Paleoenvironments and high-frequency cyclicity from Cenozoic South-East Asian shallow-water carbonates: a case study from the Oligo-Miocene buildups of Malampaya (Offshore Palawan, Philippines) F. Fournier * , L. Montaggioni, J. Borgomano Centre de Se ´dimentologie–Pale ´ontologie, UMR-CNRS 6019, Dynamique des re ´cifs et des plates-formes carbonate ´es, case 67, Universite ´ de Provence, 3, place Victor Hugo, F-13331 Marseille cedex 03, France Received 20 June 2003; received in revised form 21 November 2003; accepted 28 November 2003 Abstract The combination of core and cuttings analyses and well-log data from the Late Oligocene – Early Miocene carbonate buildup in the Malampaya gas field (Offshore Palawan, Philippines) allowed the recognition of shelf depositional paleoenvironments and the definition of high-frequency, metre-scale, subtidal cycles, usually bounded by exposure surfaces. This is amongst the first documentation of short-term platform-top evolution of Oligo-Miocene carbonates in response to high-frequency relative sea-level change described in South-East Asia. During the Late Oligocene, the Malampaya carbonate buildup is interpreted as an isolated platform rimmed by a barrier-reef system that has developed vertically according to the ‘keep-up’ mode, with no significant evidence for retardation in reef development during rises in relative sea-level. During the Early Miocene, the isolated platform was affected by more open conditions in its inner part due to the inability of the reef to catch up with relative sea-level rises. Strontium isotope-derived estimations of cycle durations, despite a relatively high degree of uncertainty in their calculation, are consistent with 4th to 5th-order cyclicity (10 – 1000 ky). q 2003 Elsevier Ltd. All rights reserved. Keywords: Cycles; Carbonate platforms; Paleoenvironment; Diagenesis; Oligocene; Miocene; South-East Asia 1. Introduction The development of Oligo-Miocene carbonate systems from South-East Asia has been mostly studied from seismic or outcrop data, with special reference to 3rd-order depositional sequences (e.g. Cucci & Clark, 1993; Epting, 1989; Gro ¨tsch & Mercadier, 1999; Kusumastuti, Van Rensbergen, & Warren, 2002; Saller & Vijaya, 2002; Sun & Esteban, 1994; Wilson, Bosence, & Limbong, 2000; Wilson, Chambers, Evans, Moss, & Nas, 1999; Wilson & Evans, 2002). In contrast, studies on short-term evolution of shallow-water carbonates in response to 4th or 5th-order relative sea-level fluctuations (sensu Goldhammer, Dunn, & Hardie, 1990) are rare, and high-frequency sequences have only been hinted at in a very few cases (Gibbson-Robinson and Soedirdja, 1986; Kusumastuti et al., 2002; Park, Matter, & Tonkin, 1995). In the Early Miocene Batu Raja limestones from the Sunda basin (South-East Sumatra, Indonesia), Park et al. (1995) defined wireline log cycles, whose frequency could be compared with 5th-order cyclicity. In addition, these authors evoked the existence of repeated exposure events related to high-frequency cyclicity and emphasized their significant role on the enhancement of reservoir quality. Repeated hiatuses associated perhaps with subaerial exposures are hinted at in other South-East Asian carbonates of Cenozoic age; some of them could be related to high- frequency cyclicity, e.g. the Early Miocene carbonate platform of the Madura Strait, east Java (Kusumastuti et al., 2002), the Miocene Kais platform, Irian Jaya (Gibbson-Robinson & Soedirdja, 1986) and the Early Miocene Gomantong Limestone, eastern Borneo (Noad, 2001). In most cases, however, the regional development of carbonate reservoirs is thought to be mainly related to 3rd- order sea-level falls and consequent meteoric diagenesis (Sun & Esteban, 1994). Often the effect of 4th to 5th-order cycles on facies and reservoir property distribution has not been considered in detail or perhaps underestimated. Indeed, the very few studies on Oligo-Miocene outcrops 0264-8172/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpetgeo.2003.11.012 Marine and Petroleum Geology 21 (2004) 1–21 www.elsevier.com/locate/marpetgeo * Corresponding author. Tel.: þ 33-4911-06178; fax: þ 33-4911-08523. E-mail address: [email protected] (F. Fournier).
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Paleoenvironments and high-frequency cyclicity from Cenozoic
South-East Asian shallow-water carbonates: a case study from the
Oligo-Miocene buildups of Malampaya (Offshore Palawan, Philippines)
F. Fournier*, L. Montaggioni, J. Borgomano
Centre de Sedimentologie–Paleontologie, UMR-CNRS 6019, Dynamique des recifs et des plates-formes carbonatees, case 67, Universite de Provence, 3,
place Victor Hugo, F-13331 Marseille cedex 03, France
Received 20 June 2003; received in revised form 21 November 2003; accepted 28 November 2003
Abstract
The combination of core and cuttings analyses and well-log data from the Late Oligocene–Early Miocene carbonate buildup in the
Malampaya gas field (Offshore Palawan, Philippines) allowed the recognition of shelf depositional paleoenvironments and the definition of
high-frequency, metre-scale, subtidal cycles, usually bounded by exposure surfaces. This is amongst the first documentation of short-term
platform-top evolution of Oligo-Miocene carbonates in response to high-frequency relative sea-level change described in South-East Asia.
During the Late Oligocene, the Malampaya carbonate buildup is interpreted as an isolated platform rimmed by a barrier-reef system that has
developed vertically according to the ‘keep-up’ mode, with no significant evidence for retardation in reef development during rises in relative
sea-level. During the Early Miocene, the isolated platform was affected by more open conditions in its inner part due to the inability of the
reef to catch up with relative sea-level rises. Strontium isotope-derived estimations of cycle durations, despite a relatively high degree of
uncertainty in their calculation, are consistent with 4th to 5th-order cyclicity (10–1000 ky).
q 2003 Elsevier Ltd. All rights reserved.
