra Campos Basins, offshore Brazil, in M. Mello, P. Yilmaz,
Velha Formation, Santos Basin, Brazilian Pre-salt: Marine and B. Katz, eds., The supergiant Lower Cretaceous Pre-
and Petroleum Geology, v. 113, 104176, doi:10.1016/j. Salt petroleum system of the Santos Basin, Brazil: AAPG
marpetgeo.2019.104176. Memoir 124, p. 215–256.
Gomez, F. J., C. Mlewski, F. J. Boidi, M. E. Farıas, and E. Kenter, J. A. M., 1990, Carbonate platform flanks: Slope
Gerard, 2018, Calcium carbonate precipitation in angle and sediment fabric: Sedimentology, v. 37, no. 5, p.
diatom-rich microbial mats: the Laguna Negra hypersa- 777–794, doi:10.1111/j.1365-3091.1990.tb01825.x.
line lake, Catamarca, Argentina: Journal of Sedimentary Klimchouk, A. 2007, Hypogene speleogenesis: Hydrogeologi-
Research, v. 88, no. 6, p. 727–742, doi:10.2110/jsr.2018. cal and morphogenetic perspective: Carlsbad, New Mex-
ico, National Cave and Karst Research Institute, Special
37.
Grimes, K. G., 2006, Syngenetic karst in Australia: A review: Paper no. 1, 106 p.
Kuchuk, F., and D. Biryukov, 2015, Pressure-transient tests
Helictite, v. 39, no. 2, p. 27–38.
and flow regimes in fractured reservoirs: SPE Reservoir
Harder, H., 1965, Experimente zur “Ausf€allung” der Kie-
Evaluation & Engineering, v. 18, no. 2, p. 82–97, doi:10.
sels€aure: Geochimica et Cosmochimica Acta, v. 29, no.
2118/166296-PA.
5, p. 429–442, doi:10.1016/0016-7037(65)90037-2.
Laubach, S. E., R. W. Baumgardner, Jr, E. R. Monson, E.
Herlinger, R., Jr., E. E. Zambonato, and L. F. De Ros, 2017,
Hunt, and K. J. Meador, 1988, Fracture detection in low-
Influence of diagenesis on the quality of lower Creta-
permeability reservoir sandstone: A comparison of
ceous Pre-Salt lacustrine carbonate reservoirs from
BHTV and FMS logs to core: Society of Petroleum Engi-
northern Campos basin, offshore Brazil: Journal of Sedi-
neers Annual Technical Conference and Exhibition,
mentary Research, v. 87, no. 12, p. 1285–1313, doi:10.
Houston, Texas, October 2–5, 1988, SPE-18119-MS,
2110/jsr.2017.70.
doi:10.2118/18119-MS.
Heubeck, C., K. Story, P. Peng, C. Sullivan, and S. Duff,
Lee, J., 1982, Well testing: Richardson, Texas, Society of
2004, An integrated reservoir study of the Liuhua 11-1
Petroleum Engineers Textbook Series, v. 1, 159 p.
field using a high-resolution three-dimensional seismic Lima, B. E. M., and L. F. De Ros, 2019, Deposition, diage-
data set, in G. P. Eberli, J. L. Massaferro, and R. F. Sarg, netic and hydrothermal processes in the Aptian Pre-Salt
eds., Seismic imaging of carbonate reservoirs and systems: lacustrine carbonate reservoirs of the northern Campos
AAPG Memoir 81, p. 149–168. Basin, offshore Brazil: Sedimentary Geology, v. 383,
Hunt, D. W., P. H. Viera de Luca, L. E. Gomis-Cartesio, F. p. 55–81, doi:10.1016/j.sedgeo.2019.01.006.
Lapponi, I. Masiero, A. Waldum, A. S. Chandler, et al., Lopez, J., A. Mateeva, D. Chalenski, and J. Przybysz-Jarnut,
2019, A very different Barremian-Aptian lacustrine Pre- 2017, Valuation of distributed acoustic sensing VSP for
Salt facies association: Biotic self-organization in BMC- frequent monitoring in deepwater: Society of Exploration
33, outer Basin Campos Basin, Brazil: First European Geophysicists International Exposition and 87th Annual
Association of Geoscientists and Engineers Workshop on Meeting, Houston, Texas, September 24-29, 2017,
Pre-Salt Reservoir: From Exploration to Production, Rio p. 6044–6048, doi:10.1190/segam2017-17685577.1.
de Janeiro, Brazil, December 5–6, 2019, 5 p., doi:10. Loucks, R. G., 1999, Paleocave carbonate reservoirs: Origins,
3997/2214-4609.201982102. burial-depth modifications, spatial complexity, and
Ibrayev, F., F. Fernandez-Ibanez, and J. M. DeGraff, 2016, reservoir implications: AAPG Bulletin, v. 83, no. 11,
Using a genetic-based approach to enhance natural frac- p. 1795–1834.
ture characterization in a giant carbonate field: Society of Macedo, J. M., 1989, Evoluç~ao tect^ onica da Bacia de Santos e
Petroleum Engineers Annual Caspian Technical Confer- areas continentais adjacentes: Boletim de Geociências da
ence and Exhibition, Astana, Kazakhstan, November Petrobras, v. 3, no. 3, p. 159–173.
1–3, 2016, 15 p., doi:10.2118/182565-MS. Medlin, L. K., 2002, Why silica or better yet why not silica?
Jones, G. D., 2015, Flow-based effective properties of sub- Speculations as to why the diatoms utilise silica as
seismic karst: Society of Petroleum Engineers Middle their cell wall material: Diatom Research, v. 17, no. 2,
East Oil & Gas Show and Conference, Manama, Bahrain, p. 453–459, doi:10.1080/0269249X.2002.9705562.

