You are on page 1of 9

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8

Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA

QUANTITATIVE CHARACTERIZATION OF CARBONATE PORE SYSTEMS ON


JONGGRANGAN FORMATION USING DIGITAL IMAGE ANALYSIS (DIA)
Ari Wibowo1*
Nana Higiana Hasanah2
1

Department of Geological Engineering, UPN Veteran Yogyakarta


Department of Geophysical Engineering, UPN Veteran Yogyakarta
JL. SWK 104 (Lingkar Utara) CondongCatur, Yogyakarta 55283
*corresponding author: ariwibowo_mr@yahoo.co.id

ABSTRACT
Jonggrangan Formation is characterized by major occurrence of reef limestone and small portion of
bioclastic limestone that were well exposed at Samigaluh Area, Kulon Progo Regency, Daerah
Istimewa Yogyakarta Province. These carbonates have a wide variety of pore systems that imprint
different petrophysical properties, which are more difficult to predict than in siliciclastics. Digital
Image Analysis (DIA) can be applied to response the carbonates complexity by calculating pore value
and also characterizing the pore shape in digital form. The method is based on images from rock thin
sections taken under an optical microscope (OM) and also core analysis results to be compared.
Sixteen rock thin section samples have been analyzed with several observation representatively in
order to mitigate the effect of area selection problems. Blue-dyed liquid also has been added to those
thin section samples, so the pore could be identified clearly. Based on further examination, both
primary and secondary pore systems are well developed. There are six pore shapes are identified
within different pore values. Crossplots of pore values indicate the pore variety depends on several
rock parameters. In summary, understanding characterization of carbonate pore system on
Jonggrangan Formation by using DIA method is fast and accurate, useful to encourage and enrich
carbonate petrophysical analysis concept.
Keywords: Jonggrangan Formation, Carbonate Pore Systems, DIA.

I.

INTRODUCTION
The carbonates are influenced greatly by
sea level fluctuation which recorded on rock
textures. Pore system in carbonate is much
more complex than siliciclastics, as a result
of overhelming biological origin of
carbonate sediments that reflected by the
occurance of porosity within grains, growth
framework porosity within reefs, and the
common development of secondary
porosity due to pervasive diagenetic
processes (Choquette and Pray, 1970). In
addition to traditional descriptive and
qualitative porosity evaluations, there exists
a need for quantitative methods that
characterize the various aspects of pore
space and enable a quantitative assessment
of the distribution of porosity and other
physical properties (Anselmetti et al., 1998).
This Paper discussess about evaluating pore
characteristic quantitatively on carbonates
using digital image analysis (DIA) which

well-established method of quantifying


pore space from images of thin sections.
Numerous researches were conducted over
time to classify pore structure. Anselmeti et
al. (1998) quantified the pore shape in
carbonates from digital images obtained by
optical microscope and Scanning electron
microscope. Castro (2013) determined
Gamma parameter (ratio between pore
perimeter and area) and correlated it with
permeability measurements. In this study,
authors try to enlight about relationship
several parameters toward cabonate facies
and pore type based on thin section and
supported by core analysis.
Jonggrangan Formation is one of many
carbonate facies that well exposed in
Indonesia, particularly in Java island, has
been elected to be analyzed. Carbonate reef
on Jonggrangan Formation may lead us to
understand more about the development of
carbonate pore system.

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA

II.