Keywords: Cycles; Carbonate platforms; Paleoenvironment; Diagenesis; Oligocene; Miocene; South-East Asia
1. Introduction
The development of Oligo-Miocene carbonate systems
from South-East Asia has been mostly studied from seismic or
outcrop data, with special reference to 3rd-order depositional
water-energy, nature of carbonate producers, periodicity
and amplitude of relative sea-level oscillations); and (4) to
propose a model of short-term, sedimentologic and
diagenetic evolution of the carbonate buildup in response
to relative high-frequency sea-level fluctuations.
2. Location and geological setting of the Malampaya
carbonate buildup
The Malampaya–Camago oil and gas accumulations are
situated within the block SC 38 (Fig. 1), offshore northwest
Palawan Island (Philippines) below 850–2000 m of water
depth (Grotsch & Mercadier, 1999; Neuhaus et al., 2003). In
this area, the presence and overall development of carbonate
buildups are mainly controlled by earlier block faulting. The
Malampaya carbonate buildup is located in the North
Palawan block, along the outer trend of carbonate prospects
that formed along a SW–NE trending extensional fault
zone. The Nido Limestone reservoir has a maximum
thickness of 700 m in the area.
The Malampaya carbonate buildup developed on the
crest of a tilted block (Fig. 2), during the Late Eocene rifting
phase of the South-China Sea (Hall, 2002; Holloway, 1982).
The break-up event related to this rifting phase was dated by
mid-Oligocene magnetic anomaly 11 (Briais, Patriat, &
Tapponnier, 1993). The spreading of the South-China Sea
led to the southward drifting of the Calamian–North
Palawan–North Borneo micro-continent throughout the
Late Oligocene and Early Miocene. During the Late Early
Miocene, collision took place between this micro-continent
and the accretion wedge of the Paleogene subduction zone
of North Cagayan (Schluter, Hinz, & Block, 1996),
promoting the obduction of the collision belt on the North
Palawan block and ceasing seafloor spreading (Briais et al.,
1993). Many carbonate buildups in the area drowned due to
the downwarping of the north-western part of the block and
the important clastic supply from the uplifted Palawan
island (Fulthorpe & Schlanger, 1989). The carbonate
buildups of the Block SC 38 are sealed by the Early to
Middle Miocene basinal Pagasa shale.
The Malampaya buildup was previously studied by
Grotsch and Mercadier (1999) on the basis of three-
dimensional (3D) seismic data and relatively sparse core
and side-wall samples from four wells (MA-1 to MA-4).
These authors dated the carbonates as Late Eocene to the
Early Miocene in age using strontium isotope analyses of
bulk-rock samples. The step-like shape of their Sr-isotope
curve in well MA-3, located in the buildup flank, resulted
from important time-rock gaps (0.5–2 my) in the Nido
sedimentary record. The age of the Nido Limestone is also
constrained by the nannofossil and planktonic dating of the
overlying Pagasa shales (Grotsch & Mercadier, 1999). The
nannofossils in the shales provided a NN5 (Langhian) age in
well MA-1 and a NN3 (Late Burdigalian) age for MA-2,
whereas the planktonic foraminiferal assemblages indicate
Fig. 1. Depth (in metres subsea) of the top Nido Limestone and well
locations, in the Malampaya and Camago gas field (after Grotsch &
Mercadier, 1999) within Block SC 38, offshore western Palawan,
Philippines.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–212
a Burdigalian age in MA-3. During the Early Oligocene, the
Malampaya carbonate system prograded southward with
buildup aggradation and subsequent backstepping during
the Late Oligocene and Early Miocene (Grotsch &
Mercadier, 1999). The buildup finally drowned during the
Late Early Miocene.
3. Material and methods
Since the initial study by Grotsch and Mercadier
(1999), the Malampaya carbonate buildup has been
penetrated by six additional wells (named MA-5 to MA-
10). It is covered by a 3D seismic survey that has been
porosity-inverted (Neuhaus et al., 2003). The average well
spacing is about 500 m. The present study is based on
new rock material and uses a different analytical approach
than that of Grotsch and Mercadier work. Our work uses a
compilation of various subsurface data from all of the
available wells and seismic lines combined with petro-
graphical, micropaleontological and geochemical analyses
of core and side-wall material from the wells MA-5 and
MA-7, drilled in 2000. Well-log data and cutting samples
were also taken into account to complete the information
in intervals where no core or side-wall samples were
available. The total length of cores in MA-5 is 72 m
distributed into two main intervals (Fig. 3) and 15 m in
MA-7. Thin-sections from core samples were prepared
with an average spacing of 0.50 m.
Detailed microscopic analysis of thin-sections pro-
vided the backbone of this study. Thin-sections were
impregnated with a blue-dyed resin and half-stained with
an alizarine-red S þ potassium ferrocyanide solution for
identification of carbonate minerals. All thin-sections
were point-counted on the basis of 300 points.
The abundance of large coral fragments was estimated
from direct macroscopic examination of the core
sections.
Additional analyses included measurements of carbon,
oxygen and strontium isotope ratios and uranium
concentrations. All of the analyses were made on
selected whole-rock samples. Carbon and oxygen isotope
ratios were measured at the University of Erlangen as
follows: carbonate powders were reacted with 100%
phosphoric acid (density .1.9) at 75 8C in an on-line
carbonate preparation line (Carbo-Kiel-single sample acid
bath) connected to a Finnigan Mat 252 mass spec-
trometer. Results are reported in permil relative to V-
PDB. Reproducibility was checked by replicate analysis
of laboratory standards and is better than ^0.02‰ for
d13C and ^0.03‰ for d18O. Sr-isotope measurements
were performed at the Vrije Universiteit in Amsterdam.