Menezes, C. P., F. H. R. Bezerra, F. Balsamo, M. Mozafari, modeling: AAPG Bulletin, v. 105, no. 1, p. 135–155, doi:
M. M. Vieira, N. K. Srivastava, and D. L. de Castro, 10.1306/07072019196.
2019, Hydrothermal silicification along faults affecting Nolting, A., C. K. Zahm, C. Kerans, and M. A. Nikolinakou,
carbonate-sandstone units and its impact on reservoir 2018, Effect of carbonate platform morphology on syn-
quality, Potiguar Basin, Brazil: Marine and Petroleum depositional deformation: Insights from numerical
Geology, v. 110, p. 198–217, doi:j.marpetgeo.2019.07. modeling: Journal of Structural Geology, v. 115, p. 91–
018. 102, doi:10.1016/j.jsg.2018.07.003.
Mercedes-Martın, R., C. Ayora, J. Tritlla, and M. Sanchez- Oliveira, T. A. S., N. M. S. M. Cruz, J. M. N. Cruz, R. S.
Roman, 2019, The hydrochemical evolution of alkaline Cunha, and M. Matos, 2019, Faults, fractures and karst
volcanic lakes: a model to understand the South Atlantic zones characterization in a Pre-Salt reservoir using geo-
Pre-salt mineral assemblages: Earth-Science Reviews, v. metric attributes: 16th International Congress of the Bra-
198, 102938, doi:10.1016/j.earscirev.2019.102938. zilian Geophysical Society, Rio de Janeiro, Brazil, August
Minzoni, M., A. Cantelli, J. Thornton, and B. Wignall, 2020, 19–22, 2019, 5 p.
Seismic-scale geometries and sequence-stratigraphic Paterson, R. J., F. F. Whitaker, P. L. Smart, G. D. Jones, and
architecture of Early Cretaceous syn-post rift carbonate D. Oldham, 2008, Controls on early diagenetic over-
systems, Pre-Salt section, Brazil: in J. Hendry, P. Burgess, printing in icehouse carbonates: Insights from modeling
D. Hunt, X. Janson, and V. Zampetti, eds., Seismic char- hydrological zone residence times using CARB3D+:
acterization of carbonate platforms and reservoirs: Geo- Journal of Sedimentary Research, v. 78, no. 4, p. 258–
logical Society, London, Special Publications 2020, v. 281, doi:10.2110/jsr.2008.029.
509, p. 105–126, doi:10.1144/SP509-2019-78. Pe~
na, W. D. L., F. L. G. Pereira, P. A. L. P. Firme, D. Roehl,
Moczydlower, B., M. C. Salom~ao, C. C. Branco, R. K. Romeu, and L. C. Pereira, 2019, Casing integrity assessment of a
T. R. Homem, L. C. de Freitas, and H. A. Lima, 2012, wellbore crossing a karst cavity in a carbonate reservoir:
53rd US Rock Mechanics/Geomechanics Symposium,
Development of the Brazilian pre-Salt fields-When to
New York City, June 23–26, 2019, ARMA-2019-1567,
pay for information and when to pay for flexibility: Soci-
7 p.
ety of Petroleum Engineers Latin America and Caribbean
Pietzsch, R., D. M. Oliveira, L. R. Tedeschi, J. V. Q. Neto,
Petroleum Engineering Conference, Mexico City, Mex-
M. F. Figueiredo, J. C. Vazquez, and R. S. de Souza,
ico, April 16–18, 2012, SPE-152860-MS, 11 p., doi:10.
2018, Palaeohydrology of the Lower Cretaceous pre-salt
2118/152860-MS.
lacustrine system, from rift to post-rift phase, Santos
Moreira, J. L. P., C. V. Madeira, J. A. Gil, and M. A. P.
Basin, Brazil: Palaeogeography, Palaeoclimatology, Palae-
Machado, 2007, Bacia de Santos: Boletim de Geociências
oecology, v. 507, p. 60–80, doi:10.1016/j.palaeo.2018.
da Petrobras, v. 15, no. 2, p. 531–549.
06.043.
Muniz, M. C., and D. W. J. Bosence, 2015, Pre-salt microbia-
Playton, T. E., X. Janson, and C. Kerans, 2010, Carbonate
lites from the Campos Basin (offshore Brazil): Image log
slopes, in N. P. James and R. W. Dalrymple, eds., Facies
facies, facies model and cyclicity in lacustrine carbonates: models 4, GEOtext 6, St. John’s, Newfoundland, Geo-
Geological Society of London, Special Publications 2015, logical Association of Canada, p. 449–476.
v. 418, no. 1, p. 221–242, doi:10.1144/SP418.10. Pontes, T., F. Barbosa, and S. Cosendey, 2018, Managed pres-
Murray, R. C., 1990, Diagenetic silica stratification in a paleo- sure drilling in deepwater Brazil presalt severe-loss sce-
silcrete, North Texas: Journal of Sedimentary Research, nario: Society of Petroleum Engineers/International
v. 60, no. 5, p. 717–720. Association of Drilling Contractors Managed Pressure
Mylroie, J., and J. Mylroie, 2017, Bahamian flank margin Drilling and Underbalanced Operations Conference and
caves as hypogene caves, in A. Klimchouk, A. N. Palmer, Exhibition, New Orleans, Louisiana, April 17–18, 2018,
J. De Waele, A. Auler, and P. Audra, eds., Hypogene SPE-189998-MS, 6 p., doi:10.2118/189998-MS.
karst regions and caves of the world: New York, Springer, Ramsay, J. G., and M. I. Huber, 1987, The techniques of
p. 757–768, doi:10.1007/978-3-319-53348-3_51 modern structural geology: Volume 2 —Folds and frac-
Narr, W., and E. Flodin, 2012, Fractures in steep-rimmed car- tures: Cambridge, Massachusetts, Academic Press,
bonate platforms: Comparison of Tengiz reservoir, p. 309–700.
Kazakhstan, and outcrops in Canning Basin, NW Austra- Rezende, M. F., and M. C. Pope, 2015, Importance of deposi-
lia: Search and Discovery article 20161, accessed Febru- tional texture in pore characterization of subsalt
ary 16, 2019, http://www.searchanddiscovery.com/ microbialite carbonates, offshore Brazil.: Geological Soci-
documents/2012/20161narr/ndx_narr.pdf. ety, London, Special Publications 2015, v. 418, no. 1,
Nolting, A., and F. Fernandez-Iban ~ ez, 2021, Stress states of p. 193–207, doi:10.1144/SP418.2.
isolated carbonate platforms: Implications for develop- Ringrose, P., and M. Bentley, 2015, Reservoir model design—
ment and reactivation of natural fractures post burial: A practitioner’s guide: Dordrecht, the Netherlands,
Marine and Petroleum Geology, v. 128, 105039, 15 p., Springer, 249 p.
doi:10.1016/j.marpetgeo.2021.105039. Rosa, M. B., J. S. d. A. Cavalcante, T. M. Miyakawa, and
Nolting, A., P. J. Moore, J. Homburg, and F. Fernandez- L. C. S. de Freitas, 2018, The giant Lula Field: World’s
~ ez, 2021, The preservation of water-table caves at
Iban largest oil production in ultra-deep water under a fast-
depth: Observations from subsurface data and numerical track development: Offshore Technology Conference,