REGIONAL GEOLOGY

The study area belongs to Kulonprogo Subbasin, the eastern part of South Central Java
Basin. This basin is located on the southern
portion of modern Java Volcanic Range and
could be identified as well as Basin lies on
Magmatic Arc. Volcanic activity which is
occured was the result of subduction
beneath the Hindia Ocean. Thus, this basin
is classified as fore arc-basin since it's
location is in front of Java Volcanic
mountain range. The oldest Formation in
this area is Middle Eosen Nanggulan
Fm. which is well exposed on Nanggulan
Village. This Formation is typified by the
occurance of intercalation of sandstone,
shale, claystone and lignit which is
deposited
on
paralic
depositional
environment. Lies conformably above
Nanggulan Formation is Kaligesing Fm.
consisted of volcanic breccia and intrusion.
Since no fossil could be attained from this
formation, it is assumed as Oligosen Age
formation. Another formation which is
deposited on the approximate same time
with this formation is the polymict
breccia Dukuh Fm. Dukuh Formation
consists of polymict calcareous breccia with
interbedded sandstone and claystone,
interpreted as shallow marine sediments.
Dukuh Formation and Kaligesing Formation
has interfingering stratigraphic relation. Lies
unconformably above Kaligesing Formation
are two distinctive carbonate formation,
Sentolo Fm. and Jonggrangan Fm. Sentolo
Fm, typified as clastic carbonate formation,
consist of intercalation of calcarenite,
calcilutite, and marl. The abundant fossil
content in this formation allowed a firm
depositional environment interpretation of
this formation, as well as the Age of this
formation. Sentolo is Early Miocene
carbonate formation deposited on Neriticupper Bathyal environment. On the other
hand Jonggrangan Fm., the focus of this
study is characterized by the occurence of
reefal limestone which depict various
carbonate facies. Jonggrangan Fm. is

interpreted as a patch reef deposited on


shelf area. Sentolo and Jonggrangan have
interfingering relation. (Pringgoprawiro and
Riyanto, 1968).

III.

SAMPLE
METHOD

AND

RESEARCH

Sixteen rock thin section samples have been


analyzed with
several
observations
representatively in order to mitigate the
effect of area selection problems. Blue-dyed
liquid also has been added to those thin
section samples, so the pore could be
identified clearly. The DIA (digital image
analysis) method is based on images from
rock thin sections taken under an optical
microscope (OM) and also core analysis
results to be compared. The method
comprises of three steps: image acquisition,
segmentation, and calculation of pore
parameters. Image acquisition takes
approximately 10 15 minutes per sample,
with additional time needed for image
analyses and statistical evaluations. The
segmentation process is the separation of a
specific feature from the background and it
was performed on colored digital images
acquired in standard RGB (red-green-blue)
which were converted to a binary 8-bit BW
(black and white) format to identify
respectively,
the
pore
and
the
matrix/cement. In this study, the object
feature of segmentation was the rock pore
that was quantified and its geometry
parameterized. In addition, the image
analyses program authors used (Image J) is
freely available as public domain software.
Then calculation process uses mathematical
programs in order to create crossplots for
several parameters.

IV.

DATA AND ANALYSIS

Several parameters defined in the DIA to


quantify pore space. Two DIA parameters,
Dominant pore size (DOM) and Perimeter
over area (PoA), are proved to best describe
several aspects of pore system, namely

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA
facies, pore type,
permeability (K).

porosity

()

and

Theoretically, dominant pore size (DOM) is


determined as the upper boundary of pore
sizes of which 50% of the porosity on a thin
section composed. Half of the pore space of
an area is composed of pores as big as or
smaller than the DOM. Whereas, Perimeter
over area (PoA) is the ratio of the sum of
total perimeter of the pores and total pore
area identified in the thin section. It
describes how complex is the porous
system independent of total porosity.
Based on further examination from outcrop,
core (polished slab) and thin section data,
eight lithofacies have been determined
(Figure 1 &2). Six pore types also have been
observed which are Vuggy, Moldic,
Intraparticle, Interparticle, Channel, and
Growth-Framework.
The
quantitative
results show that visible porosity () varied
between 1.87 34.7 % and permeability (K)
varied between 44.68 151.68 md (Table 1).
Based on crossplot both visible and
measured porosity vs perm, there is no
relationship between these parameters.
DOM varied between 157 - 1100 m2.
Crossplot of porosity vs permeability with
according to DOM values in figure 3 shows
that for a given porosity, the high DOM
values indicate the presence of large pores
in the thin section as observed in Figure 4.A,
where a grainstone facies with vuggy pore
type displays DOM 1100 m2. At the same
porosity, the smaller value DOM, smaller
pore size is and the larger number, larger
the pore size is.
The PoA varied between 75.5 - 166 mm-1.
Figure 5 shows crossplot of porosity vs
permeability with according to PoA values.
Figure 4.B shows the high PoA value, 139.5
mm-1, for a thin section of a framestone
facies with growth-framework porosity. At
the same porosity, the smaller PoA, the
simpler pore geometry is and the larger
number, the more intricate pore geometry.