The samples were dissolved in 5N acetic acid and the Sr
was separated using a ion exchange column. The Sr-
isotope ratio was measured by a Finnigan MAT 261
mass spectrometer. The reproducibility of measurements
is very good with two sigma errors between ^6 £ 1026
and ^10 £ 1026.
Petrographical analysis has allowed the recognition of
distinct microfacies that have been interpreted in terms of
depositional environments by reference to modern and
ancient analogues. The identification of diagenetic features
from thin-sections combined with whole-rock d13C and
d18O measurements allowed the reconstruction of the
diagenetic history of the series. The interpretation of the
data in terms of sequence stratigraphy was based finally on
the vertical succession of depositional environments and
diagenetic sequences in cores. Large benthic foraminiferal
biostratigraphy and strontium isotope stratigraphy
were used to constrain the stratigraphic framework of the
Nido Limestone and to help quantify the duration of
sedimentary cycles.
Fig. 2. (a) Regional seismic cross-section showing the structural setting of the Malampaya carbonate buildup, at the crest of a tilted block; (b) cross-section
from the 3D seismic reflectivity data showing the overall morphology of the carbonate buildup, with location of wells MA-1 and MA-5.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 3
4. Results
4.1. New stratigraphical data on the Nido Limestone
The biostratigraphical study of newly collected rock
material from wells MA-5 and MA-7 as well as strontium
isotope measurements on MA-1 allowed the stratigraphic
framework of the Malampaya platform to be clarified.
The Nido Limestone interval was dated on the basis of
benthic foraminiferal stratigraphy and strontium isotope
analyses of bulk-rock samples. The benthic foraminiferal
stratigraphy was based on the East Indian Letter Classifi-
1999). Biostratigraphic and strontium-derived ages are
presented in Table 1.
The shifts in Sr-isotope ratios are related to major
sequence boundaries delimiting 40–130 m thick sedimen-
tary units: MC1 and MC2 for the Late Oligocene, MM1,
MM2, MM3 and MM4 for the Early Miocene (Fig. 3). Well
correlations are constrained by biostratigraphy, seismic and
well-log data. At the base of the Nido Limestone interval, a
progradational unit of Rupelian age was defined; the top of
this unit corresponds to the lowest exposure surface
observed in the Nido Limestone.
In well MA-1, strontium ages derived from Oslick,
Miller, Feigenson, and Wright (1994) regressions are in
accordance with the biostratigraphy-based ages; however,
the age of the MC2 unit remains uncertain since
biostratigraphy gives a Lower Te age (Chattian) and Sr-
isotope dating yields ages ranging from the latest Chattian to
the earliest Aquitanian. At Eniwetak Atoll, Ludwig, Halley,
Simmons, and Peterman (1988) showed that meteoric
diagenetic alteration did not significantly alter the original
Miocene to the Pleistocene strontium isotope ratios, because
calcite precipitated in soil and fresh-water within the
interval was derived from adjacent depositional carbonates.
For the Nido Limestone, the very narrow range of variation
in strontium-isotope ratios within a given unit and the lack
of correlation with the d13C signal (Fig. 3) similarly suggest
that there was no significant alteration of the original
strontium-isotope signal during meteoric diagenesis. Later
diagenetic alteration of the isotopic signal affecting the
series below the Intra-Nido Marker unconformity (top MC2
unit) is believed to be responsible for the relative dispersion
of the measurements in this interval.
4.2. Facies analysis and paleoenvironmental interpretations
The modal analysis of bioclastic components, the
composition of foraminiferal assemblages and the sedimen-
tological features observed in cores and thin-sections
allowed four Late Oligocene facies and four Early Miocene
Fig. 3. Correlation panel of wells MA-7, MA-1 and MA-5, showing the major sedimentary units of the Malampaya buildup, the location of rock data used and
the main stable isotope results.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–214
facies to be recognized. The bioclastic composition is
expressed as a percentage of the whole non-scleractinian
biota. These facies are interpreted in terms of depositional
and Sphaerogypsina; heterosteginids, soritids and arenac-
eous foraminifera are also present, generally in the form of
broken specimens.
Paleoenvironmental interpretation. The absence of mud
is regarded as indicative of moderate to high bottom current
Table 1
Stratigraphic framework of the Nido Limestone interval based on benthic foraminiferal stratigraphy (Letter Stage Classification) from MA-1 and MA-5 and Sr-
isotope ages obtained in well MA-1
Unit Letter stage age Sr-derived age (my)
Min.a Max.a Age errora Ageb
MM4 Lower Tf1 (Burdigalian to Early Langhian) 19.90 19.97 0.80 Burdigalian
MM3 Upper Te–Lower Tf1 (Aquitanian to Early Langhian) 21.28 21.71 0.77 Aquitanian
MM2 Upper Te–Lower Tf1 (Aquitanian to Early Langhian) 22.06 22.47 0.78 Aquitanian
MM1 Upper Te–Lower Tf1 at top Undifferentiated Te at base (Chattian to Early
Langhian)
22.48 23.00 0.76 Aquitanian
MC2 Undifferentiated Te at top Lower Te (Chattian) at base 23.37 24.61 1.24 Latest Chattian to earliest Aquitanian
MC1 Lower Te (Chattian) 26.24 – 1.20 Chattian
MR Tc–Td (Rupelian) – – – –
Age errors include measurement errors and errors associated with Oslick et al. (1994) regression, considering a 95% confidence interval.a Using Oslick et al. (1994) regressions; min. and max. represent, respectively, the minimal and maximal age calculated for the given interval.b After Berggren, Kent, Swisher, & Aubry (1995).
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 5
Table 2
A summary of the main facies from the Late Oligocene and Early Miocene of the Malampaya shelf top: sedimentologic features, bioclastic components and paleoenvironmental interpretation
F.
Fo
urn
ieret
al.