Houston, Texas, April 30–May 3, 2018, OTC-29043- Denver, Colorado, October 5–8, 2003, SPE-84459-MS,
MS, 13 p., doi:10.4043/29043-MS. 15 p., doi:10.2118/84459-MS.
Saka, G., I. Troth, C. Franca, F. Ramalho, S. Hoffman, J. Toll- Sweep, M.N., J. M. Bailey, and C. R. Stone, 2003, Closed
efsrud, V. Haugse, D. Torres, and L. Hansen, 2019, Fast hole circulation drilling: Case study of drilling a high-
model update and scenario based modelling for decision pressure fractured reservoir - Tengiz Field, Republic of
making in a fast track pre-salt development: First Euro- Kazakhstan: Society of Petroleum Engineers/Interna-
pean Association of Geoscientists and Engineers Work- tional Association of Drilling Contractors Drilling Con-
shop on Pre-Salt Reservoir: from Exploration to ference, Amsterdam, Netherlands, February 19–21,
Production, Rio de Janeiro, Brazil, December 5–6, 2019, 2003, SPE-79850-MS, 21 p., doi:10.2118/79850-MS.
4 p. Tang, T., K. Kisslinger, and C. Lee, 2014, Silicate deposition
Salom~ao, M. C., D. R. Marçon, M. B. Rosa, T. C. de Salles during decomposition of cyanobacteria may promote
Pessoa, and A. C. Capeleiro Pinto, 2015, Broad strategy export of picophytoplankton to the deep ocean: Nature
to face with complex reservoirs: Expressive results of pro- Communications, v. 5, 4143, 7 p., doi:10.1038/
duction in Pre-Salt area, Offshore Brasil: Offshore Tech- ncomms5143.
nology Conference, Houston, Texas, May 4–7, 2015, Tankersley, T., W. Narr, G. King, R. Camerlo, A. Zhumagu-
OTC-25712-MS, 13 p., doi:10.4043/25712-MS. lova, M. Skalinski, and Y. Pan, 2010, Reservoir modeling
Sartorato, A. C. L., S. N. Tonietto, and E. Pereira, 2020, Silici- to characterize dual porosity, Tengiz Field, Republic of
fication and dissolution features in the Brazilian pre-salt Kazakhstan: Society of Petroleum Engineers Caspian
Barra Velha formation: Impacts in the reservoir quality Carbonates Technology Conference, Atyrau, Kazakh-
and insights for 3D geological modeling: Rio Oil & Gas stan, November 8–10, 2010, SPE-139836-MS, 15 p.,
Expo and Conference, Rio de Janeiro, Brazil, December doi:10.2118/139836-MS.
2020, doi:10.48072/2525-7579.rog.2020.068. Thiede, D. S., and P. M. Vasconcelos, 2010, Parana flood
Saurbayev, I., J. Reedy, A. Bukharbayeva, C. Hatiboglu, and basalts: Rapid extrusion hypothesis confirmed by new
A. Massingill, 2018, Challenges and value of interference 40Ar/39Ar results: Geology, v. 38, no. 8, p. 747–750,
doi:10.1130/G30919.1.
testing during early production of Kashagan Field: Society
Tiercelin, J. J., A. S. Cohen, M. J. Soreghan, and K.-E. Lezzar,
of Petroleum Engineers Annual Caspian Technical Con-
1994, Pleistocene-modern deposits of the Lake Tangan-
ference and Exhibition, Astana, Kazakhstan, October
yika Rift Basin, East Africa: A modern analog for lacus-
31–November 2, 2018, SPE-192536-MS, 13 p., doi:10.
trine source rocks and reservoirs, in A. J. Lomando, B. C.
2118/192536-MS.
Schreiber, and P. M. Harris, eds, Lacustrine reservoir and