V.

DISCUSSION

Linkage of measured pore with visible pore


Both visible porosity and measured porosity
show the different values on several
samples (Figure 6). It is result of the effect
of area selection for 2 cm x 2 cm thin
section samples, does not represent the
whole porosity characteristic of carbonate
rock. The difference value is a possibility,
because on carbonate facies, the secondary
porosity (eg: vuggy, moldic, and channel)
are subsequently well developed on
Jonggrangan Formation. Simply put, scale of
observation and area selection take an
important play.
Linkage of facies with porosity and
permeability
Facies is known as one factor which control
petrophysical values such as porosity and
permeability. However the extent of facies
control on porosity is constraint by several
aspects. In order to prove the hypothese,
the cross-plot of lithofacies toward porosity
and permeability has been done (Figure 7).
It shows that the lithofacies influence on
porosity and permeability is not linear.
Simply put, there are several samples of the
same lithofacies type which possessed
different petrophysical values. For example,
grainstone (red symbol) present the lowest
porosity and lowest permeability but the
other grainstone present higher result with
different value significantly. These condition,
by far are controlled by diagenetic
processes. A carbonate rock has its own
sedimentation history both syn depositional
or post depositional. The petrophysical
values can be better when dissolution
happen but it can be worst while
recrystallization occur, such as porosity, will
be enhance by dissolution process, while
recrystalisation does the opposite.
Linkage of pore type with DOM and PoA
In general, grainstones with large secondary
pores (vuggy) had high DOM and low PoA.
Otherwise, fine grained facies which are
packstones and wackestones supported by
microporosity had high PoA and low DOM.

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA
Figure 8 shows a summary of these result,
crossplot of DOM vs PoA plot for all thin
sections analyzed and a few images of the
sections indicating the respective PoA and
DOM parameters.

VI.

CONCLUSIONS

This study concludes that:

Carbonate pore system is more complex


than siliciclastics.
Digital image analysis method is useful
to encourage and enrich carbonate
petrophysical analysis concept.
The geometric DIA parameters that
varied better able to quantify the

VII.

differences in pore type of the thin


sections study are dominant size (DOM)
and perimeter of area (PoA) which
could be used to identify high and low
porosity.
Integration of all available data while
observing carbonate pore system is a
must because its vast secondary
porosity development.

ACKNOWLEDGEMENT

Authors would like to acknowledge to


Geological Engineering Department of UPN
Veteran Yogyakarta for the support and
also to Sedimentology Lab. for facilitating
the laboratory analysis.

REFERENCES
Anselmetti, F.S., Luthi, S., Eberli, G.P., 1998. Quantitative Characterization of Carbonate Pore Systems
by Digital Image Analysis. AAPG Bulletin, V.82, No. 10 p. 1-15.
Castro, D.D., Rocha, P.L.F., 2013. Quantitative Parameters of Pore Types in Carbonate Rocks. Revista
Brasileira de Geofisica.
Choquette, P.W. dan Pray, L.C. 1970, Geologic Nomenclature and Classification of Porosity in
Sedimentary Carbonates. The American Association of Petroleum Geologist Bulletin, Vol. 54, No. 2,
Ferbruary, 1970. p.222 224.
Ditya, A., Eberli, G.P., McNeil, D., Klaus, J., 2015. Segregating Porosity-Permeability Transform Using
Velocity and Pore Structure In Carbonate Rock. Proceedings Indonesian Petroleum Association 39th
Annual Convention & Exhibition. Jakarta.
Pringgoprawiro, H and Riyanto, 1968. Geologi Regional KulonProgo. ITB, Bandung.
Ramadhan, G.C., 2013, Organism Variety Effect on Carbonate Rock Porosity of Jonggrangan
Formation: Alternative Approach To Predict Porosity Complex. MajalahGeologi Indonesia, Vol. 28, 1
April 2013. p.29-40.
Widijana, A.R.T., Latief, F.D.E., Anissofira, A., Fatkhan, Handoyo, 2015. Modelling of The Sandstone
Ngrayong Formation Using Digital Image Analysis. Proceedings Indonesian Petroleum Association
39th Annual Convention & Exhibition. Jakarta.
Zhao, L., Nasser, M., Han, D., 2013. Quantitative Geophysical Pore-Type Characterization and Its
Geological Implication in Carbonate Reservoir. European Association of Geoscientists & Engineers,
Geophysical Prospecting p. 1-15.