/M
arin
ea
nd
Petro
leum
Geo
log
y2
1(2
00
4)
1–
21
6
conditions, an interpretation that is further supported by the
abundance of robust benthic foraminifera such as alveoli-
nids (Borelis), rotaliids, large miliolids (Austrotrillina),
Sphaerogypsina, amphisteginids. Morever, the Austrotril-
lina–Borelis association has been described by Chaproniere
(1975) in the Australian Oligocene as typical of high-energy
subseagrass environments in sheltered metahaline con-
ditions. Epiphytic forms as Heterostegina borneensis and
the soritids could reflect derivation from adjacent seagrass
communities. The abundant fine-grained coral debris
probably come from a nearby reef rim and/or from inner-
shelf patch reef frameworks. The biotic components and
the grainstone texture of the C2 facies therefore could be
interpreted as reflecting deposition in high-energy sand
shoals migrating across a backreef zone colonized by
seagrass meadows and possibly patch reefs. However, as
suggested by Pomar (2001) grainstone bodies, composed
largely of shallow water-derived components, can also be
deposited in deeper settings by currents. Nevertheless, the
hypothesis of a shallow sand shoal environment is
preferentially retained, because of the complete lack of
(Fig. 5h) and micritic mottles (Fig. 5e), and 2–3 m thick
Fig. 4. Late Oligocene and Early Miocene microfacies from the Malampaya shelf (well MA-5). Scale bar ¼ 1 mm. (a) Coralline algal crusts (facies C1a). (b)
litization and microfracturing during burial. Stages 3 and 4
could have been partly coeval.
4.3.2.2. B. Exclusive meteoric phreatic diagenetic sequence.
(1) Micritization and/or rim cement precipitation (mostly in
high-energy setting) during early marine diagenetic stage.
(2) Moderate leaching and partial to complete occlusion of
pores by drusy calcite during successive meteoric phreatic
phases. (3) Matrix neomorphosis, stylolitization and
Fig. 5. Meteoric diagenetic features from the Late Oligocene and Early Miocene limestones of the Malampaya shelf. (a) Foraminifer-rich grainstone showing
intense leaching of miliolid foraminifera (mil) and drusy calcite cementation (dc) occluding primary intergranular pores (drusy calcite cements are more
developed in the intergranular space than in the biomoulds); a leached mollusk fragment (mol) is recognizable by its residual micrite envelope; (Late
Oligocene, MA-5), scale bar ¼ 1 mm. (b) View of a cored interval (3202.45–3202.13 m) from well MA-7, Early Miocene. (c) Details of a caliche profile; note
the dark colour of the calcrete crust. (d) Coated pisoids (pis), scale bar ¼ 1 mm. (e) Bottom: calcrete mottles (mot) and glaebules (gl); the inter-mottle space is
occluded by drusy calcite cements; top: massive, structureless calcrete crust, scale bar ¼ 1 mm. (f) Alveolar septal structures (alv); the inter-septal space is
partially filled by drusy calcite cements, scale bar ¼ 1 mm. (g) Details of an alveolar septal structure showing the good preservation of needle calcite needles
and local partial recrystallization into smaller micrite-sized crystals, scale bar ¼ 50 mm. (h) Complex network of alveolar septal structures (alv), between
peloids (pel); the inter-septal space is partially filled by drusy calcite cements, scale bar ¼ 1 mm.
Q
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–2110
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 11
microfracturing during burial. Maximum drusy cementation
is often observed just below the water table.
5. Discussion
5.1. High-frequency sedimentary cycles
Based on the vertical succession of facies and diagenetic
features, the Late Oligocene–Early Miocene interval can be
subdivided into 3–10 m thick depositional cycles (Figs. 6
and 7). These high-frequency cycles (or parasequences) are
considered to be the smallest set of genetically related facies
formed during a single sea-level cycle. Some of these cycles
are bounded by unconformities that are, in most cases,
exposure surfaces. In this case, the term ‘high-frequency
sequence’ is applicable (Lehrmann & Goldhammer, 1999;
Van Wagoner et al., 1988).
Such sequences are usually capped by exposure surfaces
commonly associated with paleosols. Typical intertidal or
supratidal features that locally occur in the Indo-Pacific recent
1986) and sea-level oscillations related to Milankovitch
astronomical rhythms (Goldhammer, Dunn, & Hardie,
1987; Koerschner & Read, 1989). The presence of well-
developed exposure surfaces suggesting prolonged
exposures, associated with several metres-thick meteoric
vadose zones implies a relative sea-level fall and therefore
excludes the possibility of an autocyclic model. Episodic
tectonic events and/or eustatic oscillations more satisfac-
torily explain the prolonged exposure periods, followed by
platform flooding at high-frequency scales.
Figs. 8 and 9 present conceptual subtidal cycles of Late
Oligocene and Early Miocene age, recorded in inner-shelf
areas.
5.1.1. Late Oligocene cycles (Fig. 8)
These display the following features:
(1) gradual upward replacement of coralline algal facies
C1a or coralline algal–echinoderm facies C1b by coral–
Fig. 6. Simplified core log section, biotic zonation, diagenetic features, geochemical results and sequence interpretation of two Late Oligocene cores (in unit
MC1) from the well MA-5.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–2112
coralline algal–foraminiferal packstone facies C3. The
grainstone facies C2 is generally interbedded between
C1a (or C1b) and C3 or within the seagrass facies C3;
(2) upward-increase in the amount of framework-derived
elements (corals) and coarsening grain texture upward;
(3) 13C-depleted values and occurrences of scattered
pedogenic structures indicative of exposure events;
most of the Late Oligocene sequences observed in
cores are capped by an exposure surface.
The hypothesis of a permanent coral reef rim or other form
of barrier surrounding the inner-platform area is supported by
the general occurrence of relatively protected environments.