Simo, T., J. Sitgreaves, E. Smith, U. Bayram, and J. Stewart,
depositional systems: SEPM Core Workshop, v. 19, p.
2019, Styles of Pre-Salt carbonate platforms, Brazil San-
37–59, doi:10.2110/cor.94.01.0037.
tos Basin: (abs.): AAPG Annual Convention and Exhibi-
Tosca, N. J., and V. P. Wright, 2018, Diagenetic pathways
tion, San Antonio, Texas, May 19–22, 2019, accessed
linked to labile Mg-clays in lacustrine carbonate reser-
July 7, 2020, http://www.searchanddiscovery.com/ voirs: a model for the origin of secondary porosity in the
abstracts/html/2019/ace2019/abstracts/1645.html. Cretaceous pre-salt Barra Velha Formation, offshore Bra-
Sims, P. A., D. G. Mann, and L. K. Medlin, 2006, Evolution zil, in P. J. Armitage, A. R. Butcher, J. M. Churchill,
of the diatoms: insights from fossil, biological and molec- A. E. Csoma, C. Hollis, R. H. Lander, J. E. Omma, and
ular data: Phycologia, v. 45, no. 4, p. 361–402, doi:10. R. H. Worden, eds., Reservoir quality of clastic and car-
2216/05-22.1. bonate rocks: Analysis, modelling and prediction: Geo-
Solum, J. G., and B. A. H. Huisman, 2017, Toward the crea- logical Society, London, Special Publications 2018, v.
tion of models to predict static and dynamic fault-seal 435, p. 33–46, doi:10.1144/SP435.1.
potential in carbonates: Petroleum Geoscience, v. 23, no. Trice, R., 2005, Challenges and insights in optimising oil pro-
1, p. 70–91, doi:10.1144/petgeo2016-044. duction form Middle Eastern karst reservoirs: Society of
Stephens, N. P., B. Yergaliyeva, A. Zhazbayeva and B. Uap, Petroleum Engineers Middle East Oil & Gas Show and
2020, Detailed sector model to reconcile surveillance Conference, Kingdom of Bahrain, March 12–15, 2005,
results and high permeability, geologic features: Example SPE-93679-MS, 26 p., doi:10.2118/93679-MS.
from Kashagan Field, Kazakhstan, Society of Petroleum Turazzi Naveiro, J., and D. Haimson, 2015, Sapinho€ y field,
Engineers Annual Caspian Technical Conference, Vir- Santos Basin Pre-Salt: From conceptual design to project
tual, October 21–22, 2020, SPE-202532-MS, 14 p., doi: execution and results: Offshore Technology Conference
10.2118/202532-MS. Brasil, Rio de Janeiro, Brazil, October, 27–29, 2015,
Sun, H., T. Jiang, and X. Deng, 2019, Dynamic description OTC-26320-MS, 19 p., doi:10.4043/26320-MS.
technology of fractured vuggy carbonate gas reservoirs: Vaughan, R. L., 2015, Practical fracture modeling workflows:
Cambridge, Massachusetts, Gulf Professional Publishing, An example from a large carbonate field development:
351 p. Society of Petroleum Engineers Middle East Oil & Gas
Sun, S. Q., and R. Sloan, 2003, Quantification of uncertainty Show Conference, Manama, Bahrain, March 8–11, 2015,
in recovery efficiency predictions: Lessons learned from SPE-172652-MS, 9 p., doi:10.2118/172652-MS.
250 mature carbonate fields: Society of Petroleum Engi- Vice, G. S., J. E. Faulds, W. J. Ehni, and M. F. Coolbaugh,
neers Annual Technical Conference and Exhibition, 2007, Structural controls of a blind geothermal system in