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA

Table 1. The summary of analysis result


Digital Image Analysis

Sample
Code

Lithofacies

3D
Visible
(%)
(%)

Ari 1

Wackestone

5,7

15,63

Vuggy

202

Ari 2

Framestone

5,07

3,44

Moldic

704

Ari 3

Packstone

3,59

8,96

Intraparticle

110

166

84

Ari 4

Framestone

3,18

22,59

Growth-Framework

366

137,5

69

Ari 5

Bafflestone

2,58

11,25

Growth-Framework

575

116.5

52,79

Ari 6

Floatstone

2,47

1,87

Intraparticle

476

136

76,55

Ari 7

Floatstone

5,1

6,28

Vuggy

353

138,75

46,29

Ari 8

Grainstone

3,16

2,50

Intraparticle

157

175

44,68

Ari 9

Marl

6,43

22,38

Interparticle

580

106,5

Ari 10

Packstone

6,43

13,91

Moldic

445

114,75

83,92

Ari 11

Grainstone

8,87

34,70

Vuggy

1110

105

75,43

Ari 12

Packstone

6,6

10,74

Channel

345

139,5

118,26

Ari 13

Rudstone

7,16

8,25

Intraparticle

262,5

97, 5

93,22

Ari 14

Floatstone

7,66

8,49

Intraparticle

272,5

121,5

120,56

Ari 15

Bafflestone

2,77

13,12

Interparticle

840

92,5

100,48

Ari 16

Packstone

4,94

4,87

Intraparticle

483

145,75

Main
Pore Type

DOM
(m2)

PoA
(mm-1)

Perm
(md)

144

49,98

75,5

: porosity; DOM: dominant pore size; PoA: perimeter over area; Perm: permeability.

151,68

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA

Figure 1. (From left to right) Integration of outcrop, core (polished slab) and thin section data.

Figure 2. Core-plug sample and tools for measuring porosity analysis to support DIA result.

Figure 3. Crossplot of porosity vs permeability with according to DOM values. The high DOM values
indicate the presence of large pore size, and lower DOM values indicate the opposite.

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA
A

Lithofacies
Main Pore Type
Avr. Porosity

= Grainstone
= Vuggy
= 34.7 %

DOM
PoA
K

= 1101m2
= 105 mm-1
= 75.43 md

= Framestone
= Growth-Framework
= 22.59 %

DOM
PoA
K

= 366 m2
= 137.5 mm-1
= 69 md

Lithofacies
Main Pore Type
Avr. Porosity
C

Lithofacies
= Packstone
DOM = 345 m2
Main Pore Type = Channel
PoA
= 139.5 mm-1
Avr. Porosity
= 10.74 %
K
=118.26 md
Figure 4. Photomicrographs show three samples section of pore system (left) in which porosity is
shown as blue, and digital image analysist result in BW-image (right) which porosity is shown as black
and matrix is white. (A) Grainstone facies, (b) Framestone facies, and (c) Packstone facies.

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA

Figure 5. Crossplot of porosity vs permeability with according to PoA values.

Figure 6. Comparison of measured porosity values obtained from core-plug analysis and porosity
values obtained from digital image analysis for various pore type . There is no linear caused by
nonrepresentative observation area of view under optical microscope.

PROCEEDING, SEMINAR NASIONAL KEBUMIAN KE-8


Academia-Industry Linkage
15-16 OKTOBER 2015; GRHA SABHA PRAMANA

Figure 7. Crossplots of measured pore vs permeability (left) and visible pore vs permeability (right)
for various facies. Both crossplots ilustrates that the extent of facies control on carbonate pore
system is constraint by several aspect.

Figure 8. Crossplot of dominant size (DOM) vs perimeter of area (PoA) plot for all thin sections
analyzed and a few images of the sections indicating the respective DOM and PoA parameters.