The presence of a reefal barrier is also inferred from the facies
types described in the Late Oligocene interval from the
Malampaya shelf top, the abundance of scleractinian
remains, the composition of foraminiferal assemblages and,
finally, by the overall morphology of the buildup as observed
from the seismic data. However, no true reef frameworks
have been drilled to date. A ‘keep-up’ reef barrier (i.e. coral
growth having kept pace with relative sea-level) and a ‘catch-
up’ inner-platform (i.e. deposition having caught up with
relative sea-level rise) are inferred for the Late Oligocene
time span. The deepening-upward hemi-cycles generally are
limited in thickness and correspond to an ‘empty bucket’
stage in the sense of Schlager (1981), punctuated by
occasional low oxygen conditions as suggested by the higher
uranium content near the base of the cycles (Fig. 6, at 3633.50
and 3630.00 m). The development of shallowing-up stages
during sea-level highstands is enhanced by relatively high
sediment accumulation rates that are considered to be
controlled by carbonate production in seagrass beds and,
possibly, from mid-shelf patch reefs. Seagrass environments
are known to shelter dense populations of carbonate
producers and to trap great quantities of derived particles
(Brasier, 1975). The subtidal nature of all of the facies
encountered indicates that accommodation space has not
been completely filled up during the relative sea-level
highstands; the exposure surfaces that truncate subtidal
deposits at the top of the cycles without intervening intertidal
Fig. 7. Simplified core log section, biotic zonation, diagenetic features, geochemical results and sequence interpretation of an Early Miocene cored interval
from the well MA-5. The base of unit MM3 is characterized by a water deepening event as evidenced by the nature of biological associations, and the strong
shift in C-isotope values in relation to a change of cycle type (exposure-bounded cycles or not).
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 13
or subtidal stage suggests a rapid relative sea-level drop.
During the lowstands of relative sea-level, carbonates
deposited at the shelf top were submitted to intense leaching,
cementation and pedogenesis during subaerial exposure. It is
noteworthy that possible accumulation of sediment may have
occurred inter- to supratidally but could have been removed
by erosion during the stage of subaerial exposure or during
the subsequent marine transgression. Subsequent sea-level
rise has resulted in the inundation of the platform once more
and thus favoured the deposition of subtidal material.
Fig. 8. Conceptual model of the Late Oligocene high-frequency cycles from the Malampaya inner-shelf (example of an exposure-capped cycle). (A) Model
describing the idealized cycle controlled by sea-level fluctuations (the effect of sediment compaction is not considered in this model). (B) Growth model of the
Malampaya buildup in response to short-term, relative sea-level fluctuations.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–2114
5.1.2. Early Miocene cycles (Fig. 9)
These cycles are characterized by the following features:
(1) gradual upward replacement of echinoderm and
coralline algal facies M1, by coralline algal–
echinoderm-foraminiferal facies M2a and M2b, and
by coral–coralline algal–foraminiferal facies M3
successively;
(2) increasing content of framework-derived elements
(corals) and grain coarsening from base to top;
Fig. 9. Conceptual model of the Early Miocene high-frequency cycles from the Malampaya inner-shelf (case of an exposure-capped cycle). (A) Model
describing the idealized cycle controlled by sea-level fluctuations (the effect of sediment compaction is not considered in this model). (B) Growth model of the
Malampaya buildup in response to short-term relative sea-level fluctuations.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 15
(3) common occurrence of pedogenic structures and/or
highly negative carbon isotope ratios at the top of the
sequences, indicating exposure events.
The Early Miocene sequences are not systematically
overlain by exposure surfaces; the non-exposed sequences
only exhibit a M1–M2 facies succession and the seagrass-
related facies M3 is lacking.
The occurrence of an Early Miocene coral reef barrier
remains unclear in Malampaya. The facies described on
the shelf top indicate relatively shallow water with open
marine environments at the base of the sequence
(deposition of facies M1). This basal facies is interpreted
as deposited in a transgressive system tract or an early
highstand system tract. More restricted conditions seem to
have occurred during deposition of the upper part of the
M1 facies. These conditions were enhanced when the M2
and M3 facies were deposited. The establishment of more
sheltered conditions could be related either to the
development of a reef barrier or sand shoals at the shelf
margin or to local variations in bottom topography.
However, the maximum abundance of scleractinians that
is close to the top of sequence, strongly suggests that a
reef barrier and/or reef patches grew coevally with the
later part of the HST.
The complete cycles (including the M1 –M2– M3
intervals) have been generally exposed at their tops. Similar
to the Late Oligocene deposits, the sediments may not have
aggraded to sea-level. The exposure surface is consequent
upon a rapid sea-level drop, mostly characterized by the
development of 1–5 m thick paleosoils, intense weathering
of subtidal inner-shelf carbonates and the partial occlusion
of pores by low-magnesian drusy calcite.
In the cycles devoid of subaerial signatures, the lowstand
or stillstand deposits cannot be distinguished from the
highstand deposits and are generally represented by the
coralline algal–foraminiferal-echinoderm facies M2b. In
this case, there is no evidence of sea-level drop. At the top of
the cycle, the sequence boundary is marked by a sudden
change in depositional environments recorded by the
recurrence of the M1 facies.
5.2. Subtidal nature of the cycles
Antecedent topography features, combined with water-
energy factors, probably played a key-role in the control of
sediment deposition. Three scenarios can be proposed to
support the apparent absence of inter- or supratidal deposits
in the upper parts of the Late Oligocene to Early Miocene
cycles of the Malamapaya buildups. (1) The deposition of
purely subtidal facies might result from the existence of an
energy threshold that has prevented the sediment to aggrade
above this limit and, consequently, prevented the total infill
of the available accommodation space (Osleger, 1991). The
subtidal nature of high-frequency cycles from Malampaya
therefore could reflect current control on the transport
and deposition of carbonate particles, as suggested by
Grotsch and Mercadier (1999). These authors pointed out
that the western margin of the buildup had been subjected to
strong currents and documented the existence of prevailing
winds from the North-East sector. Important offbank
transport from the buildup top to the leeward flank and the
subsequent formation of progradational lobes have been
advocated by these authors for the Oligocene interval.