the northern Pyramid Lake area, northwestern Nevada: Winter, W., R. Jahnert, and A. B. França, 2007, Bacia de
Transactions—Geothermal Resources Council, v. 31, p. Campos: Boletim de Geociências da Petrobras, v. 15,
133–137. p. 511–529.
Wan, Y., Y. Liu, and N. Wu, 2020, Numerical pressure tran- Wright, V. P., 2020, The mantle CO2 and the giant Aptian
sient analysis for unfilled-caved carbonate reservoirs chemogenic lacustrine carbonate factory of the South
based on Stokes-Darcy coupled theory: Journal of Petro- Atlantic: Some carbonates are made not born: Sedimen-
leum Science Engineering, v. 190, 107085, doi:10.1016/ tology, p. 1–27, doi:10.1111/sed.12835.
j.petrol.2020.107085. Wright, V. P., and A. J. Barnett, 2015, An abiotic model for
Warren, J. E., and P. J. Root, 1963, The behavior of naturally the development of textures in some South Atlantic early
fractured reservoirs: Society of Petroleum Engineers Fall Cretaceous lacustrine carbonates, in D. W. J. Bosence,
Meeting, Los Angeles, California, October 7–10, 1963, K. A. Gibbons, D. P. Le Heron, W. A. Morgan, T. Pritch-
SPE-426-PA, 11 p., doi:10.2118/426-PA. ard, and B. A. Vining, eds., Microbial carbonates in space
Warrlich, G. M. D., E. W. Adams, A. Ryba, T. Tam, K. K. and time: Implications for global exploration and produc-
Ting, and H.-K. Tang, 2019, What matters for flow and tion: Geological Society, London, Special Publications
recovery in carbonate gas reservoirs: Insights from the 2015, v. 418, p. 209–219, doi:10.1144/SP418.3.
mature Central Luconia Province, offshore Sarawak, Yaqoob, T., O. Bello, and J. Oppelt, 2017, Application of
Malaysia: AAPG Bulletin, v. 103, no. 3, p. 691–721, doi: controlled-influx mud cap drilling techniques in exploit-
10.1306/09051817219. ing the marine Pre-Salt carbonates in Santos Basin: Case
Wennberg, O. P., G. McQueen, P. H. Vieira de Luca, D. study: Society of Petroleum Engineers SPE/IATMI Asia
Hunt, A. S. Chandler, A. Waldum, and F. Lapponi, Pacific Oil & Gas Conference and Exhibition, Jakarta,
2019, Open fractures in pre-salt reservoirs in the Indonesia, October 17–19, 2017, SPE-186305-MS, 11
Campos Basin: Examples from silicified carbonates in p., doi:10.2118/186305-MS.
BM-C-33: European Association of Geoscientists and Yose, L. A., S. Brown, T. L. Davis, T. Eiben, G. S. Kompanik,
Engineers First EAGE Workshop on Pre-Salt Reservoir: and S. R. Maxwell, 2001, 3-D geologic model of a frac-
From Exploration to Production, Rio de Janeiro, Brazil, tured carbonate reservoir, Normal Wells Field, NWT,
December 5–6, 2019, 5 p., doi:10.3997/2214-4609. Canada: Bulletin of Canadian Petroleum Geology, v. 49,
201982010. no. 1, p. 86–116, doi:10.2113/49.1.86.