(2) The absence of inter- and supratidal structures at the top
of cycles could have resulted from a fast sea-level drop that
did not allow beach-rocks or microbial mats to develop.
(3) The third alternative is that tidal deposits may have been
removed by subaerial erosion during the phase of platform
exposition and/or by marine erosion during the subsequent
phase of flooding.
5.3. High-frequency cycle stacking pattern and recognition
of composite depositional sequences in the Early Miocene
interval
Through analysis of the vertical cycle succession in the
Early Miocene interval, the high-frequency cycles have
been grouped into composite depositional sequences,
correlatable from well to well, within the inner-platform
areas. The composite depositional sequences are com-
posed of at least three high-frequency cycles and are
bounded by exposure surfaces. These sequences are
defined by sharp changes in facies and cycle types: their
lower parts are composed of non-exposed cycles, and
comprise the M1–M2 facies succession only. A marked
rise in relative sea-level is thought to be responsible for
the sharp changes in facies and cycle type. Fig. 7 shows
the base of the sequence MM3-1, at 3338 m deep, that is
also the base of unit MM3. Sedimentary units group up to
three sequences. The unit boundaries are clearly sequence
boundaries and correspond to major time-rock gaps in the
sedimentary record, as evidenced by Sr-isotope measure-
ments (Fig. 3).
Fig. 10 presents the cycle thickness evolution and the
vertical facies distribution in the upper part of well Ma-5
(upper unit MM2 and unit MM3). These plots show that the
bases of sequences can display sharp changes in cycle
thickness. However, there is no clear thinning upward
evolution and no correlation between cycle thickness, the
proportion of shallower facies (M3) versus deeper facies
(M1, M2a, M2b) in the cycle and the type of sequence top
(occurrence of an exposure surface or not). The lack of
correlation between cycle thickness and accommodation in
subtidal cycles has been published previously by Boss and
Rasmunssen (1995). Since sediments have not aggraded up
to sea-level. Variations in cycle thickness are therefore more
likely to reflect variations in the rate of carbonate
accumulation rather than variations in the accommodation
space driven by eustacy.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–2116
5.4. Period and amplitude of the high-frequency relative
sea-level cycles (Table 3)
Strontium-derived ages were calculated from the MA-1
well where there is a paucity of paleontologic control, in
particular in the Early Miocene interval. Duration of
stratigraphic intervals was estimated for units MC2, MM1,
MM2 and MM3 (Table 1) on the basis of Sr-isotope ratios
within the given interval using the regressions of Oslick et al.
(1994). The power of resolution of the 87Sr/86Sr regressions
is relatively low (^640 ky for the Early Miocene and
^1080 ky for the Late Oligocene with a 95% confidence
interval) and consequently only permits rough estimates of
time durations. The duration of the Early Miocene sedimen-
tary units using Oslick et al. (1994) regressions at site 747A
ranges from 406 ^ 1570 to 521 ^ 1526 ky. The age errors
take into account the internal error on 87Sr/86Sr measure-
ments and the error associated with Oslick et al. (1994)
regressions. By dividing the maximum calculated duration of
the unit by the number of high-frequency sequences, the
cycle periods were estimated as averaging 35 ^ 102 to
48 ^ 177 ky. In spite of the large range of uncertainty of
these estimations, the averages are compatible with Milan-
kovitch frequency bands. Orbital forcing is now known to be
important in dictating climate changes during the last 60 my
(Palike & Shackleton, 2000). The importance of eccentricity
and obliquity cycles has also been demonstrated for the Late
Oligocene and Early Miocene (Zachos, Shackleton, Reve-
tectonic control of these cycles cannot be ruled out.
Moreover, some cycles may not have been recorded, in
particular during short-term lowstands. This could result in
minimizing the number of cycles per unit, thus leading to an
overestimation of the cycle period.
The relative scatter of Sr isotopic ratios in the Late
Oligocene below the intra-Nido marker unconformity, in
particular, makes the duration of the composite sequence
difficult to determine. In this regard, the duration of
1235 ^ 2493 ky obtained for MC2 by considering the
maximal range of Sr-ages calculated for this interval, is
probably overestimated.
Due to the high uncertainty in paleo-water-depths
estimations in the basal facies (M1, C1a and C1b),
amplitudes of relative sea-level cycles are difficult to
evaluate. For the Late Oligocene, the maximal cycle
thickness of 10 m is suggested as a minimum value for
the relative sea-level cycle amplitude. For the Early
Miocene, the deepest facies (M1 or M2a–b) was probably
deposited in water-depth greater than 20 m, thus indicating a
relative sea-level cycle amplitude greater than 20 m.
5.5. Rates of accumulation
Considering the mid-point value of unit durations,
accumulation rates are estimated as averaging
0.10–0.19 m/ky for the Late Oligocene and Early Miocene
(Table 3). In Neogene carbonates from South-East Asia,
calculated accumulation rates range from 0.3 to 1 m/ky
(Wilson, 2002). For Holocene lagoons, a wide range of
values is in evidence. For example, in the Mayotte lagoon
(Western Indian Ocean), sedimentation rates vary from 0.12
to 4.39 m/ky (Zinke, Reijmer, & Thomassin, 2003). In the
shallowing-upward cycles of Davies lagoon (Eastern Aus-
tralia), rates range from 1.4 to 3.4 m/ky (Tudhope, 1989).
Fig. 10. Plot of cycle thickness and facies proportion, from the upper unit MM2 to the top of unit MM3, in the well Ma-5. Bases of composite sequences are
characterized by high-frequency cycles not bounded by an exposure surface: they also correspond to jumps in cycle thickness. However, there is no clear
correlation between cycle thickness, the nature of the sequence boundary (occurrence of an exposure surface or not) and the proportion of shallower facies
(M3) versus deeper facies (M1, M2a, M2b) in the cycle.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 17
Similar ranges of accumulation rates are given for lagoonal
environments in different locations: 0.65 m/ky in Pacific
0.5–0.6 m/ky in Belize atoll lagoons (Gischler, 2003),
0.7–2.1 m/ky in the Great Barrier Reef (Smith, Frankel, &
Jell, 1998). In the Malampaya inner-shelf, changes in
carbonate producers from the base to the top of sequence
suggest changes in production/accumulation rates during a
given cycle. The basal transgressive echinoderm–coralline
algal facies is probably related to lower production/
accumulation rates than the sequence top seagrass facies
dominated by benthic foraminifera, corals and coralline
algae. Indeed, in the modern reef environments, the
respective potential carbonate production of these organisms
is as follows: up to 9000 g m22 yr21 for corals (Heiss, 1995),
up to 3300 g m22 yr21 for coralline algae (Chisholm, 2000),
up to 900 g m22 yr21 for benthic foraminifers (Harney,
Hallock, Fletcher, & Richmond, 1999). Very few studies are
available concerning the potential carbonate production by
echinoderms. For the echinoid genus Diadema, according to
the weight and the age of adults, carbonate production rates
can be estimated to about 10 g yr21 per individual (Bauer,
1976), their density averaging commonly 4–5 individuals
per m2 in modern reef environments (Peyrot-Clausade, pers.
comm.). Values of 40–50 g yr21 m22 can be considered
reasonable estimates for potential carbonate production of
Diadema. Although such rates are probably strongly
dependent on echinoderm taxa, echinoderm–coralline
algal dominated facies are inferred to have lower potential
carbonate production than that of the coral–foraminiferal–
coralline algal facies.
These accumulation rates may be affected: (1) by the
relatively low resolution of Sr-isotope stratigraphy that
biases an estimation of cycle durations; and (2) by a period
of cycle formation that does not coincide with the period of
active sedimentation. For example, cycle durations include
the duration of exposure during sea-level lowstands.
Additionally, Grotsch and Mercadier (1999) suggested
that a large proportion of the sediment produced in the
platform top was probably exported basinward and
deposited in the form of leeward progradational wedges.
In brief, the mean rates obtained from cycles are likely to
underestimate the actual accumulation rates during the
interval of active sedimentation.
5.6. Comparison with ancient and modern analogues
The Upper Oligocene Malampaya sequence differs from
that of the Lower Miocene through the apparent absence of
attributes characterizing deposition in an open inner-
platform during sea-level transgressive and early highstand
stages. In modern reef environments, sequences related to
both stages are present and many authors have defined the
timing and development patterns during postglacial time
(last 23,000 yr).
For example, Tudhope (1989) described Holocene
shallowing-upward sedimentation in the Davies Reef
lagoon (Great Barrier Reef, Australia), as partly subject
to open marine conditions at the base (7500 yr BP) and
followed by restricted lagoonal sedimentation afterwards.
The ‘high-energy window’ evidenced at the base of the
sequence was interpreted as resulting from a rapid sea-
level rise that outstripped the capacity of the reef to
catch-up (Hopley, 1984). The onset of sheltered inner-
platform conditions occurred as soon as the open
circulation ‘window’ was closed by barrier reef accretion.
Such a scenario can be applied to the upward M1–M2
facies transition from open to protected inner-shelf
environments. The starting growth phase of the outer
reef rim may have been postponed due to a variety of
environmental constraints. Such a retardation in reef
initiation was also reported from the Holocene fringing
reefs in New Caledonia (Cabioch, Montaggioni, & Faure,
1995).
Similar to modern shallow-water carbonate environ-
ments, the differences in the depositional patterns between
the Malampaya Late Oligocene and Early Miocene
sequences might reflect differences in water quality and
conditions of larval recruitment (Cabioch, Camoin, &
Montaggioni, 1999; Montaggioni, 2000). But, the open
platform stage in the Early Miocene cycles could have
resulted from faster changes in relative sea-level rises,
possibly in conjunction with episodic tectonic events or with
higher-frequency eustatic oscillations. However, the
Table 3
Ranges of sedimentary unit durations, cycle periods and accumulation rates for units MC2, MM1, MM2 and MM3 in well MA-5
Unit Number
of cycles
Calculated deposition
ime of units using
Oslick et al. (1994)
regressions (ky)
Duration of
cycles (ky)
Average cycle
thickness (m)
Average accumulation
rate (m/ky)
MM3 9 430 ^ 1540a 48 ^ 177a 5.50 0.02b 0.11c
MM2 10 410 ^ 1570a 41 ^ 157a 4.00 0.02b 0.1c
MM1 15 520 ^ 1520a 35 ^ 102a 5.30 0.04b 0.15c
MC2 14 1240 ^ 2490a 88 ^ 178a 8.50 0.003b 0.1c
a Age errors include measurement errors and errors associated with Oslick et al. (1994) regression, considering a 95% confidence interval.b Value calculated using the maximal expected value of cycle duration.c Value calculated using the average expected value of cycle duration.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–2118
difference between Oligocene and Miocene cycles could be
related to the development of the East Asian Monsoon in the
earliest Miocene (Guo et al., 2002). The increased
storminess may have affected the nature and distribution
of barriers along the platform margins and caused
significant variations in the protected versus open signatures
in the Malampaya inner-shelf.