Fernandez-Ibanezetal.2022 AAPG XSPerm

Uploaded by

Copyright:

Available Formats

Fernandez-Ibanezetal.2022 AAPG XSPerm

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fernandez-Ibanezetal.2022 AAPG XSPerm

Uploaded by

Copyright:

Available Formats

Excess permeability in the AUTHORS

AAPG Bulletin, v. 106, no. 4 (April 2022), pp. 701–738 701

702 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 703

The lacustrine Aptian carbonates of the South Atlan-

704 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 705

706 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 707

708 Excess Permeability in the Brazil Pre-Salt

Non-BBFs—Non-BBFs (otherwise referred to as

FERNÁNDEZ-IBÁÑEZ ET AL. 709

710 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 711

712 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 713

imaged because their texture and high amplitude val- LCZs

714 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 715

716 Excess Permeability in the Brazil Pre-Salt

A “V-shape” response in the derivative is the

FERNÁNDEZ-IBÁÑEZ ET AL. 717

718 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 719

720 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 721

722 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 723

724 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 725

726 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 727

728 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 729

730 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 731

732 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 733

734 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 735

736 Excess Permeability in the Brazil Pre-Salt

FERNÁNDEZ-IBÁÑEZ ET AL. 737

738 Excess Permeability in the Brazil Pre-Salt

You might also like