The most common feature of the Late Oligocene and
Early Miocene cycles is that they probably have all
developed subtidally and are frequently bounded by an
exposure surface at the top. In ancient carbonate systems,
diverse examples of such subtidal cycles have been
described: Triassic of Latemar, Italian Alps (Goldhammer
et al., 1987, 1990), Oligocene of Suwanee, Florida
(Hammes, 1992), Pleistocene of South Florida (Perkins,
1977). All of them exhibit shallowing-upward trends, and
caliche beds at the top of sequences. Such cycles are thought
to have been driven by high-frequency changes in
accommodation space, with rapid falls in sea-level.
6. Conclusions
Petrographical and geochemical analyses of subsurface
material from the Malampaya buildup provide new
information about the depositional patterns of Oligo-
Miocene carbonates from South-East Asia, particularly
with regard to the short-term evolutionary history of isolated
shallow-water carbonate systems in response to high-
frequency variations in relative sea-level.
– The vertical and lateral distribution of facies appears
to be strongly controlled by short-term relative sea-
level changes. In the inner-shelf area, the main types
of carbonate producer have varied throughout a
cycle in response to changes in water-depth and to
the degree of marine circulation exchanges with the
open sea. Connectivity to the open sea probably has
been mainly dependent on sea-level control of reef
development at the shelf margin. Relative sea-level
falls caused subaerial exposure and the consequent
intense alteration of cycle top carbonates. In this
way, short-term changes in accommodation space
controlled the reservoir properties of the Nido
Limestone.
– Despite a relatively high degree of uncertainty, the
cycles have the same time duration range as that of
orbital forcing events; however, episodic tectonic
controls could have mimiked Milankovitch-band
frequency cycles.
– The response of Malampaya Oligo-Miocene, reef-
bearing buildups to high-frequency variations in sea-
level is partly similar to that observed in Recent reef
systems, particularly in the timing of reef growth
relative to flooding of the antecedent substrate.
Acknowledgements
This work was funded by Shell Philippines Explora-
tion B.V. (SPEX). Their support and approval to publish
this paper are gratefully acknowledged. We especially
thank D. Neuhaus (SPEX) and J. Borgomano (Shell
Research International B.V., Rijwijk, The Netherland) for
the initiation of this project. This paper largely benefited
from the experience of F. Abbots-Guardiola (Shell
International, Houston), C. Mercadier, P. Cassidy and
G. Warrlich from the Shell Carbonate Team (Rijwijk,
The Netherland). We are also grateful to D. Bosence, L.
Pomar, M. Wilson, G. Conesa, J.P. Margerel and J.P.
Masse, for very helpful discussions. We thank
M. Joachimski (Erlangen University) for carbon and
oxygen isotope analyses and the Laboratory of Isotope
Geochemistry at the Vrije Universiteit (Amsterdam) for
Sr-isotope ratio measurements.
References
Adams, C. G. (1970). A reconsideration of the East Indian Letter
Classification of the Tertiary. British Museum of Natural History.
Geology, 19, 87–137.
Adams, C. G. (1984). Neogene larger Foraminifera, evolutionary and
geological events in context of the datum planes. In N. Ikebe, & R.
Tsuchi (Eds.), Pacific neogene datum planes (pp. 47–68).
Aronson, R. B., Blake, D. B., & Oji, T. (1997). Retrograde community
structure in the late Eocene of Antartica. Geology, 25(10), 903–906.
Bathurst, R. G. C. (1975). Carbonate sediments and their diagenesis.
Amsterdam: Elsevier, 658 pp.
Bauer, J. C. (1976). Growth aggregation and maturation in the echinoid
Diadema antillarum. Bulletin of Marine Science, 26(2), 273–277.
Berggren, W. A., Kent, D. V., Swisher, C. C., & Aubry, M.-P. (1995). A
revised Cenozoic geochronology and chronostratigraphy. Geochronol-
ogy, time scales and global stratigraphic correlation. SEPM Special
Publication 54, Tulsa, OK: Society for Sedimentary Geology, pp. 129–
212.
Borgomano, J., van Konijnenburg, J. H., & Jauffred, J. C. (2001). Anatomy
of carbonate bodies for hydrocarbon reservoir modeling: Applications
and future developments. Geologie Mediterraneenne, XXVIII, 1–2. (see
also pp. 23–26).
Bosence, E. D. W. J. (1983a). Coralline algal reef frameworks. Journal of
Geological Society of London, 140, 365–376.
Bosence, D. W. J. (1983b). The occurrence and ecology of recent
rhodoliths. In T. M. Peryt (Ed.), Coated grains (pp. 225–242). Berlin:
Springer.
Boss, S. K., & Rasmunssen, K. A. (1995). Misuse of Fischer plots as sea-
level curves. Geology, 23(3), 221–224.
Boudagher-Fadel, M. K., & Banner, F. T. (1999). Revision of the
stratigraphic significance of the Oligocene–Miocene ‘Letter-Stages’.
Revue de Micropaleontologie, 42(2), 93–97.
Brasier, M. D. (1975). An outline history of seagrass communities.
Palaeontology, 18, 681–702.
Briais, A., Patriat, P., & Tapponnier, P. (1993). Updated interpretation of
magnetic anomalies and sea floor spreading stages in the South China
Sea: Implications for the Tertiary tectonics of southeast Asia. Journal of
Geophysical Research, 98, 6299–6328.
Cabioch, G., Camoin, G. F., & Montaggioni, L. F. (1999). Postglacial
growth history of a French Polynesian barrier reef tract, Tahiti, central
Pacific. Sedimentology, 46, 985–1000.
F. Fournier et al. / Marine and Petroleum Geology 21 (2004) 1–21 19
Cabioch, G., Montaggioni, L. F., & Faure, G. (1995). Holocene initiation
and development of New Caledonian fringing reefs, South-West
Pacific. Coral Reefs, 14, 131–140.
Chaproniere, G. C. H. (1975). Palaeoecology of Oligo-Miocene Forami-